Universal Serial Bus

Device Class Definition

for
Video Devices

Revision 1.5
August 9, 2012
 USB Device Class Definition for Video Devices

Contributors
Hans van Antwerpen Cypress Semiconductor
Eric Luttmann Cypress Semiconductor
David Roh Dolby Labs
Choon Chng Google Inc.
Pawel Osciak Google Inc.
Ville-Mikko Rautio Google Inc.
Van Duros Immedia Semiconductor Inc.
Abdul R. Ismail Intel Corp.
Bradley Saunders Intel Corp.
Ygal Blum Jungo
Yoav Nissim Jungo
Jean-Michel Chardon Logitech Inc.
Olivier Lechenne Logitech Inc.
Geraud Mudry Logitech Inc.
Chandrashekhar Rao Logitech Inc.
Remy Zimmermann Logitech Inc.
Chris Yokum MCCI Corporation
Stephen Cooper Microsoft Corp.
Maribel Figuera Microsoft Corp.
Richard Webb Microsoft Corp.
Anand Ganesh Microsoft Corp.
David Goll Microsoft Corp.
Hiro Kobayashi Microsoft Corp.
Bertrand Lee Microsoft Corp.
Jeff Zhu Microsoft Corp.
Andrei Jefremov Microsoft Corp.
Tim Vlaar Point Grey Research
Mark Bohm SMSC
John Sisto SMSC
Will Harris Texas Instruments
Grant Ley Texas Instruments
Anshuman Saxena Texas Instruments
Paul E. Berg USB-IF

Revision 1.5 August 9, 2012 ii

 USB Device Class Definition for Video Devices

Revision 1.5 August 9, 2012 iii

 USB Device Class Definition for Video Devices

Copyright © 2012, USB Implementers Forum, Inc.

All rights reserved.

A LICENSE IS HEREBY GRANTED TO REPRODUCE THIS SPECIFICATION FOR

INTERNAL USE ONLY. NO OTHER LICENSE, EXPRESS OR IMPLIED, BY
ESTOPPEL OR OTHERWISE, IS GRANTED OR INTENDED HEREBY.

USB-IF AND THE AUTHORS OF THIS SPECIFICATION EXPRESSLY DISCLAIM

ALL LIABILITY FOR INFRINGEMENT OF INTELLECTUAL PROPERTY RIGHTS,
RELATING TO IMPLEMENTATION OF INFORMATION IN THIS
SPECIFICATION. USB-IF AND THE AUTHORS OF THIS SPECIFICATION ALSO
DO NOT WARRANT OR REPRESENT THAT SUCH IMPLEMENTATION(S) WILL
NOT INFRINGE THE INTELLECTUAL PROPERTY RIGHTS OF OTHERS .

THIS SPECIFICATION IS PROVIDED "AS IS” AND WITH NO WARRANTIES,

EXPRESS OR IMPLIED, STATUTORY OR OTHERWISE. ALL WARRANTIES ARE
EXPRESSLY DISCLAIMED. NO WARRANTY OF MERCHANTABILITY, NO
WARRANTY OF NON-INFRINGEMENT, NO WARRANTY OF FITNESS FOR ANY
PARTICULAR PURPOSE, AND NO WARRANTY ARISING OUT OF ANY
PROPOSAL, SPECIFICATION, OR SAMPLE.

IN NO EVENT WILL USB-IF OR USB-IF MEMBERS BE LIABLE TO ANOTHER FOR

THE COST OF PROCURING SUBSTITUTE GOODS OR SERVICES, LOST PROFITS,
LOSS OF USE, LOSS OF DATA OR ANY INCIDENTAL, CONSEQUENTIAL,
INDIRECT, OR SPECIAL DAMAGES, WHETHER UNDER CONTRACT, TORT,
WARRANTY, OR OTHERWISE, ARISING IN ANY WAY OUT OF THE USE OF THIS
SPECIFICATION, WHETHER OR NOT SUCH PARTY HAD ADVANCE NOTICE OF
THE POSSIBILITY OF SUCH DAMAGES.

All product names are trademarks, registered trademarks, or service marks of their respective owners.

Please send comments via electronic mail to <video-chair>@usb.org

Revision 1.5 August 9, 2012 iv

 USB Device Class Definition for Video Devices

Revision History

Version Date Description

1.0 September 4, Initial release
2003
1.0a December 4, Table A.5: Added extension reference row
2003 Table A.6: Added USB extension row
Table A.6: Defined VS_FORMAT_MPEG4SL as 0x0B
Added section A,9.8 for selectors extensibility
Table B.1, B.2, B3 & B4: Added extension row
Added Appendix D. Revision history
1.1 June 1st, 2005 Added GET_INFO request for VideoStreaming interface
Updated Table 4-56 Stream Error Code Control
Updated Table 3-8 to correct the wMaxMultiplier description
Add support for multiple clock frequencies (RR0033)
Latency optimizations for Stream-based formats (RR0041)
Define Probe/Commit controls for Render (RR0042)
Add Analog Video Standard and Status Control (RR0044)
Define constraints on RES value (RR0047)
Detail behavior of Multiplier and Limit Multiplier (RR0048)
Cosmetic changes to Terminal and Unit Descriptors (RR0049)
Cosmetic and Functional changes(RR0050)
Clarify usage of bits in GET_INFO (RR0051)
Specify the meanings of Request Error Codes (RR0053)
Allows support for a revision of a Payload specification to be made
independently of the Core Specification.
Added three fields to the Video Probe And Commit Control
(RR0054).
Specification of Absolute and Relative Control relationship.
(RR057).
Specification of Asynchronous Controls’ behavior (RR0059).
Allow 0 for Payload Version in Probe and Commit Control
(RR0060).
Modified Table 2-6 Extended Field of the Payload Header, Table 3-
15 Payload Format Descriptor and Section 2.4.2.4 Still Image
Capture. (RR0066)
Remove “Driver” from Terms and abbreviation and added UVC.
Updated section 2.4.3.7, change VDC to UVC. Update section 3.7.2
and Table 3-3: Change bcdVDC with bcdUVC. (RR0064)
Corrected Statement regarding multiple clock support (RR0069).
Removed auto-update side-effect from Probe/Commit Frame
Interval field (RR0070).
Updated Descriptor Size inTable 3-8 Processing Unit Descriptor
(RR0072).
Updated range of reserved values in Table 4-43 Analog Video

Revision 1.5 August 9, 2012 v

 USB Device Class Definition for Video Devices

Standard Control (RR0072).

Remove reference to “Vendor Unique Payload Format” in Table 4-
47 Video Probe and Commit Controls (RR0072).
Marked obsolete format and Frame descriptor type values as
reserved (VS_FORMAT_MPEG1, VS_FORMAT_MPEG2PS,
VS_FORMAT_MPEG4, VS_FORMAT_VENDOR,
VS_FRAME_VENDOR) in Table A- 6 Video Class-Specific VS
Interface Descriptor Subtypes (RR0072).
Added new format and Frame descriptor types
(VS_FORMAT_FRAME_BASED,VS_FRAME_FRAME_BASED,
VS_FORMAT_STREAM_BASED) in Table A- 6 Video Class-
Specific VS Interface Descriptor Subtypes (RR0072).
Added new Processing Unit Controls
(PU_ANALOG_VIDEO_STANDARD_CONTROL,
PU_ANALOG_LOCK_STATUS_CONTROL) in Table A- 13
Processing Unit Control Selectors (RR0072).
1.5 June 6, 2012 Updated bcdUVC to 1.5.
Updated Probe and Commit to include features of temporally
encoded video that impact bandwidth negotiation, increasing the
size of the VS_PROBE_CONTROL and
VS_COMMIT_CONTROL.
Added twenty Encoding Unit controls to facilitate control of device
based video encoders.
Added three new controls to the Camera Terminal:
CT_FOCUS_SIMPLE_CONTROL, CT _WINDOW_CONTROL,
and CT_REGION_OF_INTEREST_CONTROL.
Added one new control to the Processing Unit:
PU_CONTRAST_AUTO_CONTROL.
Added solution for backward compatibility of UVC 1.5 devices.
Added bit to GET_INFO to declare control as “Disabled for current
Commit state”
Removed reference section Device_Qualifier Descriptor and
Other_Speed_Configuration Descriptor.
Updated Data Format Classes (section 2.4.3.7) with temporally
encoded video formats.
Added Notations section.
Added references to USB Specification 3.0.
Moved discussion on Optical Zoom to appendix.

Revision 1.5 August 9, 2012 vi

 USB Device Class Definition for Video Devices

Table of Contents
1 Introduction ............................................................................................................................. 1
1.1 Purpose ............................................................................................................................. 1
1.2 Scope ................................................................................................................................ 1
1.3 Related Documents .......................................................................................................... 1
1.4 Document Conventions .................................................................................................... 1
1.4.1 Notations .................................................................................................................. 2
1.5 Terms and Abbreviations ................................................................................................. 2
2 Functional Characteristics ....................................................................................................... 5
2.1 Video Interface Class ....................................................................................................... 5
2.2 Video Interface Subclass and Protocol............................................................................. 5
2.3 Video Function Topology ................................................................................................ 6
2.3.1 Input Terminal .......................................................................................................... 8
2.3.2 Output Terminal ....................................................................................................... 9
2.3.3 Camera Terminal ...................................................................................................... 9
2.3.4 Selector Unit ........................................................................................................... 10
2.3.5 Processing Unit....................................................................................................... 10
2.3.6 Encoding Unit......................................................................................................... 11
2.3.7 Extension Unit ........................................................................................................ 12
2.4 Operational Model.......................................................................................................... 12
2.4.1 Video Interface Collection ..................................................................................... 13
2.4.2 VideoControl Interface ........................................................................................... 13
2.4.2.1 Control Endpoint ............................................................................................. 14
2.4.2.2 Status Interrupt Endpoint ................................................................................. 14
2.4.2.3 Hardware Trigger Interrupts ............................................................................ 17
2.4.2.4 Still Image Capture .......................................................................................... 17
2.4.2.5 Optical Zoom vs Digital Zoom ........................................................................ 18
2.4.3 VideoStreaming Interface....................................................................................... 18
2.4.3.1 Stream Bandwidth Selection ........................................................................... 19
2.4.3.2 Video and Still Image Samples ....................................................................... 20
2.4.3.2.1 Sample Bulk Transfers ................................................................................. 23
2.4.3.2.2 Sample Isochronous Transfers...................................................................... 25
2.4.3.3 Video and Still Image Payload Headers .......................................................... 29
2.4.3.4 Stream Synchronization and Rate Matching ................................................... 32
2.4.3.4.1 Latency ......................................................................................................... 32
2.4.3.4.2 Clock Reference ........................................................................................... 33
2.4.3.4.3 Presentation Time ......................................................................................... 33
2.4.3.5 Dynamic Frame Interval Support .................................................................... 34
2.4.3.6 Device Initiated Dynamic Format Change Support ........................................ 34
2.4.3.7 Data Format Classes ........................................................................................ 34
2.4.4 Control Transfer and Request Processing .............................................................. 35
3 Descriptors ............................................................................................................................. 43
3.1 Descriptor Layout Overview .......................................................................................... 44
3.2 Device Descriptor ........................................................................................................... 44
3.3 Device_Qualifier Descriptor (deprecated) ..................................................................... 45
3.4 Configuration Descriptor................................................................................................ 45

Revision 1.5 August 9, 2012 vii

 USB Device Class Definition for Video Devices

3.4.1 Backwards Compatibility ....................................................................................... 45

3.5 Other_Speed_Configuration Descriptor (deprecated) .................................................... 45
3.6 Interface Association Descriptor .................................................................................... 45
3.7 VideoControl Interface Descriptors ............................................................................... 46
3.7.1 Standard VC Interface Descriptor .......................................................................... 46
3.7.2 Class-Specific VC Interface Descriptor ................................................................. 47
3.7.2.1 Input Terminal Descriptor ............................................................................... 49
3.7.2.2 Output Terminal Descriptor............................................................................. 50
3.7.2.3 Camera Terminal Descriptor ........................................................................... 51
3.7.2.4 Selector Unit Descriptor .................................................................................. 53
3.7.2.5 Processing Unit Descriptor .............................................................................. 54
3.7.2.6 Encoding Unit Descriptor ................................................................................ 56
3.7.2.7 Extension Unit Descriptor ............................................................................... 58
3.8 VideoControl Endpoint Descriptors ............................................................................... 59
3.8.1 VC Control Endpoint Descriptors .......................................................................... 59
3.8.1.1 Standard VC Control Endpoint Descriptor ...................................................... 59
3.8.1.2 Class-Specific VC Control Endpoint Descriptor ............................................. 59
3.8.2 VC Interrupt Endpoint Descriptors ........................................................................ 59
3.8.2.1 Standard VC Interrupt Endpoint Descriptor .................................................... 59
3.8.2.2 Class-specific VC Interrupt Endpoint Descriptor ............................................ 60
3.9 VideoStreaming Interface Descriptors ........................................................................... 60
3.9.1 Standard VS Interface Descriptor........................................................................... 61
3.9.2 Class-Specific VS Interface Descriptors ................................................................ 61
3.9.2.1 Input Header Descriptor .................................................................................. 61
3.9.2.2 Output Header Descriptor ................................................................................ 64
3.9.2.3 Payload Format Descriptors ............................................................................ 65
3.9.2.4 Video Frame Descriptor .................................................................................. 66
3.9.2.5 Still Image Frame Descriptor .......................................................................... 66
3.9.2.6 Color Matching Descriptor .............................................................................. 68
3.10 VideoStreaming Endpoint Descriptors........................................................................... 69
3.10.1 VS Video Data Endpoint Descriptors .................................................................... 69
3.10.1.1 Standard VS Isochronous Video Data Endpoint Descriptor ........................... 69
3.10.1.2 Standard VS Bulk Video Data Endpoint Descriptor ....................................... 70
3.10.2 VS Bulk Still Image Data Endpoint Descriptors .................................................... 71
3.10.2.1 Standard VS Bulk Still Image Data Endpoint Descriptor ............................... 71
3.11 String Descriptors ........................................................................................................... 72
4 Class-Specific Requests ......................................................................................................... 73
4.1 Request Layout ............................................................................................................... 73
4.1.1 Set Request ............................................................................................................. 73
4.1.2 Get Request ............................................................................................................ 74
4.2 VideoControl Requests .................................................................................................. 76
4.2.1 Interface Control Requests ..................................................................................... 77
4.2.1.1 Power Mode Control ....................................................................................... 77
4.2.1.2 Request Error Code Control ............................................................................ 79
4.2.2 Unit and Terminal Control Requests ...................................................................... 80
4.2.2.1 Camera Terminal Control Requests ................................................................ 81

Revision 1.5 August 9, 2012 viii

 USB Device Class Definition for Video Devices

4.2.2.1.1 Scanning Mode Control ................................................................................ 81

4.2.2.1.2 Auto-Exposure Mode Control ...................................................................... 82
4.2.2.1.3 Auto-Exposure Priority Control ................................................................... 82
4.2.2.1.4 Exposure Time (Absolute) Control .............................................................. 83
4.2.2.1.5 Exposure Time (Relative) Control ............................................................... 83
4.2.2.1.6 Focus (Absolute) Control ............................................................................. 84
4.2.2.1.7 Focus (Relative) Control .............................................................................. 84
4.2.2.1.8 Focus, Simple Range .................................................................................... 85
4.2.2.1.9 Focus, Auto Control ..................................................................................... 86
4.2.2.1.10 Iris (Absolute) Control................................................................................ 86
4.2.2.1.11 Iris (Relative) Control ................................................................................. 87
4.2.2.1.12 Zoom (Absolute) Control ........................................................................... 87
4.2.2.1.13 Zoom (Relative) Control ............................................................................ 88
4.2.2.1.14 PanTilt (Absolute) Control ......................................................................... 89
4.2.2.1.15 PanTilt (Relative) Control .......................................................................... 89
4.2.2.1.16 Roll (Absolute) Control .............................................................................. 91
4.2.2.1.17 Roll (Relative) Control ............................................................................... 91
4.2.2.1.18 Privacy Control ........................................................................................... 92
4.2.2.1.19 Digital Window Control ............................................................................. 92
4.2.2.1.20 Digital Region of Interest (ROI) Control ................................................... 93
4.2.2.2 Selector Unit Control Requests ....................................................................... 94
4.2.2.3 Processing Unit Control Requests ................................................................... 94
4.2.2.3.1 Backlight Compensation Control ................................................................. 96
4.2.2.3.2 Brightness Control ........................................................................................ 96
4.2.2.3.3 Contrast Control ........................................................................................... 96
4.2.2.3.4 Contrast, Auto Control ................................................................................. 97
4.2.2.3.5 Gain Control ................................................................................................. 97
4.2.2.3.6 Power Line Frequency Control..................................................................... 98
4.2.2.3.7 Hue Control .................................................................................................. 98
4.2.2.3.8 Hue, Auto Control ........................................................................................ 98
4.2.2.3.9 Saturation Control......................................................................................... 99
4.2.2.3.10 Sharpness Control ....................................................................................... 99
4.2.2.3.11 Gamma Control ........................................................................................ 100
4.2.2.3.12 White Balance Temperature Control ........................................................ 100
4.2.2.3.13 White Balance Temperature, Auto Control .............................................. 100
4.2.2.3.14 White Balance Component Control .......................................................... 101
4.2.2.3.15 White Balance Component, Auto Control ................................................ 102
4.2.2.3.16 Digital Multiplier Control ......................................................................... 102
4.2.2.3.17 Digital Multiplier Limit Control ............................................................... 103
4.2.2.3.18 Analog Video Standard Control ............................................................... 103
4.2.2.3.19 Analog Video Lock Status Control .......................................................... 103
4.2.2.4 Encoding Units .............................................................................................. 104
4.2.2.4.1 Encoding Units Operational Model ............................................................ 104
4.2.2.4.2 Select Layer Control ................................................................................... 112
4.2.2.4.3 Video Resolution Control ........................................................................... 114
4.2.2.4.4 Profile and Toolset Control ........................................................................ 114

Revision 1.5 August 9, 2012 ix

 USB Device Class Definition for Video Devices

4.2.2.4.5 Minimum Frame Interval Control .............................................................. 115

4.2.2.4.6 Slice Mode Control ..................................................................................... 116
4.2.2.4.7 Rate Control Mode Control ........................................................................ 116
4.2.2.4.8 Average Bit Rate Control ........................................................................... 120
4.2.2.4.9 CPB Size Control........................................................................................ 120
4.2.2.4.10 Peak Bit Rate Control ............................................................................... 121
4.2.2.4.11 Quantization Parameter Control ............................................................... 121
4.2.2.4.12 Quantization Parameter Range Control .................................................... 122
4.2.2.4.13 Synchronization and Long Term Reference Frame Control .................... 123
4.2.2.4.14 Long-Term Buffer Control ....................................................................... 125
4.2.2.4.15 Long-Term Reference Picture Control ..................................................... 126
4.2.2.4.16 Long-Term Reference Validation Control ............................................... 127
4.2.2.4.17 SEI Messages Control .............................................................................. 127
4.2.2.4.18 Priority Control ......................................................................................... 129
4.2.2.4.19 Start or Stop Layer Control ...................................................................... 129
4.2.2.4.20 Level IDC Control .................................................................................... 130
4.2.2.4.21 Error Resiliency Control ........................................................................... 130
4.2.2.5 Extension Unit Control Requests .................................................................. 131
4.3 VideoStreaming Requests ............................................................................................ 132
4.3.1 Interface Control Requests ................................................................................... 132
4.3.1.1 Video Probe and Commit Controls ............................................................... 133
4.3.1.1.1 Probe and Commit Operational Model ....................................................... 144
4.3.1.1.2 Stream Negotiation Examples .................................................................... 146
4.3.1.2 Video Still Probe Control and Still Commit Control .................................... 150
4.3.1.3 Synch Delay Control ..................................................................................... 152
4.3.1.4 Still Image Trigger Control ........................................................................... 152
4.3.1.5 Generate Key Frame Control ......................................................................... 153
4.3.1.6 Update Frame Segment Control .................................................................... 153
4.3.1.7 Stream Error Code Control ............................................................................ 154
Appendix A. Video Device Class Codes .................................................................................... 156
A.1. Video Interface Class Code ............................................................................................. 156
A.2. Video Interface Subclass Codes ...................................................................................... 156
A.3. Video Interface Protocol Codes ...................................................................................... 156
A.4. Video Class-Specific Descriptor Types .......................................................................... 156
A.5. Video Class-Specific VC Interface Descriptor Subtypes ................................................ 157
A.6. Video Class-Specific VS Interface Descriptor Subtypes ................................................ 157
A.7. Video Class-Specific Endpoint Descriptor Subtypes ...................................................... 158
A.8. Video Class-Specific Request Codes .............................................................................. 158
A.9. Control Selector Codes .................................................................................................... 158
Appendix B. Terminal Types ...................................................................................................... 162
B.1. USB Terminal Types ....................................................................................................... 162
B.2. Input Terminal Types ...................................................................................................... 162
B.3. Output Terminal Types .................................................................................................... 163
B.4. External Terminal Types ................................................................................................. 163
Appendix C. Video and Still Image Formats .............................................................................. 164
C.1. Supported video and still image formats ......................................................................... 164

Revision 1.5 August 9, 2012 x

 USB Device Class Definition for Video Devices

C.2. Proprietary video formats ................................................................................................ 164

Appendix D. Optical and Digital Zoom ...................................................................................... 165
D.1. Optical Zoom ................................................................................................................... 165
D.2. Digital Zoom ................................................................................................................... 167
D.3. Relationship between Optical and Digital Zoom ............................................................ 169
D.4. Absolute vs. Relative Zoom ............................................................................................ 169

Revision 1.5 August 9, 2012 xi

 USB Device Class Definition for Video Devices

List of Tables
Table 2-1 Status Packet Format 15
Table 2-2 Status Packet Format (VideoControl Interface as the Originator) 16
Table 2-3 Status Packet Format (VideoStreaming Interface as the Originator) 16
Table 2-4 Summary of Still Image Capture Methods 18
Table 2-5 Format of the Payload Header 29
Table 2-6 Extended Fields of the Payload Header 30
Table 3-1 Standard Video Interface Collection IAD 46
Table 3-2 Standard VC Interface Descriptor 47
Table 3-3 Class-specific VC Interface Header Descriptor 48
Table 3-4 Input Terminal Descriptor 50
Table 3-5 Output Terminal Descriptor 51
Table 3-6 Camera Terminal Descriptor 52
Table 3-7 Selector Unit Descriptor 53
Table 3-8 Processing Unit Descriptor 54
Table 3-10 Extension Unit Descriptor 58
Table 3-11 Standard VC Interrupt Endpoint Descriptor 59
Table 3-12 Class-specific VC Interrupt Endpoint Descriptor 60
Table 3-13 Standard VS Interface Descriptor 61
Table 3-14 Class-specific VS Interface Input Header Descriptor 62
Table 3-15 Class-specific VS Interface Output Header Descriptor 64
Table 3-16 Payload Format Descriptor 65
Table 3-17 Defined Video Frame Descriptor Resources 66
Table 3-18 Still Image Frame Descriptor 67
Table 3-19 Color Matching Descriptor 68
Table 3-20 Standard VS Isochronous Video Data Endpoint Descriptor 70
Table 3-21 Standard VS Bulk Video Data Endpoint Descriptor 70
Table 3-22 Standard VS Bulk Still Image Data Endpoint Descriptor 71
Table 4-1 Set Request 73
Table 4-2 Get Request 74
Table 4-3 Defined Bits Containing Capabilities of the Control 75
Table 4-4 Interface Control Requests 77
Table 4-5 Power Mode Control 77
Table 4-6 Device Power Mode 78
Table 4-7 Request Error Code Control 79
Table 4-8 Unit and Terminal Control Requests 80
Table 4-9 Scanning Mode Control 82
Table 4-10 Auto-Exposure Mode Control 82
Table 4-11 Auto-Exposure Priority Control 83
Table 4-12 Exposure Time (Absolute) Control 83
Table 4-13 Exposure Time (Relative) Control 84
Table 4-14 Focus (Absolute) Control 84
Table 4-15 Focus (Relative) Control 85
Table 4-17 Focus, Auto Control 86
Table 4-18 Iris (Absolute) Control 86
Table 4-19 Iris (Relative) Control 87

Revision 1.5 August 9, 2012 xii

 USB Device Class Definition for Video Devices

Table 4-20 Zoom (Absolute) Control 87

Table 4-21 Zoom (Relative) Control 88
Table 4-22 PanTilt (Absolute) Control 89
Table 4-23 PanTilt (Relative) Control 90
Table 4-24 Roll (Absolute) Control 91
Table 4-25 Roll (Relative) Control 91
Table 4-26 Privacy Shutter Control 92
Table 4-29 Selector Unit Control Requests 94
Table 4-30 Backlight Compensation Control 96
Table 4-31 Brightness Control 96
Table 4-32 Contrast Control 96
Table 4-34 Gain Control 97
Table 4-35 Power Line Frequency Control 98
Table 4-36 Hue Control 98
Table 4-37 Hue, Auto Control 99
Table 4-38 Saturation Control 99
Table 4-39 Sharpness Control 99
Table 4-40 Gamma Control 100
Table 4-41 White Balance Temperature Control 100
Table 4-42 White Balance Temperature, Auto Control 100
Table 4-43 White Balance Component Control 101
Table 4-44 White Balance Component, Auto Control 102
Table 4-45 Digital Multiplier Control 102
Table 4-46 Digital Multiplier Limit Control 103
Table 4-47 Analog Video Standard Control 103
Table 4-48 Analog Video Lock Status Control 103
Table 4-73 Extension Unit Control Requests 131
Table 4-74 Interface Control Requests inside a Particular VideoStreaming Interface 133
Table 4-75 Video Probe and Commit Controls 134
Table 4-76 VS_PROBE_CONTROL Requests 145
Table 4-77 VS_COMMIT_CONTROL Requests 146
Table 4-78 Video Still Probe Control and Still Commit Control 150
Table 4-79 VS_STILL_PROBE_CONTROL Requests 151
Table 4-80 VS_STILL_COMMIT_CONTROL Requests 151
Table 4-81 Synch Delay Control 152
Table 4-82 Still Image Trigger Control 152
Table 4-83 Generate Key Frame Control 153
Table 4-84 Update Frame Segment Control 154
Table 4-85 Stream Error Code Control 154
Table A- 1 Video Interface Class Code 156
Table A- 2 Video Interface Subclass Codes 156
Table A- 3 Video Interface Protocol Codes 156
Table A- 4 Video Class-Specific Descriptor Types 156
Table A- 5 Video Class-Specific VC Interface Descriptor Subtypes 157
Table A- 6 Video Class-Specific VS Interface Descriptor Subtypes 157
Table A- 7 Video Class-Specific Endpoint Descriptor Subtypes 158

Revision 1.5 August 9, 2012 xiii

 USB Device Class Definition for Video Devices

Table A- 8 Video Class-Specific Request Codes 158

Table A- 9 VideoControl Interface Control Selectors 159
Table A- 10 Terminal Control Selectors 159
Table A- 11 Selector Unit Control Selectors 159
Table A- 12 Camera Terminal Control Selectors 159
Table A- 13 Processing Unit Control Selectors 160
Table A- 15 Extension Unit Control Selectors 161
Table A- 16 VideoStreaming Interface Control Selectors 161
Table B- 1 USB Terminal Types 162
Table B- 2 Input Terminal Types 162
Table B- 3 Output Terminal Types 163
Table B- 4 External Terminal Types 163

Revision 1.5 August 9, 2012 xiv

 USB Device Class Definition for Video Devices

List of Figures
Figure 2-3 Input Terminal Icon 9
Figure 2-4 Output Terminal Icon 9
Figure 2-5 Selector Unit Icon (2 input pins) 10
Figure 2-6 Processing Unit Icon 11
Figure 2-8 Extension Unit Icon 12
Figure 2-9 Stream Bandwidth Selection 20
Figure 2-10 Protocol Layering and Abstraction 21
Figure 2-11 A Payload Transfer 22
Figure 2-12 Sample Bulk Read (Multiple Transfers per Sample) 23
Figure 2-13 Sample Bulk Read (Single Transfer per Sample) 24
Figure 2-14 Sample Bulk Write (Single Transfer per Sample) 24
Figure 2-15 Sample Isochronous Transfer, IN endpoint 25
Figure 2-16 Sample Isochronous Transfer, OUT endpoint 26
Figure 2-17 Sample Isochronous Transfer, IN endpoint 27
Figure 2-18 Sample Isochronous Transfer, OUT endpoint 28
Figure 2-19 Control Transfer Example (Case 1) 38
Figure 2-20 Control Transfer Example (Case 2) 39
Figure 2-21 Control Transfer Example (Case 3) 40
Figure 2-22 Control Transfer Example (Case 4) 41
Figure 2-23 Control Transfer Example (Case 5) 42
Figure 3-1 Video Camera Descriptor Layout Example 44
Figure 4-5 Successful USB Isochronous Bandwidth Negotiation 147
Figure 4-6 Failed USB Isochronous Bandwidth Negotiation 148
Figure 4-7 Dynamic Stream Settings Modification while Streaming 149

Revision 1.5 August 9, 2012 xv

 USB Device Class Definition for Video Devices

1 Introduction

1.1 Purpose
This document describes the minimum capabilities and characteristics that a video streaming
device must support to comply with the USB Video Class specification.

It defines and standardizes video streaming functionality on the USB, and contains all necessary
information for a designer to build a USB-compliant device that incorporates video streaming
functionality. It specifies the standard and class-specific descriptors that must be present in each
USB video function. It further explains the use of class-specific requests that allow for full video
streaming control. Finally, it explains how devices can be compliant with multiple versions of
this specification to enable backwards compatibility.

Devices that conform to this specification will be referred to as USB Video Class devices.

1.2 Scope
The USB Device Class Definition for Video Devices applies to all devices or functions within
composite devices that are used to manipulate video and video-related functionality. This would
include devices such as desktop video cameras (or "webcams"), digital camcorders, analog video
converters, analog and digital television tuners, and still-image cameras that support video
streaming. This specification also applies to Video Devices that compress video using temporal
encoders.

1.3 Related Documents

USB Specification Revision 3.0, November 12, 2008, www.usb.org
USB Specification Revision 2.0, April 27, 2000, www.usb.org
Interface Association Descriptor ECN, www.usb.org

1.4 Document Conventions

The following typographic conventions are used:
 Italic Documents references
 Bold Request fields
 UPPERCASE Constants

The following terms are defined:

 Expected
a keyword used to describe the behavior of the hardware or software in the design models
assumed by this specification. Other hardware and software design models may also be
implemented
 May/Could
a keyword that indicates flexibility of choice with no implied preference.

Revision 1.5 August 9, 2012 1

 USB Device Class Definition for Video Devices

 Shall/Must
keywords indicating a mandatory requirement. Designers are required to implement all such
mandatory requirements.
 Should
a keyword indicating flexibility of choice with a strongly preferred alternative. Equivalent to
the phrase is recommended.

1.4.1 Notations
The following notations are used in this specification and all associated video payload and
example documents.
Notation Description
SET_INTERFACE (n) This notation indicates a SET_INTERFACE request as
defined section 9.4.10 of USB Specification 2.0 and USB
Specification 3.0 where wValue = n.
Control_Name(request_type) This notation indicates a specific request_type being
issued to a specific Control_Name.
EU_*_CONTROL(request_type) This notation indicates a specific request_type issued to
any Encoding Unit.

1.5 Terms and Abbreviations

Term Description
Configuration A collection of one or more interfaces that may be selected on a USB
device.
Control A logical object within an Entity that is used to manipulate a specific
property of that Entity.
CT Camera terminal.
Descriptor Data structure used to describe a USB device capability or
characteristic.
Device USB peripheral.
Endpoint Source or sink of data on a USB device.
Entity A Unit, Terminal or Interface within the video function, each of which
may contain Controls.
EOS End of Slice. A bit marker in the video frame header that indicates the
associated video payload contains the last bit in the current Slice.
GUID Globally Unique Identifier. Also known as a universally unique
identifier (UUID). The Guidgen.exe command line program from
Microsoft is used to create a GUID. Guidgen.exe never produces the
same GUID twice, no matter how many times it is run or how many
different machines it runs on. Entities such as video formats that need to
be uniquely identified have a GUID. Search www.microsoft.com for
more information on GUIDs and Guidgen.exe.

Revision 1.5 August 9, 2012 2

 USB Device Class Definition for Video Devices

Host Computer system where a Host Controller is installed.

Host Controller Hardware that connects a Host to the USB.
Host Software Generic term for a collection of drivers, libraries and/or applications that
provide operating system support for a device.
IAD Interface Association Descriptor. This is used to describe that two or
more interfaces are associated to the same function. An ‘association’
includes two or more interfaces and all of their alternate setting
interfaces.
IDR Instantaneous Decoder Refresh. An I-Frame where no frames after the
IDR frame can rely on frames before the IDR frame.
Interface An Entity representing a collection of zero or more endpoints that
present functionality to a Host.
IT Input Terminal.
Dependency Layer The same as dependency representation in the H.264 specification and
in the VP8 payload specification.
Multicast Streaming Two or more independent video streams originating from the same
source (Video Function) sent over separate video streaming interfaces..
OT Output Terminal.
Payload Transfer In the context of the USB Video Class, a Payload Transfer is a unit of
data transfer common to bulk and isochronous endpoints. Each Payload
Transfer includes a Payload Header followed by Payload Data. For
isochronous endpoints, a Payload Transfer is contained in the data
transmitted during a single (micro) frame: up to 1023 bytes for a full-
speed endpoint; up to 1024 bytes for a high-speed endpoint; and up to
3072 bytes for a high-speed/high-bandwidth endpoint. For bulk
endpoints, a Payload Transfer is contained in the data transmitted in a
single bulk transfer (which may consist of multiple bulk data
transactions).
Payload Data Format-specific data contained in a Payload Transfer (excluding the
Payload Header).
Payload Header A header at the start of each Payload Transfer that provides data framing
and encapsulation information.
PTS Presentation Time Stamp. This is the source clock time in native device
clock units when the raw frame capture begins.
PU Processing Unit.
Request A mechanism supported by the video function for the host software to
interact with a Control within an Entity.
QP Quantization Parameter. This is an index used to derive a scaling matrix
for video encoding.
Sample Transfer A sample transfer is composed of one or more payload transfer(s)
representing a video sample.
SCR Source Clock Reference. This is a two part value that contains (1) the
source clock time when the video sample leaves the device and (2) the

Revision 1.5 August 9, 2012 3

 USB Device Class Definition for Video Devices

current 1 KHz SOF counter.

SEI Supplemental Enhancement Information as defined in the H.264
specification.
Simulcast Stream Simulcast Streams are multiple concurrent, independently coded bit
streams from the same source streamed over single Video Streaming
Interface.
SOF Start of Frame. Transmitted by the USB host every millisecond, the SOF
token contains an 11 bit incremental frame number.
STC Source Time Clock. The clock used by the data source that governs the
sampling of video (or related) data.
SU Selector Unit.
TD Terminal Descriptor.
Terminal An Entity representing a source (Input Terminal) or sink (Output
Terminal) for data flowing into or out of a video function.
UD Unit Descriptor.
Unit An Entity representing a transformation of data flowing through a video
function.
USB Universal Serial Bus.
USB Transaction See USB 2.0 Chapter 5.
USB Transfer See USB 2.0 Chapter 5.
UVC USB Video Class.
VC VideoControl; refers to the interface used for video function control.
VIC Video Interface Collection; refers to the collection of VideoControl and
VideoStreaming interfaces within the same video function.
VS VideoStreaming; refers to the interface(s) used for video stream
transport.
XU Extension Unit.

Revision 1.5 August 9, 2012 4

 USB Device Class Definition for Video Devices

2 Functional Characteristics
The video function is located at the interface level in the device class hierarchy. It consists of a
number of interfaces grouping related pipes that together implement the interface to the video
function.

Video functions are addressed through their video interfaces. Each video function has a single
VideoControl (VC) interface and can have several VideoStreaming (VS) interfaces. The
VideoControl (VC) interface is used to access the device controls of the function whereas the
VideoStreaming (VS) interfaces are used to transport data streams into and out of the function.
The collection of the single VideoControl interface and the VideoStreaming interfaces that
belong to the same video function is called the Video Interface Collection (VIC). An Interface
Association Descriptor (IAD) is used to describe the Video Interface Collection.

2.1 Video Interface Class

The Video Interface class groups all functions that can interact with USB-compliant video data
streams. All functions that convert between analog and digital video domains can be part of this
class. In addition, those functions that transform USB-compliant video data streams into other
USB-compliant video data streams can be part of this class. Even analog video functions that are
controlled through USB belong to this class.

In fact, for a video function to be part of this class, the only requirement is that it exposes one
VideoControl Interface. No further interaction with the function is mandatory, although most
functions in the video interface class will support one or more optional VideoStreaming
interfaces for consuming or producing one or more video data streams.

The Video Interface class code is assigned by the USB. For details, see section A.1 "Video
Interface Class Code".

2.2 Video Interface Subclass and Protocol

The Video Interface class is divided into subclasses as identified by the Interface Subclass code.
The following two interface subclasses are defined in this specification.
 VideoControl Interface
 VideoStreaming Interface

The following Function Subclass is used in the Interface Association Descriptor.

 Video Interface Collection

The assigned codes can be found in sections A.2, "Video Interface Subclass Codes" and A.3,
"Video Interface Protocol Codes" of this specification. All other subclass codes are unused and
reserved except code 0xFF, which is reserved for vendor-specific extensions.

Revision 1.5 August 9, 2012 5

 USB Device Class Definition for Video Devices

2.3 Video Function Topology

To be able to manipulate the physical properties of a video function, its functionality must be
divided into addressable entities. The following two generic entities are identified:
 Units
 Terminals

Units provide the basic building blocks to fully describe most video functions. Video functions
are built by connecting together several of these Units. A Unit has one or more Input Pins and a
single Output Pin, where each Pin represents a cluster of logical data streams inside the video
function.

Unit Unit

Figure 2-1: One or more Input Pins, single Output Pin

Units are wired together by connecting their I/O Pins according to the required topology. A
single Output Pin can be connected to one or more Input Pins (fan-out allowed). However, a
single Input Pin can only be connected to one Output Pin (fan-in disallowed). Loops or cycles
within the graph topology are not allowed.

Unit Unit

Unit Unit

Unit Unit

Figure 2-2: Fan-out Allowed / Fan-in Disallowed

In addition, the concept of Terminal is introduced. There are two types of Terminals. An Input
Terminal (IT) is an entity that represents a starting point for data streams inside the video
function. An Output Terminal (OT) represents an ending point for data streams. From the video
function’s perspective, a USB endpoint is a typical example of an Input Terminal or Output
Terminal. It either provides data streams to the video function (IT) or consumes data streams
coming from the video function (OT). Likewise, a Charge Coupled Device (CCD) sensor, built
into the video function is represented as an Input Terminal in the video function’s model.
Connection to a Terminal is made through its single Input Pin or Output Pin.

Revision 1.5 August 9, 2012 6

 USB Device Class Definition for Video Devices

Input Pins of a Unit are numbered starting from one up to the total number of Input Pins on the
Unit. The Output Pin number is always one. Terminals have one Input or Output Pin that is
always numbered one.

The information traveling over I/O Pins is not necessarily of a digital nature. It is possible to use
the Unit model to describe fully analog or even hybrid video functions. The mere fact that I/O
Pins are connected together is a guarantee (by construction) that the protocol and format, used
over these connections (analog or digital), is compatible on both ends.

Every Unit in the video function is fully described by its associated Unit Descriptor (UD). The
Unit Descriptor contains all necessary fields to identify and describe the Unit. Likewise, there is
a Terminal Descriptor (TD) for every Terminal in the video function. In addition, these
descriptors provide all necessary information about the topology of the video function. They
fully describe how Terminals and Units are interconnected.

The descriptors are further detailed in section 3, "Descriptors" of this document.

This specification describes the following types of standard Units and Terminals that are
considered adequate to represent most video functions available today and in the near future:

 Input Terminal
 Output Terminal
 Selector Unit
 Processing Unit
 Encoding Unit
 Extension Unit

Also, there are certain special Terminals that extend the functionality of the basic Input and
Output Terminals. These special Terminals support additional Terminal Descriptor fields and
Requests that are specific to the extended features these Terminals provide. These include:

 Media Transport Terminal (defined in USB Device Class Definition for Video Media
Transport Terminal specification)
 Camera Terminal

The types of Units defined in this specification could be extended in future revisions, or via
companion specifications. For example, a Tuner Unit could be added as a companion
specification to accommodate devices with TV Tuners.

Inside a Unit or Terminal, functionality is further described through Video Controls. A Control
typically provides access to a specific video property. Each Control has a set of attributes that
can be manipulated or that present additional information about the behavior of the Control.
Controls have attributes, which might include:

Revision 1.5 August 9, 2012 7

 USB Device Class Definition for Video Devices

 Current setting
 Minimum setting
 Maximum setting
 Resolution
 Size
 Default

Consider a Brightness Control inside a Processing Unit. By issuing the appropriate requests, the
Host software can obtain values for the Brightness Control’s attributes and, for instance, use
them to correctly display the Control in a User Interface. Setting the Brightness Control’s current
setting attribute allows the Host software to change the brightness of the video that is being
streamed.

The ensemble of Unit Descriptors, Terminal Descriptors and Video Controls provide a full
description of the video function to the Host. A generic class driver shall be able to fully control
the video function. When functionality is represented by Extension Units, the class driver shall
permit access to vendor-specific extensions via a pass-through mechanism. The implementation
details of such a class driver are beyond the scope of this specification.

2.3.1 Input Terminal

The Input Terminal (IT) is used as an interface between the video function’s "outside world" and
other Units inside the video function. It serves as a receptacle for data flowing into the video
function. Its function is to represent a source of incoming data after this data has been extracted
from the data source. The data may include audio and metadata associated with a video stream.
These physical streams are grouped into a cluster of logical streams, leaving the Input Terminal
through a single Output Pin.

An Input Terminal can represent inputs to the video function other than USB OUT endpoints. A
CCD sensor on a video camera or a composite video input is an example of such a non-USB
input. However, if the video stream is entering the video function by means of a USB OUT
endpoint, there is a one-to-one relationship between that endpoint and its associated Input
Terminal. The class-specific Output Header descriptor contains a field that holds a direct
reference to this Input Terminal (see section 3.9.2.2, “Output Header Descriptor”). The Host
needs to use both the endpoint descriptors and the Input Terminal descriptor to get a full
understanding of the characteristics and capabilities of the Input Terminal. Stream-related
parameters are stored in the endpoint descriptors. Control-related parameters are stored in the
Terminal descriptor.

The symbol for the Input Terminal is depicted in the following figure.

Revision 1.5 August 9, 2012 8

 USB Device Class Definition for Video Devices

Figure 2-3 Input Terminal Icon

2.3.2 Output Terminal

The Output Terminal (OT) is used as an interface between Units inside the video function and
the "outside world". It serves as an outlet for video information, flowing out of the video
function. Its function is to represent a sink of outgoing data. The video data stream enters the
Output Terminal through a single Input Pin.

An Output Terminal can represent outputs from the video function other than USB IN endpoints.
A Liquid Crystal Display (LCD) screen built into a video device or a composite video out
connector are examples of such an output. However, if the video stream is leaving the video
function by means of a USB IN endpoint, there is a one-to-one relationship between that
endpoint and its associated Output Terminal. The class-specific Input Header descriptor contains
a field that holds a direct reference to this Output Terminal (see section 3.9.2.1, “Input Header
Descriptor”). The Host needs to use both the endpoint descriptors and the Output Terminal
descriptor to fully understand the characteristics and capabilities of the Output Terminal. Stream-
related parameters are stored in the endpoint descriptors. Control-related parameters are stored in
the Terminal descriptor.

The symbol for the Output Terminal is depicted in the following figure.

Figure 2-4 Output Terminal Icon

2.3.3 Camera Terminal

The Camera Terminal (CT) controls mechanical (or equivalent digital) features of the device
component that transmits the video stream. As such, it is only applicable to video capture devices
with controllable lens or sensor characteristics. A Camera Terminal is always represented as an
Input Terminal with a single output pin. It provides support for the following features.

 Scanning Mode (Progressive or Interlaced)

 Auto-Exposure Mode
 Auto-Exposure Priority
 Exposure Time
 Focus
 Auto-Focus
 Simple Focus
 Iris
 Zoom

Revision 1.5 August 9, 2012 9

 USB Device Class Definition for Video Devices

 Pan
 Roll
 Tilt
 Digital Windowing
 Region of Interest

Support for any particular control is optional. The Focus control can optionally provide support
for an auto setting (with an on/off state). If the auto setting is supported and set to the on state,
the device will provide automatic focus adjustment, and read requests will reflect the
automatically set value. Attempts to programmatically set the Focus control when in auto mode
shall result in protocol STALL with an error code of bRequestErrorCode = “Wrong State”.
When leaving Auto-Focus mode (entering manual focus mode), the control shall remain at the
value that was in effect just before the transition.

2.3.4 Selector Unit

The Selector Unit (SU) selects from n input data streams and routes them unaltered to the single
output stream. It represents a source selector, capable of selecting among a number of sources. It
has an Input Pin for each source stream and a single Output Pin.

The symbol for the Selector Unit is depicted in the following figure.

Figure 2-5 Selector Unit Icon (2 input pins)

2.3.5 Processing Unit

The Processing Unit (PU) controls image attributes of the video being streamed through it. It has
a single input and output pin. It provides support for the following features:

User Controls
 Brightness
 Hue
 Saturation
 Sharpness
 Gamma
 Digital Multiplier (Zoom)

Auto Controls
 White Balance Temperature
 White Balance Component
 Backlight Compensation

Revision 1.5 August 9, 2012 10

 USB Device Class Definition for Video Devices

 Contrast

Other
 Gain
 Power Line Frequency
 Analog Video Standard
 Analog Video Lock Status

Support for any particular control is optional. In particular, if the device supports the White
Balance function, it shall implement either the White Balance Temperature control or the White
Balance Component control, but not both. The User Controls indicate properties that are
governed by user preference and not subject to any automatic adjustment by the device. The
Auto Controls will provide support for an auto setting (with an on/off state). If the auto setting
for a particular control is supported and set to the on state, the device will provide automatic
adjustment of the control, and read requests to the related control will reflect the automatically
set value. Attempts to programmatically set the Focus control when in auto mode shall result in
protocol STALL with an error code of bRequestErrorCode = “Wrong State”. When leaving an
auto mode, the related control shall remain at the value that was in effect just before the
transition.

The symbol for the Processing Unit is depicted in the following figure.

Figure 2-6 Processing Unit Icon

2.3.6 Encoding Unit

The Encoding Unit controls attributes of the encoder that encodes the video being streamed
through it. It has a single input and multiple output pins. It provides support for the following
features which can be used before or after streaming has started.
 Select Layer
 Video Resolution
 Profile and Toolset
 Minimum Frame Interval
 Slice Mode
 Rate Control Mode
Average Bitrate Control
CPB Size Control
 Peak Bit Rate
 Quantization Parameter
 Synchronization and Long Term Reference Frame

Revision 1.5 August 9, 2012 11

 USB Device Class Definition for Video Devices

 Long Term Reference Buffers

 Long Term Picture
 Valid Long Term Pictures
 LevelIDC
 SEI Message
 QP Range
 Priority ID
 Start or Stop Layer
 Error Resiliency

Support for the Encoding Unit control is optional and only applicable to devices with onboard
video encoders. The Select Layer control also allows control of individual streams for devices
that support simulcast transport of more than one stream. Individual payloads may specialize the
behavior of each of these controls to align with the feature set defined by the associated encoder,
e.g. H.264. This specialized behavior is defined in the associated payload specification.

The symbol for the Encoding Unit is depicted in the following figure.

EU
Figure 2-7 Encoding Unit Icon

2.3.7 Extension Unit

The Extension Unit (XU) is the method provided by this specification to add vendor-specific
building blocks to the specification. The Extension Unit can have one or more Input Pins and has
a single Output Pin.

Although a generic host driver will not be able to determine what functionality is implemented in
the Extension Unit, it shall report the presence of these extensions to vendor-supplied client
software, and provide a method for sending control requests from the client software to the Unit,
and receiving status from the unit.

The symbol for the Extension Unit is depicted in the following figure.

Figure 2-8 Extension Unit Icon

2.4 Operational Model

A device can support multiple configurations. Within each configuration can be multiple
interfaces, each possibly having alternate settings. These interfaces can pertain to different

Revision 1.5 August 9, 2012 12

 USB Device Class Definition for Video Devices

functions that co-reside in the same composite device. Several independent video functions can
exist in the same device. Interfaces that belong to the same video function are grouped into a
Video Interface Collection described by an Interface Association Descriptor. If the device
contains multiple independent video functions, there must be multiple Video Interface
Collections (and hence multiple Interface Association Descriptors), each providing full access to
their associated video function.

As an example of a composite device, consider a desktop camera equipped with a built in

microphone. Such a device could be configured to have one interface collection dealing with
configuration and control of the audio function, while another interface collection deals with its
video aspects. One of those, the VideoControl interface, is used to control the inner workings of
the function, whereas the other, the VideoStreaming interface, handles the data traffic received
from the camera video subsystem.

Video Interface Collections can be dynamic in devices that support multiple operating modes.
Because the VideoControl interface, together with its associated VideoStreaming interface(s),
constitutes the ‘logical interface’ to the video function, they must all come into existence at the
same moment in time. Changing the operating mode of a device causes the previous Video
Interface Collection to be replaced with a new Video Interface Collection, followed by re-
initialization of the host software. This specification does not provide a mechanism for the host
to initiate such a mode change, which is typically initiated via a physical switch on the device.

As stated earlier, video functionality is located at the interface level in the device class hierarchy.
The following sections describe the Video Interface Collection, containing a single VideoControl
interface and optional VideoStreaming interfaces, together with their associated endpoints that
are used for video function control and for data stream transfer.

2.4.1 Video Interface Collection

A device must use an Interface Association Descriptor to describe a Video Interface Collection
for each device function that requires a VideoControl Interface and one or more VideoStreaming
interfaces. The Interface Association Descriptor must always be returned as part of the device’s
complete configuration descriptor in response to a GetDescriptor (Configuration) request. The
Interface Association Descriptor must be located before the VideoControl Interface and its
associated VideoStreaming Interfaces (including all alternate settings). All of the interface
numbers in the set of associated interfaces must be contiguous.

2.4.2 VideoControl Interface

To control the functional behavior of a particular video function, the Host can manipulate the
Units and Terminals inside the video function. To make these objects accessible, the video
function must expose a single VideoControl interface. This interface can contain the following
endpoints.

Revision 1.5 August 9, 2012 13

 USB Device Class Definition for Video Devices

 A control endpoint for manipulating Unit and Terminal settings and retrieving the state of
the video function. This endpoint is mandatory, and the default endpoint 0 is used for this
purpose.
 An interrupt endpoint for status returns. This endpoint is optional, but may be mandatory
under certain conditions. See section 2.4.2.2, "Status Interrupt Endpoint" for further
information.

The VideoControl interface is the single entry point to access the internals of the video function.
All requests that are concerned with the manipulation of certain Video Controls within the video
function’s Units or Terminals must be directed to the VideoControl interface of the video
function. Likewise, all descriptors related to the internals of the video function are part of the
class-specific VideoControl interface descriptor.

This specification defines a single alternate setting for the VideoControl interface, the default
alternate setting zero.

2.4.2.1 Control Endpoint

The video interface class uses endpoint 0 (the default pipe) as the standard way to control the
video function using class-specific requests. These requests are always directed to one of the
Units or Terminals that make up the video function. The format and contents of these requests
are detailed further in this document.

2.4.2.2 Status Interrupt Endpoint

A USB VideoControl interface can support an optional interrupt endpoint to inform the Host
about the status of the different addressable entities (Terminals, Units, interfaces and endpoints)
inside the video function. The interrupt endpoint, if present, is used by the entire Video Interface
Collection to convey status information to the Host. It is considered part of the VideoControl
interface because this is the anchor interface for the Collection.

This interrupt endpoint is mandatory if:

 The device supports hardware triggers for still image capture (see section 2.4.2.3,
"Hardware Trigger Interrupts").
 The device implements any AutoUpdate controls (controls supporting device initiated
changes).
 The device implements any Asynchronous controls (see section 2.4.4, "Control Transfer
and Request Processing").

The interrupt packet is a variable size data structure depending on the originator of the interrupt
status. The bStatusType and bOriginator fields contain information about the originator of the
interrupt. The bEvent field contains information about the event triggering the interrupt. If the
originator is the Video Control interface, the bSelector field reports the Control Selector of the
control that issued the interrupt. Any addressable entity inside a video function can be the
originator.

Revision 1.5 August 9, 2012 14

 USB Device Class Definition for Video Devices

The contents of the bOriginator field must be interpreted according to the code in D3..0 of the
bStatusType field. If the originator is the VideoControl interface, the bOriginator field contains
the Terminal ID or Unit ID of the entity that caused the interrupt to occur. If the bOriginator
field is set to zero, the virtual entity interface is the originator. This can be used to report global
VideoControl interface changes to the Host. If the originator is a VideoStreaming interface, the
bOriginator field contains the interface number of the VideoStreaming interface. This scheme is
unambiguous because Units and Terminals are not allowed to have an ID of zero.

If the originator is the VideoControl interface, the bAttribute field indicates the type of Control
change.

The contents of the bEvent field must also be interpreted according to the code in D3..0 of the
bStatusType field. If the originator is the VideoStreaming interface, there are additional button
press events defined as described in the table below.

For all originators, there is a Control Change event defined. Controls that support this event will
trigger an interrupt when a host-initiated or externally-initiated control change occurs. The
interrupt shall only be sent when the operation corresponding to the control change is completed
by the device.

A Control shall support Control Change events if any of the following is true:
 The Control state can be changed independently of host control.
 The Control can take longer than 10ms from the start of the Data stage through the
completion of the Status stage when transferring to the device (SET_CUR operations).

If a control is required to support Control Change events, the event shall be sent for all
SET_CUR operations, even if the operation can be completed within the 10ms limit. The device
indicates support for the Control Change event for any particular control via the GET_INFO
attribute (see section 4.1.2, "Get Request"). Section 2.4.4, "Control Transfer and Request
Processing" describes in detail the interaction of Control Transfers (Requests) and Control
Change events.

The following tables specify the format of the status packet.

Table 2-1 Status Packet Format
Offset Field Size Value Description
0 bStatusType 1 Bitmap/Number D7..4: Reserved
D3..0: Originator
0 = Reserved
1 = VideoControl interface
2 = VideoStreaming interface
1 bOriginator 1 Number ID of the Terminal, Unit or Interface that
reports the interrupt

Revision 1.5 August 9, 2012 15

 USB Device Class Definition for Video Devices

When the originator is a Video Control Interface, the rest of structure is:
Table 2-2 Status Packet Format (VideoControl Interface as the Originator)
Offset Field Size Value Description
2 bEvent 1 Number 0x00: Control Change
0x01 – 0xFF: Reserved
3 bSelector 1 Number Control Change
Report the Control Selector of the control that
issued the interrupt.
4 bAttribute 1 Number Specify the type of control change:
0x00: Control value change
0x01: Control info change
0x02: Control failure change
0x03: Control min change
0x04: Control max change
0x05 – 0xFF: Reserved
5 bValue n See control request description in section 4.2
"VideoControl Requests".

bAttribute: Description:
0x00 Equivalent to the result of a
GET_CUR request
0x01 Equivalent to the result of a
GET_INFO request
0x02 Equivalent to the result of a
GET_CUR request on
VC_REQUEST_ERROR_
CODE_CONTROL
0x03 Equivalent to the result of a
GET_MIN request
0x04 Equivalent to the result of a
GET_MAX request

When the originator is a Video Streaming Interface the rest of the structure is:
Table 2-3 Status Packet Format (VideoStreaming Interface as the Originator)
Offset Field Size Value Description
2 bEvent 1 Number All originators:
0x00 = Button Press
0x01 – 0xFF = Stream Error
3 bValue n Number Button Press: (n=1)
0x00: Button released
0x01: Button pressed

Revision 1.5 August 9, 2012 16

 USB Device Class Definition for Video Devices

2.4.2.3 Hardware Trigger Interrupts

One of the defined usages of the Status Interrupt Endpoint is for hardware triggers to notify host
software to initiate still image capture. When the hardware detects a button press, for example,
the Status Interrupt Endpoint will issue an interrupt originating from the relevant
VideoStreaming interface. The event triggering the interrupt (button press or release) is indicated
in the interrupt packet. The default, initial state of the button is the "release" state.

The device will have to specify whether it supports hardware triggers, and how the Host software
should respond to hardware trigger events. These are specified in the class-specific descriptors
within the relevant VideoStreaming interface. See section 3, "Descriptors".

2.4.2.4 Still Image Capture

A common feature of video cameras is the support of still image capture associated with a video
stream. This can be initiated either by programmatic software triggers or hardware triggers.

Depending on the method used, the still image frame may have to be the same size as the video
frames that are being streamed. There are several supported methods of capturing the still image,
and the device will have to specify which method it supports in the class-specific descriptors
within the relevant VideoStreaming interface.

Method 1 - The host software will extract the next available video frame from the active video
pipe in the relevant VideoStreaming interface upon receiving the hardware trigger event. The
hardware does not interrupt or alter the video stream in this case. For this method, the still image
frame is always the same size as the video frames being streamed.

Method 2 – If the device supports higher-quality still images, it has the option of streaming still-
image-specific packets across the active video pipe. In this case, the host software will
temporarily suspend video streaming, select the optimal bandwidth alternate setting based on the
still probe/commit negotiation (subject to bandwidth availability), send a
VS_STILL_IMAGE_TRIGGER_CONTROL Set request with the "Transmit still image" option
(see section 4.3.1.4, "Still Image Trigger Control"), and prepare to receive the still image data.
The device transmits the still image data marked as such in the payload header (see section
2.4.3.2.2, "Sample Isochronous Transfers"). Once the complete still image is received, the host
software will then revert back to the original alternate setting, and resume video streaming.

Method 3 – This method enables the capture of higher-quality still images from a dedicated bulk
still image pipe. By doing so, the active streams would continue uninterrupted. There are two
cases covered by this method.

In the first case, the host software initiates the still image capture from the device. It does so by
issuing a VS_STILL_IMAGE_TRIGGER_CONTROL Set request with the "Transmit still image
via dedicated bulk pipe" option (see section 4.3.1.4, "Still Image Trigger Control"). In this case,

Revision 1.5 August 9, 2012 17

 USB Device Class Definition for Video Devices

after issuing the request, the host will start receiving the still image from the bulk still image
endpoint of the relevant VideoStreaming interface. The device captures the high-quality still
image and transmits the data to the bulk still image endpoint. While transmission is occurring,
the bTrigger field of the VS_STILL_IMAGE_TRIGGER_CONTROL control shall remain as
"Transmit still image via dedicated bulk pipe". After transmission is complete, the device shall
reset the control to "Normal operation" and trigger a control change interrupt via the Status
Interrupt endpoint.

In the second case, the device initiates the still image transmission after detecting a hardware
trigger. When the hardware detects a button press, the Status Interrupt endpoint will issue an
interrupt originating from the relevant VideoStreaming interface. If the bTriggerUsage field of
the selected Format descriptor is set as initiating still image capture, the device shall set the
bTrigger field of the VS_STILL_IMAGE_TRIGGER_CONTROL control to “Transmit still
image via dedicated bulk pipe”. The Host software should then begin receiving still image data
that was captured by the device after it received the interrupt. After transmission is complete, the
device shall reset the bTrigger field to “Normal operation”. The host software can abort data
transmission by issuing a VS_STILL_IMAGE_TRIGGER_CONTROL request with the “Abort
still image transmission” option. In either case, the device shall trigger a control change interrupt
via the Status Interrupt endpoint

The following table summarizes endpoint usage for the various methods of still image capture.
Table 2-4 Summary of Still Image Capture Methods
Isochronous video data pipe Bulk video data pipe
Method 1 1 Isochronous (Video) 1 Bulk (Video)
Method 2 1 Isochronous (Video/Still) 1 Bulk (Video/Still)
Method 3 1 Isochronous (Video) 1 Bulk (Video)
1 Bulk (Still) 1 Bulk (Still)

2.4.2.5 Optical Zoom vs Digital Zoom

Users expect to use a single control to traverse the entire range of optical and digital zoom.
Further, users expect that digital zoom will not be applied until full optical zoom has been
realized. It is recommended that the device enforce this behavior. A solution with details on how
this can be achieved is presented in Appendix D.

2.4.3 VideoStreaming Interface

VideoStreaming interfaces are used to interchange digital data streams between the Host and the
video function. They are optional. A video function can have zero or more VideoStreaming
interfaces associated with it, each possibly carrying data of a different nature and format. Each
VideoStreaming interface can have one isochronous or bulk data endpoint for video, and an
optional dedicated bulk endpoint for still images related to the video (only for method 3 of still
image transfer. See section 2.4.2.4 "Still Image Capture"). This construction guarantees a one-to-

Revision 1.5 August 9, 2012 18

 USB Device Class Definition for Video Devices

one relationship between the VideoStreaming interface and the single data stream related to the
endpoint.

A VideoStreaming interface with isochronous endpoints must have alternate settings that can be
used to change certain characteristics of the interface and underlying endpoint(s). A typical use
of alternate settings is to provide a way to change the bandwidth requirements an active
isochronous pipe imposes on the USB. All devices that transfer isochronous video data must
incorporate a zero-bandwidth alternate setting for each VideoStreaming interface that has an
isochronous video endpoint, and it must be the default alternate setting (alternate setting zero). A
device offers to the Host software the option to temporarily relinquish USB bandwidth by
switching to this alternate setting. The zero-bandwidth alternate setting does not contain a
VideoStreaming isochronous data endpoint descriptor.

A VideoStreaming interface containing a bulk endpoint for streaming shall support only alternate
setting zero. Additional alternate settings containing bulk endpoints are not permitted in a device
that is compliant with the Video Class specification. This restriction does not prohibit the mix of
bulk and isochronous endpoints when the bulk endpoints are used solely for Still Image Transfer
Method 3. In that case, each alternate setting will include the descriptors for both an isochronous
endpoint and a bulk endpoint.

If a VideoStreaming interface with an isochronous endpoint supports a set of video parameter

combinations (including video format, frame size and frame rate) that utilize significantly
varying amounts of bandwidth across all combinations, it is recommended that the
VideoStreaming interface support a range (greater than two) of alternate interface settings with
varying maximum packet sizes. By doing so, the host would be able to select an appropriate
alternate setting for the given video parameter combination that makes most efficient use of bus
bandwidth.

For device implementers, the process of determining the number of alternate settings to be
provided and the maximum packet size for the video data endpoint in each alternate setting is
implementation dependent, and would depend on the bandwidth usage across the range of video
parameter combinations that the VideoStreaming interface is capable of supporting.

2.4.3.1 Stream Bandwidth Selection

The bandwidth required by a video stream can be satisfied by a USB bandwidth that is equal to
or greater than the function stream bandwidth. This can be illustrated as follows.

Revision 1.5 August 9, 2012 19

 USB Device Class Definition for Video Devices

USB Bandwidth
Function Bandwidth
Stream

Figure 2-9 Stream Bandwidth Selection

The optimal allocation of the USB bandwidth to match the function’s bandwidth requirement is
achieved via negotiation between the host and the device.

See section 4.3.1.1, "Video Probe and Commit Control" for a complete description of the
negotiation process.

The negotiation process allows the host to provide preferred stream parameters to the device,
while the device selects the best combination of streaming parameters and reports the maximum
bandwidth usage for those settings. The host will use the bandwidth information to identify the
optimal alternate interface. The device is responsible for choosing the live streaming parameters
once the bandwidth is allocated. These parameters may be different than originally agreed upon
during the negotiation process. However, during the negotiation process, the host provided hints
to the device indicating the preferred way to choose the live stream parameters.

Once bandwidth has been allocated and streaming started, further parameter negotiation between
the host and the device can be performed without disturbing the current stream. Streaming
parameters are set as a group so that the function will have all information available while it
attempts to determine a working set.

Still image Method 2 uses a similar mechanism (see section 2.4.2.4, “Still Image Capture”).

2.4.3.2 Video and Still Image Samples

A video (or still image) sample refers to an encoded block of video data that the format-specific
decoder is able to accept and interpret in a single transmission. A single video sample may or
may not correspond to a single decoded video frame, depending on the video format in use. For
example, a YUV video stream (which has no inter-frame compression) would have a one to one
correspondence between a video sample and video frame. However, a MPEG-2 TS data stream
will require many video samples (or TS packets) to form a decoded video frame.

A single video sample may require multiple class-defined Payload Transfers. Conversely, there
may be one or more video samples within a single Payload Transfer. In the latter case, there must
be an integral number of fixed size samples within each Payload Transfer.

Revision 1.5 August 9, 2012 20

 USB Device Class Definition for Video Devices

The VideoStreaming endpoint(s) encapsulate data with the class-defined Payload Header. This
encapsulation is identical for Payload Transfers on both isochronous and bulk endpoint types,
and applies to both the streaming and still image endpoints.

The following block diagram details the protocol layering and abstraction used in Payload
Transfers.

Video Codec

(3) Sample Transfers

Video Sample Handler

(2) Payload Transfers

Isochronous Bulk Transfers
Transfers

(1) USB Transfers

USBD

Figure 2-10 Protocol Layering and Abstraction

1. I/O Request Packet (IRP) requests from the client to the USB system software result in USB
transfers.
2. In response to IRP completion, the host software forwards the data in the form of payload
transfers. The bulk and isochronous handlers hide the transfer type differences from the
upper layers of the protocol stack.
3. The video sample handler accumulates the individual payload transfers to form a sample
transfer.

A Payload Transfer is composed of the class-defined payload header (see section 2.4.3.3 "Video
and Still Image Payload Headers") followed by the format-specific payload data.

Payload header Payload data

Revision 1.5 August 9, 2012 21

 USB Device Class Definition for Video Devices

Figure 2-11 A Payload Transfer

Revision 1.5 August 9, 2012 22

 USB Device Class Definition for Video Devices

2.4.3.2.1 Sample Bulk Transfers

The following examples show the relationship between Video Samples, Payload Transfers and
the token and data packets when exchanging bulk transfers with a device. Handshake packets are
not shown for the sake of clarity.

IN DATA0 IN DATA1 … IN DATA0/1

Payload
Payload
Header
Data

IN DATA0 IN DATA1 … IN DATA0/1 Video

Sample

Payload
Payload
Header
Data
.
.
.

IN DATA0 IN DATA1

Payload
Payload
Header
Data

Figure 2-12 Sample Bulk Read (Multiple Transfers per Sample)

Revision 1.5 August 9, 2012 23

 USB Device Class Definition for Video Devices

IN DATA0 IN DATA1 … IN DATA0/1

Payload
Payload
Header
Data

Video
Sample

Figure 2-13 Sample Bulk Read (Single Transfer per Sample)

OUT DATA0 OUT DATA1 … OUT DATA0/1

Payload
Payload
Header
Data

Video
Sample

Figure 2-14 Sample Bulk Write (Single Transfer per Sample)

Revision 1.5 August 9, 2012 24

 USB Device Class Definition for Video Devices

2.4.3.2.2 Sample Isochronous Transfers

The following examples show the relationship between Video Samples, Payload Transfers and
the token and data packets when exchanging isochronous transfers with a device. The actual
video sample size and bandwidth usage (i.e. number of data transactions and amount of data in
the last transaction of each payload) will vary according to the requirements of the device and
payload.

Figure 2-15 gives an example of a High Speed/High Bandwidth transfer over an IN endpoint.

SOF Packet

IN DATA2 IN DATA1 IN DATA0

Payload
Payload
Header
Data

SOF Packet

IN DATA2 IN DATA1 IN DATA0 Video

Sample

Payload
Payload
Header
Data
.
.
SOF Packet

IN DATA2 IN DATA1 IN DATA0

Payload Payload
Header Data

Figure 2-15 Sample Isochronous Transfer, IN endpoint

Revision 1.5 August 9, 2012 25

 USB Device Class Definition for Video Devices

Figure 2-16 gives an example of a High Speed/High Bandwidth transfer over an OUT endpoint.

SOF Packet

OUT MDATA OUT MDATA OUT DATA2

Payload Payload
Header Data

SOF Packet

OUT MDATA OUT MDATA OUT DATA2 Video

Sample

Payload Payload
Header Data
.
.
.
SOF Packet

OUT MDATA OUT DATA1

Payload Payload
Header Data

Figure 2-16 Sample Isochronous Transfer, OUT endpoint

Revision 1.5 August 9, 2012 26

 USB Device Class Definition for Video Devices

Figure 2-17 gives an example of a Full or High Speed transfer over an IN endpoint.

SOF Packet

IN DATA0

Payload
Header Payload
Data

SOF Packet

IN DATA0
Video
Sample
Payload
Header Payload
Data
.
.
.
SOF Packet

IN DATA0

Payload Payload
Header Data

Figure 2-17 Sample Isochronous Transfer, IN endpoint

Revision 1.5 August 9, 2012 27

 USB Device Class Definition for Video Devices

Figure 2-18 gives an example of a Full or High Speed transfer over an OUT endpoint.

SOF Packet

OUT DATA0

Payload
Header Payload
Data

SOF Packet

OUT DATA0 Video

Sample

Payload
Header Payload
. Data

.
.
SOF Packet

OUT DATA0

Payload Payload
Header Data

Figure 2-18 Sample Isochronous Transfer, OUT endpoint

Revision 1.5 August 9, 2012 28

 USB Device Class Definition for Video Devices

2.4.3.3 Video and Still Image Payload Headers

Every Payload Transfer containing video or still-image sample data must start with a Payload
Header.

The format of the payload header is defined as follows.

Table 2-5 Format of the Payload Header
Offset Field Size Value Description
0 bHeaderLength 1 Number Length of the payload header in bytes
including this field.
1 bmHeaderInfo 1 Bitmap Provides information on the sample data
following the header, as well as the
availability of optional header fields in this
header.
D0: Frame ID – For frame-based formats,
this bit toggles between 0 and 1 every time a
new video frame begins. For stream-based
formats, this bit toggles between 0 and 1 at
the start of each new codec-specific segment.
This behavior is required for frame-based
payload formats (e.g., DV) and is optional
for stream-based payload formats (e.g.,
MPEG-2 TS). For stream-based formats,
support for this bit must be indicated via the
bmFramingInfo field of the Video Probe
and Commit controls (see section 4.3.1.1,
“Video Probe and Commit Controls”).

D1: End of Frame – This bit is set if the

following payload data marks the end of the
current video or still image frame (for frame-
based formats), or to indicate the end of a
codec-specific segment (for stream-based
formats). This behavior is optional for all
payload formats. For stream-based formats,
support for this bit must be indicated via the
bmFramingInfo field of the Video Probe
and Commit Controls (see section 4.3.1.1,
“Video Probe and Commit Controls”).

D2: Presentation Time – This bit is set if the

dwPresentationTime field is being sent as
part of the header.

Revision 1.5 August 9, 2012 29

 USB Device Class Definition for Video Devices

D3: Source Clock Reference – This bit is set

if the dwSourceClock field is being sent as
part of the header.

D4: Payload specific bit. See individual

payload specifications for use.

D5: Still Image – This bit is set if the

following data is part of a still image frame,
and is only used for methods 2 and 3 of still
image capture. For temporally encoded
formats, this bit indicates the following data
is part of an intra-coded frame.

D6: Error – This bit is set if there was an

error in the video or still image transmission
for this payload. The Stream Error Code
control would reflect the cause of the error.

D7: End of header – This bit is set if this is

the last header group in the packet, where the
header group refers to this field and any
optional fields identified by the bits in this
field (Defined for future extension).

The following fields may or may not be included in the header, depending on the bits that were
specified in the bmHeaderInfo field above.

These fields are in the order in which they are specified in the bitmap header field above, in the
order of least significant bit first. Because the header itself might be extended in the future, the
offset of dwPresentationTime is also variable. The device will indicate if it supports these fields
in the Payload Format Descriptor within the class-specific VideoStreaming descriptor. See
section 3.9.2.3 "Payload Format Descriptors".
Table 2-6 Extended Fields of the Payload Header
Offset Field Size Value Description
Variable dwPresentationTime 4 Number Presentation Time Stamp (PTS).
The source clock time in native device
clock units when the raw frame capture
begins. This field may be repeated for
multiple payload transfers comprising a
single video frame, with the restriction
that the value shall remain the same

Revision 1.5 August 9, 2012 30

 USB Device Class Definition for Video Devices

throughout that video frame. The PTS is

in the same units as specified in the
dwClockFrequency field of the Video
Probe Control response.
Variable scrSourceClock 6 Number A two-part Source Clock Reference
(SCR) value
D31..D0: Source Time Clock in native
device clock units
D42..D32: 1KHz SOF token counter
D47..D43: Reserved, set to zero.
The least-significant 32 bits (D31..D0)
contain clock values sampled from the
System Time Clock (STC) at the source.
The clock resolution shall be according to
the dwClockFrequency field of the Probe
and Commit response of the device as
defined in Table 4-75 of this specification.
This value shall comply with the
associated stream payload specification.
The times at which the STC is sampled
must be correlated with the USB Bus
Clock. To that end, the next most-
significant 11 bits of the SCR (D42..D32)
contain a 1 KHz SOF counter,
representing the frame number at the time
the STC was sampled.
 STC must be captured when the
first video data of a video frame is
put on the USB bus.
 SCR must remain constant for all
payload transfers within a single
video frame.
The most-significant 5 bits (D47..D43) are
reserved, and must be set to zero.
The maximum interval between Payload
Headers containing SCR values is 100ms,
or the video frame interval, whichever is
greater. Shorter intervals are permitted.

Revision 1.5 August 9, 2012 31

 USB Device Class Definition for Video Devices

The periodic transmission of the dwPresentationTime and dwSourceClock fields is mandatory

if all of the following conditions are true.

 The device has multiple video and/or audio source functions and is sending audio and
video streams to the host.
 The video and/or audio streams are interrelated and therefore need to be kept
synchronized.
 The stream format in use does not already contain timestamp and clock reference
information (MPEG-2 TS is an example of a format that contains this information).
 The sample is part of a video frame (and not a still image frame).

For temporally encoded payloads, the dwPresentationTime and dwSourceClock fields may be
required for all video frames. See the appropriate payload specification for details.

These time information fields allow the host software to reconstruct the source clock to support
high-quality synchronization between separate data pipes (audio, video, etc.) and rate matching
between the data source and sink, as discussed in the following section.

2.4.3.4 Stream Synchronization and Rate Matching

To properly synchronize multiple audio and video streams from a media source, the media
source must provide (to the media sink) its local stream latency, periodic clock reference
information, and a way for the media sink to determine the proper presentation time for samples
from each stream (relative to the other streams).

2.4.3.4.1 Latency
The media source is required to report its internal latency (delay from data acquisition to data
delivery on the bus). This latency reflects the lag introduced by any buffering, compression,
decompression, or processing done by the stream source. Without latency information for each
stream, a media sink (or rendering device) cannot properly correlate the presentation times of
each stream.

In the case of a video source, this means that the source must guarantee that the portion of a
sample fully acquired as of SOFn (Start Of Frame n) will have been completely sent to the bus as
of SOFn+. Latency  is the source’s internal delay expressed in number of USB frames
(milliseconds). For high-speed endpoints, the resolution increases to 125 microseconds, but the
delay will continue to be expressed in number of USB frames. Every VideoStreaming interface
must report this latency value. See the description of the wDelay parameter in section 4.3.1.1,
"Video Probe and Commit Controls". By following these rules, phase jitter is limited to ±1
millisecond. It is up to the video sink to synchronize streams by scheduling the rendering of
samples at the correct moment, taking into account the internal delays of all media streams being
rendered.

Revision 1.5 August 9, 2012 32

 USB Device Class Definition for Video Devices

2.4.3.4.2 Clock Reference

Clock reference information is used by a media sink to perform clock rate matching. Rate
matching refers to the synchronization of the media sink’s rendering clock with the media
source’s sampling clock. Without clock rate matching, a stream will encounter buffer overrun or
underrun errors. This has not been a problem with audio streams due to the relative ease of
performing audio sample rate conversion. However, sample rate conversion is significantly more
difficult with video, so a method for rate matching is required.

To understand the problem of clocks running at slightly different rates, consider the following
example. For simplicity, assume that video buffers can be filled instantaneously, and that there is
one buffer available to be filled at any given time within the video frame interval. Also assume
that the two crystals governing the source and rendering clocks operate with 100ppm (parts per
million) accuracy. The accuracy value is a ratio that can be applied such that for every frame, the
clock will drift by a fraction of the frame that is equal to the ratio. In other words, two clocks
with accuracy of 100ppm could have a worst case drift relative to each other of 1/5,000th of a
frame (two clocks at opposite extremes of their valid operating range for a cumulative error ratio
of 2 * 100/1,000,000). Therefore, a frame glitch will occur once every 5,000 frames. At a frame
rate of 30 fps, this would equate to a glitch every 166.67 seconds. At a frame rate of 60 fps, it’s
worse, with one glitch every 83.3 seconds.

Frame glitches can be postponed, but not avoided, by adding additional buffers to hold video
frames before they are rendered. If the source clock is running slower than the rendering clock,
the buffer underrun could only be postponed by letting the extra buffers fill to a certain threshold
before rendering, resulting in unacceptable latency. Once the first glitch occurs, the extra buffers
are effectively useless, since the behavior will degrade to the single-buffer case from that point
onward.

This specification assumes that in all cases, the media sink has no control over the media source
clock, and that the source and sink do not "slave" to a common clock (the bus clock lacking
sufficient resolution). Also, due to cost constraints, additional isochronous endpoints to
communicate clock rate information will not be used. Therefore, this specification requires that a
video stream include clock reference information that can be used to adjust the rendering clock
rate. The clock reference information may be encapsulated in a transport stream, or it may be
provided via an optional field in each payload header. This field becomes required in the latter
case.

2.4.3.4.3 Presentation Time

For fixed rate streams, the presentation time can be derived from the data stream. For a fixed-rate
audio stream (e.g., PCM), the media sink can derive the presentation time from the stream offset
(typically the count of bytes since start of capture). For variable rate streams, each sample must
be accompanied by a presentation timestamp. The media sink is responsible for converting the
timestamp to native units and adjusting the timestamp to account for the local clock offset when
a stream starts, as well as accounting for source stream latency. Even though video streams
might arrive at the media sink at a fixed frame rate, if they are subject to variable rate

Revision 1.5 August 9, 2012 33

 USB Device Class Definition for Video Devices

compression and encoding, they are not considered fixed-rate streams and will require
timestamps on the samples.

2.4.3.5 Dynamic Frame Interval Support

In order to adjust to different environmental conditions, such as varying lighting conditions, it
may be necessary for a video device (such as a camera) to dynamically change the frame interval
and sensor exposure time to maintain acceptable image quality while streaming.

After bus bandwidth for the video data pipe of the corresponding VideoStreaming interface has
been allocated and streaming has commenced, the data source may dynamically vary the frame
interval (and the corresponding frame rate), as long as the new frame interval does not require
greater bus bandwidth than what was originally allocated. The data sink would determine the
new frame interval based on the Presentation Time Stamp (PTS) information included in the
video payload headers.

2.4.3.6 Device Initiated Dynamic Format Change Support

Certain devices, such as those that contain a tape media transport, are capable of dynamically
changing the video format being streamed to the host while streaming is occurring. Since the
new video format may have different bus bandwidth requirements from the old format, the host
must be notified of the format change and be allowed to perform the reconfiguration and bus
bandwidth reallocation necessary to support the new video format.

The device indicates its support for dynamic format change events through the bmInfo field of
the VideoStreaming Input Header. See section 3.9.2.1 "Input Header Descriptor".

When a dynamic format change event occurs, the following steps take place:
 Device detects dynamic format change (while streaming is occurring).
 Device begins sending empty data payloads to the host with the Error bit set in the video
stream payload header.
 Device sets the Stream Error Code Control to "Format Change" (see section 4.3.1.7
"Stream Error Code Control").
 The host queries the new stream state through a VS_PROBE_CONTROL request with
the GET_CUR attribute (see 4.3.1.1, “Video Probe and Commit Controls”).
 If the new format is acceptable by the host, it issues a VS_COMMIT_CONTROL request
with the SET_CUR attribute and, if necessary, reallocates the USB bandwidth through an
alternate interface selection standard request. If the new format is not acceptable, the host
will negotiate a new format with the stream PROBE/COMMIT controls.

2.4.3.7 Data Format Classes

For the purposes of host processing of incoming and outgoing data packets, the various video
formats supported by the USB Video Class (UVC) can be divided into two broad categories:

Revision 1.5 August 9, 2012 34

 USB Device Class Definition for Video Devices

 Frame-based video formats – These video formats require the frame/sample boundary
information to be transmitted out-of-band. Examples of such formats are uncompressed
video (formatted in various YUV variants), MJPEG, and DV. For these formats, the FID
(and optionally EOF) bits in the UVC payload headers must be supported.
 Stream-based video formats – These video formats have the frame/sample boundary
information transmitted in-band. Examples of such formats are MPEG-2 TS, MPEG-2 PS
and MPEG-1 system streams. For these formats, the FID and EOF bits are optional. If
used, the bits allow the sender to identify codec-specific segment boundaries within the
stream. The receiver would typically use this information to provide data to a decoder
with lower latency than would be possible if buffer fullness alone was used to trigger
buffer completion (see section 4.3.1.1, “Video Probe and Commit Controls”).
 Temporally Encoded video formats – While these video formats have the frame/sample
boundary information transmitted in-band, they are often managed as frames or sub-
frames by the host. The EOF and EOS bits are required to indicate these boundaries to
the host so it may generate time stamps and trigger buffer completion on these
boundaries. Examples of temporally encoded video formats are H.264 and VP8.

The following is determined by the format class under which the video format is classified:
 The default Incoming/Outgoing data processing algorithm
 Bit fields supported by default in UVC payload header (BFH[0])

The following is determined by the specific video payload format:

 Format descriptor type
 Frame descriptor type, if needed
 Support for time information fields in UVC payload header

2.4.4 Control Transfer and Request Processing

The Video Class specification’s control transfer (or Request) mechanism builds upon sections
5.5, “Control Transfers”; 8.5.3 ,“Control Transfers; 9.2.6, “Request Processing”; and 9.3, “USB
Device Requests” of the Universal Serial Bus Specification, Revision 2.0 (the USB 2.0 spec).
Those sections describe the timing and error handling of control transfers, but do not prescribe a
method for control transfer completion using interrupt pipes. The following paragraphs describe
Control Transfer operations in the context of the Video Class, including the use of the Status
Interrupt pipe to provide notification of state changes within the device.

Control transfers minimally have two transaction stages: Setup and Status. A control transfer
may optionally contain a Data stage between the Setup and Status stages. The Setup stage
contains all information necessary to address a particular entity, specify the desired operation,
and prepare for an optional Data stage. A Data stage can be host-to-device (OUT transactions),
or device-to-host (IN transactions), depending on the direction and operation specified in the
Setup stage via the bmRequestType and bRequest fields.

Revision 1.5 August 9, 2012 35

 USB Device Class Definition for Video Devices

In the context of the Video Class specification, SET_CUR requests will always involve a Data
stage from host to device, and GET_* requests will always involve a Data stage from device to
host. Although none are defined currently, an exception to this rule would be a SET_CUR
request where the bRequest field contains all information necessary to place the device into a
known state. However, “toggle” requests without a Data stage are explicitly disallowed.

The device shall use protocol STALL (not function stall) during the Data or Status stages if the
device is unable to complete the Control transfer (see section 8.5.3.4 of the USB Specification
Revision 2.0). Reasons for protocol STALL include unsupported operations, invalid target entity,
invalid control selector, unexpected Data length, or invalid Data content. The device shall update
the value of Request Error Code Control, and the host may use that control to determine the
reason for the protocol STALL (see section 4.2.1.2 "Request Error Code Control"). The device
must not NAK or STALL the SETUP transaction.

Typically, the host will serialize Control Transfers, meaning that the next Setup stage will not
begin until the previous Status stage has completed. However, in situations where the bus has
experienced errors, a Setup transaction may be sent before the completion of a previous control
transfer. The device must abandon the previous control transfer.

Due to this command serialization, it is important that the duration of control transfers (from
Setup stage through Status stage) be kept as short as possible. For this reason, as well as the
desire to avoid polling for device status, this specification defines an interrupt status mechanism
to convey status changes independently of the control transfers that caused the state change. This
mechanism is described in section 2.4.2.2, "Status Interrupt Endpoint". Any control that requires
more than 10ms to respond to a SET_CUR request (asynchronous control), or that can change
independently of any external SET_CUR request (Autoupdate control), must send a Control
Change status interrupt. These characteristics will be reflected in the GET_INFO response for
that control (see 4.1.2, “Get Request”).

If a SET_CUR request is issued to an Asynchronous Control with unsupported operations,

invalid target entity, unexpected data Length or invalid data content, the device shall use
protocol STALL since the device is unable to complete the Control Transfer (see section 8.5.3.4
of the USB Specification Revision 2.0). The device shall update the value of the Request Error
Code Control (see section 4.2.1.2 "Request Error Code Control").

In the case of a SET_CUR request with valid parameters to an Asynchronous Control, the
Control Transfer operation shall enter the Status stage immediately after receiving the data
transferred during the Data stage. Once the Status stage has successfully completed, the device
shall eventually send a Control Change Interrupt that will reflect the outcome of the request:
 If the request succeeded, the Control Change Interrupt will advertise the new value (see
section 2.4.2.2 “Status Interrupt Endpoint”).
 If the request could not be executed, the device shall send a Control Change Interrupt
using the Control Failure Change mechanism to describe the reason for the failure (see

Revision 1.5 August 9, 2012 36

 USB Device Class Definition for Video Devices

Table 2-1 in section 2.4.2.2 “Status Interrupt Endpoint” and Figure 2-23 in section 2.4.4
“Control Transfer and Request Processing”).

The amount of time between the end of a successful Status stage and the Control Change
interrupt is implementation specific. For instance, a tape transport might take 3-5 seconds to
completely change state, so the Control Change interrupt would be sent within 3-5 seconds.

The following flow diagrams show the Setup, Data and Status stages of SET_CUR Control
Transfers for controls supporting one of the two legal bit combinations with the D1 (SET) bit
enabled. These are described because they show the relationship between a SET_CUR request
and the resulting state change.

Revision 1.5 August 9, 2012 37

 USB Device Class Definition for Video Devices

SET/GET Supported

Host Device

Setup Stage

Data Stage

SET_CUR State
<=10ms Change

Status Stage

Figure 2-19 Control Transfer Example (Case 1)

Revision 1.5 August 9, 2012 38

 USB Device Class Definition for Video Devices

SET/GET/Interrupt Supported

Host Device

Setup Stage

Data Stage
SET_CUR
<=10ms
Status Stage

>10ms

State
Change
Control Change interrupt

Figure 2-20 Control Transfer Example (Case 2)

Revision 1.5 August 9, 2012 39

 USB Device Class Definition for Video Devices

SET/GET/Interrupt Supported (with error scenarios)

Host Device
Setup Stage

Data Stage
SET_CUR
<=10ms
Status Stage

Setup Stage
GET_CUR
(same Data Stage STALL
control)

Setup Stage
GET_CUR
(other Data Stage
control)
Status Stage
>10ms

Setup Stage
SET_CUR
(same Data Stage STALL
control)
State
Change
Control Change interrupt

Setup Stage
GET_CUR
(same Data Stage
control)
Status Stage

Figure 2-21 Control Transfer Example (Case 3)

Revision 1.5 August 9, 2012 40

 USB Device Class Definition for Video Devices

SET/GET/Interrupt Supported (Device busy before first SET request)

Host Device

Setup Stage
SET_CUR
Data Stage STALL

Device
Setup Stage Busy
SET_CUR
Data Stage STALL

Setup Stage

Data Stage
SET_CUR
<=10ms
Status Stage

>10ms

State
Change
Control Change interrupt

Figure 2-22 Control Transfer Example (Case 4)

Revision 1.5 August 9, 2012 41

 USB Device Class Definition for Video Devices

SET/GET/Interrupt Supported/State Change Failure

Host Device

Setup Stage

Data Stage
SET_CUR
<=10ms
Status Stage

>10ms

State
Change
Control Change interrupt

Figure 2-23 Control Transfer Example (Case 5)

Revision 1.5 August 9, 2012 42

 USB Device Class Definition for Video Devices

3 Descriptors
Descriptors are used by USB devices to report their attributes. A descriptor is a data structure
with a defined format. For information, see section 9.5 Descriptors of USB Specification
Revision 2.0.
The following sections describe the standard and class-specific USB descriptors for the Video
Interface Class.

Revision 1.5 August 9, 2012 43

 USB Device Class Definition for Video Devices

3.1 Descriptor Layout Overview

The following diagram illustrates the descriptor layout for an entire device. The example used in
this case is for a desktop video camera device with a single isochronous video pipe and a
dedicated bulk still image pipe. If the video pipe supported temporally encoded video formats,
the descriptor layout would include an Encoding Unit.

Figure 3-1 Video Camera Descriptor Layout Example

3.2 Device Descriptor

Because video functionality is always considered to reside at the interface level, this class
specification does not define a specific video device descriptor.

Revision 1.5 August 9, 2012 44

 USB Device Class Definition for Video Devices

For devices that contain a video function that only exposes a VideoControl Interface, the device
descriptor must indicate that class information is to be found at the interface level. Therefore, the
bDeviceClass field of the device descriptor must contain zero so that enumeration software
looks down at the interface level to determine the Interface Class. The bDeviceSubClass and
bDeviceProtocol fields must be set to zero.

Devices that expose one or more Video Interface Collections also indicate that class information
is to be found at the interface level. However, since the device uses an Interface Association
Descriptor in order to describe the Video Interface Collection, it must set the bDeviceClass,
bDeviceSubClass and bDeviceProtocol fields 0xEF, 0x02 and 0x01 respectively. This set of
class codes is defined as the Multi-interface Function Class codes.

All other fields of the device descriptor must comply with the definitions in section 9.6.1
"Device" of the appropriate USB specification (USB Specification Revision 2.0 or USB
Specification Revision 3.0). There is no class-specific device descriptor.

3.3 Device_Qualifier Descriptor (deprecated)

3.4 Configuration Descriptor

The configuration descriptor for a device containing a video function is identical to the standard
Configuration descriptor defined in section 9.6.3 "Configuration" of USB Specification Revision
2.0 or USB Specification Revision 3.0. There is no class-specific configuration descriptor.

3.4.1 Backwards Compatibility

Separate configurations should be used when a device wishes to support more than one version
of the USB Video Class specification. Configurations shall be listed by increasing version,
starting from index 0.

3.5 Other_Speed_Configuration Descriptor (deprecated)

3.6 Interface Association Descriptor

A device must use an Interface Association Descriptor to describe a Video Interface Collection
for each device function that requires a VideoControl Interface and one or more VideoStreaming
interfaces. The standard VIC Interface Association Descriptor is identical to the standard
Interface Association Descriptor defined in the Interface Association Descriptor ECN, except
that some fields now have dedicated values.

If the VideoControl interface is part of a Video Interface Collection, the iFunction field in the
IAD and the iInterface field in the Standard VC interface descriptor for this Video Interface
Collection must be equal.

Revision 1.5 August 9, 2012 45

 USB Device Class Definition for Video Devices

Table 3-1 Standard Video Interface Collection IAD

Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 8
1 bDescriptorType 1 Constant INTERFACE ASSOCIATION
Descriptor.
2 bFirstInterface 1 Number Interface number of the first
VideoControl interface that is
associated with this function.
3 bInterfaceCount 1 Number Number of contiguous VideoStreaming
interfaces that are associated with this
function. The count includes the first
VideoControl interface and all its
associated VideoStreaming interfaces.
4 bFunctionClass 1 Class CC_VIDEO. Video Interface Class
code (assigned by the USB). See
section A.1, "Video Interface Class
Code".
5 bFunctionSubClass 1 SubClass SC_VIDEO_INTERFACE_COLLECT
ION. Video Interface Subclass code.
Assigned by this specification. See
section A.2, "Video Interface Subclass
Codes".
6 bFunctionProtocol 1 Protocol Not used. Must be set to
PC_PROTOCOL_UNDEFINED.
7 iFunction 1 Index Index of a string descriptor that
describes this interface. This must be
used for the device (function) name and
be implemented in US English
(LANGID = 0x0409) at the minimum.

3.7 VideoControl Interface Descriptors

The VideoControl (VC) interface descriptors contain all relevant information to fully
characterize the corresponding video function. The standard interface descriptor characterizes the
interface itself, whereas the class-specific interface descriptor provides pertinent information
concerning the internals of the video function. It specifies revision level information and lists the
capabilities of each Unit and Terminal.

3.7.1 Standard VC Interface Descriptor

The standard VC interface descriptor is identical to the standard interface descriptor defined in
section 9.6.5 "Interface" of USB Specification Revision 2.0, except that some fields have now
dedicated values.

Revision 1.5 August 9, 2012 46

 USB Device Class Definition for Video Devices

Table 3-2 Standard VC Interface Descriptor

Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 9
1 bDescriptorType 1 Constant INTERFACE descriptor type
2 bInterfaceNumber 1 Number Number of interface. A zero-based value
identifying the index in the array of
concurrent interfaces supported by this
configuration.
3 bAlternateSetting 1 Number Value used to select an alternate setting for
the interface identified in the prior field.
4 bNumEndpoints 1 Number Number of endpoints used by this interface
(excluding endpoint 0). This number is 0 or
1 depending on whether the optional status
interrupt endpoint is present.
5 bInterfaceClass 1 Class CC_VIDEO. Video Interface Class code
(assigned by the USB). See section A.1,
"Video Interface Class Code".
6 bInterfaceSubClass 1 Subclass SC_VIDEOCONTROL. Video Interface
Subclass code. Assigned by this
specification. See section A.2, "Video
Interface Subclass Codes".
7 bInterfaceProtocol 1 Protocol Must be set to PC_PROTOCOL_15.
8 iInterface 1 Index Index of a string descriptor that describes
this interface. This must be used for the
device (function) name and be
implemented in US English (LANGID =
0x0409) at the minimum.

3.7.2 Class-Specific VC Interface Descriptor

The class-specific VC interface descriptor is a concatenation of all the descriptors that are used
to fully describe the video function, i.e., all Unit Descriptors (UDs) and Terminal Descriptors
(TDs).

The total length of the class-specific VC interface descriptor depends on the number of Units and
Terminals in the video function. Therefore, the descriptor starts with a header that reflects the
total length in bytes of the entire class-specific VC interface descriptor in the wTotalLength
field. The bcdUVC field identifies the release of the Video Device Class Specification with
which this video function and its descriptors are compliant. The bInCollection field indicates
how many VideoStreaming interfaces there are in the Video Interface Collection to which this
VideoControl interface belongs. The baInterfaceNr() array contains the interface numbers of all
the VideoStreaming interfaces in the Collection. The bInCollection and baInterfaceNr() fields

Revision 1.5 August 9, 2012 47

 USB Device Class Definition for Video Devices

together provide all necessary information to determine which interfaces together constitute the
entire USB interface to the video function, i.e., describe the Video Interface Collection.

The order in which the Unit and Terminal descriptors are reported is not important, because
every descriptor can be identified through its bDescriptorType and bDescriptorSubtype fields.
The bDescriptorType field identifies the descriptor as being a class-specific interface descriptor.
The bDescriptorSubtype field further qualifies the exact nature of the descriptor.

The following table defines the class-specific VC interface header descriptor.

Table 3-3 Class-specific VC Interface Header Descriptor
Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 12+n
1 bDescriptorType 1 Constant CS_INTERFACE descriptor type
2 bDescriptorSubType 1 Constant VC_HEADER descriptor subtype
3 bcdUVC 2 BCD Video Device Class Specification release
number in binary-coded decimal. (i.e. 2.10
is 210H and 1.50 is 150H)
5 wTotalLength 2 Number Total number of bytes returned for the
class-specific VideoControl interface
descriptor. Includes the combined length
of this descriptor header and all Unit and
Terminal descriptors.
7 dwClockFrequency 4 Number Use of this field has been deprecated.

The device clock frequency in Hz. This

will specify the units used for the time
information fields in the Video Payload
Headers of the primary data stream and
format.

The dwClockFrequency field of the

Video Probe and Commit control replaces
this descriptor field. A value for this field
shall be chosen such that the primary or
default function of the device will be
available to host software that implements
Version 1.0 of this specification.
11 bInCollection 1 Number The number of VideoStreaming interfaces
in the Video Interface Collection to which
this VideoControl interface belongs: n
12 baInterfaceNr(1) 1 Number Interface number of the first
VideoStreaming interface in the Collection
… … … … …

Revision 1.5 August 9, 2012 48

 USB Device Class Definition for Video Devices

12+(n- baInterfaceNr(n) 1 Number Interface number of the last

1) VideoStreaming interface in the Collection

This header is followed by one or more Unit and/or Terminal Descriptors. The layout of the
descriptors depends on the type of Unit or Terminal they represent. There is a descriptor type for
each Unit and Terminal described in section 2.3, "Video Function Topology". They are
summarized in the following sections. The first four fields are common for all Unit and Terminal
Descriptors. They contain the Descriptor Length, Descriptor Type, Descriptor Subtype, and Unit
or Terminal ID.

Each Unit and Terminal within the video function is assigned a unique identification number, the
Unit ID (UID) or Terminal ID (TID), contained in the bUnitID or bTerminalID field of the
descriptor. The value 0x00 is reserved for undefined ID, effectively restricting the total number
of addressable entities in the video function (both Units and Terminals) to 255.

Besides uniquely identifying all addressable entities in a video function, the IDs also serve to
describe the topology of the video function; i.e., the bSourceID field of a Unit or Terminal
descriptor indicates to which other Unit or Terminal this Unit or Terminal is connected.

3.7.2.1 Input Terminal Descriptor

The Input Terminal descriptor (ITD) provides information to the Host that is related to the
functional aspects of the Input Terminal.

The Input Terminal is uniquely identified by the value in the bTerminalID field. No other Unit
or Terminal within the same video function may have the same ID. This value must be passed in
the bTerminalID field of each request that is directed to the Terminal.

The wTerminalType field provides pertinent information about the physical entity that the Input
Terminal represents. This could be a USB OUT endpoint, an external Composite Video In
connection, a camera sensor, etc. A complete list of Terminal Type codes is provided in section
B.2, "Input Terminal Types".

The bAssocTerminal field is used to associate an Output Terminal to this Input Terminal,
effectively implementing a bi-directional Terminal pair. An example of this would be a tape unit
on a camcorder, which would have Input and Output Terminals to sink and source video
respectively. If the bAssocTerminal field is used, both associated Terminals must belong to the
bi-directional Terminal Type group. If no association exists, the bAssocTerminal field must be
set to zero.

The Host software can treat the associated Terminals as being physically related. In many cases,
one Terminal can not exist without the other. An index to a string descriptor is provided to
further describe the Input Terminal.

The following table presents an outline of the Input Terminal descriptor.

Revision 1.5 August 9, 2012 49

 USB Device Class Definition for Video Devices

Table 3-4 Input Terminal Descriptor

Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 8 (+ x)
1 bDescriptorType 1 Constant CS_INTERFACE descriptor type
2 bDescriptorSubtype 1 Constant VC_INPUT_TERMINAL descriptor subtype
3 bTerminalID 1 Constant A non-zero constant that uniquely identifies
the Terminal within the video function. This
value is used in all requests to address this
Terminal.
4 wTerminalType 2 Constant Constant that characterizes the type of
Terminal. See Appendix B, "Terminal Types".
6 bAssocTerminal 1 Constant ID of the Output Terminal to which this Input
Terminal is associated, or zero (0) if no such
association exists.
7 iTerminal 1 Index Index of a string descriptor, describing the
Input Terminal.
… … … … Depending on the Terminal type, certain Input
Terminal descriptors have additional fields.
The descriptors for these special Terminal
types are described in separate sections
specific to those Terminals, and in
accompanying documents.

3.7.2.2 Output Terminal Descriptor

The Output Terminal descriptor (OTD) provides information to the Host that is related to the
functional aspects of the Output Terminal.

The Output Terminal is uniquely identified by the value in the bTerminalID field. No other Unit
or Terminal within the same video function may have the same ID. This value must be passed in
the bTerminalID field of each request that is directed to the Terminal.

The wTerminalType field provides pertinent information about the physical entity the Output
Terminal represents. This could be a USB IN endpoint, an external Composite Video Out
connection, a LCD display, etc. A complete list of Terminal Type codes is provided in section
B.3, "Output Terminal Types".

The bAssocTerminal field is used to associate an Input Terminal to this Output Terminal,
effectively implementing a bi-directional Terminal pair. If the bAssocTerminal field is used,
both associated Terminals must belong to the bi-directional Terminal Type group. If no
association exists, the bAssocTerminal field must be set to zero.

Revision 1.5 August 9, 2012 50

 USB Device Class Definition for Video Devices

The bSourceID field is used to describe the connectivity for this Terminal. It contains the ID of
the Unit or Terminal to which this Output Terminal is connected via its Input Pin.
An index to a string descriptor is provided to further describe the Output Terminal.

The following table presents an outline of the Output Terminal descriptor.

Table 3-5 Output Terminal Descriptor
Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 9 (+ x)
1 bDescriptorType 1 Constant CS_INTERFACE descriptor type
2 bDescriptorSubtype 1 Constant VC_OUTPUT_TERMINAL descriptor
subtype
3 bTerminalID 1 Constant A non-zero constant that uniquely identifies
the Terminal within the video function. This
value is used in all requests to address this
Terminal.
4 wTerminalType 2 Constant Constant that characterizes the type of
Terminal. See Appendix B, "Terminal
Types".
6 bAssocTerminal 1 Constant Constant, identifying the Input Terminal to
which this Output Terminal is associated, or
zero (0) if no such association exists.
7 bSourceID 1 Constant ID of the Unit or Terminal to which this
Terminal is connected.
8 iTerminal 1 Index Index of a string descriptor, describing the
Output Terminal.
… … … … Depending on the Terminal type, certain
Output Terminal descriptors have additional
fields. The descriptors for these special
Terminal types are described in
accompanying documents.

3.7.2.3 Camera Terminal Descriptor

The Camera Terminal is uniquely identified by the value in the bTerminalID field. No other
Unit or Terminal within the same video function may have the same ID. This value must be
passed in the bTerminalID field of each request that is directed to the Terminal.

The wTerminalType field provides pertinent information about the physical entity that the Input
Terminal represents. For the Camera Terminal, this field shall be set to ITT_CAMERA
(see section B.2, “Input Terminal Types”).

Revision 1.5 August 9, 2012 51

 USB Device Class Definition for Video Devices

The bAssocTerminal field is used to associate an Output Terminal to this Input Terminal,
effectively implementing a bi-directional Terminal pair. An index to a string descriptor is
provided to further describe the Camera Terminal.

The bmControls field is a bitmap, indicating the availability of certain camera controls for the
video stream.

The layout of the Camera Terminal descriptor is detailed in the following table.
Table 3-6 Camera Terminal Descriptor
Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 18
1 bDescriptorType 1 Constant CS_INTERFACE descriptor type
2 bDescriptorSubtype 1 Constant VC_INPUT_TERMINAL
descriptor subtype
3 bTerminalID 1 Constant A non-zero constant that uniquely
identifies the Terminal within the
video function. This value is used
in all requests to address this
Terminal.
4 wTerminalType 2 Constant Constant that characterizes the type
of Terminal. This is set to the
ITT_CAMERA value.
6 bAssocTerminal 1 Constant ID of the Output Terminal to which
this Input Terminal is associated.
7 iTerminal 1 Index Index of a string descriptor that
describes the Camera Terminal.
8 wObjectiveFocalLengthMin 2 Number The value of Lmin If Optical Zoom
is not supported; this field shall be
set to 0.
10 wObjectiveFocalLengthMax 2 Number The value of Lmax If Optical Zoom
is not supported; this field shall be
set to 0.
12 wOcularFocalLength 2 Number The value of Locular If Optical
Zoom is not supported; this field
shall be set to 0.
14 bControlSize 1 Number Size in bytes of the bmControls
field: 3
15 bmControls 3 Bitmap A bit set to 1 indicates that the
mentioned Control is supported for
the video stream.
D0: Scanning Mode
D1: Auto-Exposure Mode

Revision 1.5 August 9, 2012 52

 USB Device Class Definition for Video Devices

D2: Auto-Exposure Priority

D3: Exposure Time (Absolute)
D4: Exposure Time (Relative)
D5: Focus (Absolute)
D6 : Focus (Relative)
D7: Iris (Absolute)
D8 : Iris (Relative)
D9: Zoom (Absolute)
D10: Zoom (Relative)
D11: PanTilt (Absolute)
D12: PanTilt (Relative)
D13: Roll (Absolute)
D14: Roll (Relative)
D15: Reserved
D16: Reserved
D17: Focus, Auto
D18: Privacy
D19: Focus, Simple
D20: Window
D21: Region of Interest
D22 – D23: Reserved, set to zero

3.7.2.4 Selector Unit Descriptor

The Selector Unit is uniquely identified by the value in the bUnitID field of the Selector Unit
descriptor (SUD). No other Unit or Terminal within the same video function may have the same
ID. This value must be passed with each request that is directed to the Selector Unit.

The bNrInPins field contains the number of Input Pins (p) of the Selector Unit. The connectivity
of the Input Pins is described via the baSourceID() array that contains p elements. The index i
into the array is one-based and directly related to the Input Pin numbers. baSourceID(i) contains
the ID of the Unit or Terminal to which Input Pin i is connected.

An index to a string descriptor is provided to further describe the Selector Unit.

The following table details the structure of the Selector Unit descriptor.
Table 3-7 Selector Unit Descriptor
Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 6+p
1 bDescriptorType 1 Constant CS_INTERFACE descriptor type
2 bDescriptorSubtype 1 Constant VC_SELECTOR_UNIT descriptor subtype
3 bUnitID 1 Number A non-zero constant that uniquely identifies
the Unit within the video function. This

Revision 1.5 August 9, 2012 53

 USB Device Class Definition for Video Devices

value is used in all requests to address this

Unit.
4 bNrInPins 1 Number Number of Input Pins of this Unit: p
5 baSourceID(1) 1 Number ID of the Unit or Terminal to which the first
Input Pin of this Selector Unit is connected.
… … … … …
5+(p-1) baSourceID(p) 1 Number ID of the Unit or Terminal to which the last
Input Pin of this Selector Unit is connected.
5+p iSelector 1 Index Index of a string descriptor, describing the
Selector Unit.

3.7.2.5 Processing Unit Descriptor

The Processing Unit is uniquely identified by the value in the bUnitID field of the Processing
Unit descriptor (PUD). No other Unit or Terminal within the same video function may have the
same ID. This value must be passed with each request that is directed to the Processing Unit.

The bSourceID field is used to describe the connectivity for this Processing Unit. It contains the
ID of the Unit or Terminal to which this Processing Unit is connected via its Input Pin.
bSourceID must refer to a Unit or Terminal in the same video function. The bmControls field
is a bit-map, indicating the availability of certain processing Controls for the video stream.

An index to a string descriptor is provided to further describe the Processing Unit.

The layout of the Processing Unit descriptor is detailed in the following table.
Table 3-8 Processing Unit Descriptor
Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 13
1 bDescriptorType 1 Constant CS_INTERFACE descriptor type
2 bDescriptorSubtype 1 Constant VC_PROCESSING_UNIT descriptor
subtype
3 bUnitID 1 Number A non-zero constant that uniquely identifies
the Unit within the video function. This value
is used in all requests to address this Unit.
4 bSourceID 1 Constant ID of the Unit or Terminal to which this Unit
is connected.
5 wMaxMultiplier 2 Number If the Digital Multiplier control is supported,
this field indicates the maximum digital
magnification, multiplied by 100. For
example, for a device that supports 1-4.5X
digital zoom (a multiplier of 4.5), this field
would be set to 450. If the Digital Multiplier

Revision 1.5 August 9, 2012 54

 USB Device Class Definition for Video Devices

control is not supported, this field shall be set

to 0.
7 bControlSize 1 Number Size of the bmControls field, in bytes: 3
8 bmControls 3 Bitmap A bit set to 1 indicates that the mentioned
Control is supported for the video stream.
D0: Brightness
D1: Contrast
D2: Hue
D3: Saturation
D4: Sharpness
D5: Gamma
D6: White Balance Temperature
D7: White Balance Component
D8: Backlight Compensation
D9: Gain
D10: Power Line Frequency
D11: Hue, Auto
D12: White Balance Temperature, Auto
D13: White Balance Component, Auto
D14: Digital Multiplier
D15: Digital Multiplier Limit
D16: Analog Video Standard
D17: Analog Video Lock Status
D18: Contrast, Auto
D19 – D23: Reserved. Set to zero.
11 iProcessing 1 Index Index of a string descriptor that describes this
processing unit.
12 bmVideoStandards 1 Bitmap A bitmap of all analog video standards
supported by the Processing Unit.

A value of zero indicates that this bitmap

should be ignored.

D0: None
D1: NTSC – 525/60
D2: PAL – 625/50
D3: SECAM – 625/50
D4: NTSC – 625/50
D5: PAL – 525/60
D6-D7: Reserved. Set to zero.

Revision 1.5 August 9, 2012 55

 USB Device Class Definition for Video Devices

3.7.2.6 Encoding Unit Descriptor

The Encoding Unit is uniquely identified by the value in the bUnitID field of the Encoding Unit
descriptor (EUD). No other Unit or Terminal within the same video function may have the same
ID. This value must be passed with each request that is directed to the Encoding Unit.

The bSourceID field is used to describe the connectivity for this Encoding Unit. It contains the
ID of the Unit or Terminal to which this Encoding Unit is connected via its Input Pin.
bSourceID must refer to a Unit or Terminal in the same video function. The bmControls field is
a bit-map, indicating the availability of certain encoding Controls for the video stream.
An index to a string descriptor is provided to further describe the Encoding Unit.
The layout of the Encoding Unit descriptor is detailed in the following table.

Table 3-9 Encoding Unit Descriptor

Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes. The
value must be 13.
1 bDescriptorType 1 Constant CS_INTERFACE descriptor type.
2 bDescriptorSubtype 1 Constant VC_ENCODING_UNIT descriptor
subtype.
3 bUnitID 1 Number A non-zero constant that uniquely
identifies the Unit within the video
function. This value is used in all
requests to address this Unit.
4 bSourceID 1 Constant ID of the Unit or Terminal to which this
Unit is connected.
5 iEncoding 1 Index Index of a string descriptor that describes
this encoding unit.
6 bControlSize 1 Number Size, in bytes, of the bmControls and
bmControlsRuntime fields: The value
must be 3.
7 bmControls 3 Bitmap A bit set to 1 indicates that the specified
control is supported for initialization:
D0: Select Layer
D1: Profile and Toolset
D2: Video Resolution
D3: Minimum Frame Interval
D4: Slice Mode
D5: Rate Control Mode
D6: Average Bit Rate
D7: CPB Size
D8: Peak Bit Rate
D9: Quantization Parameter

Revision 1.5 August 9, 2012 56

 USB Device Class Definition for Video Devices

D10: Synchronization and Long-Term

Reference Frame
D11: Long-Term Buffer
D12: Picture Long-Term Reference
D13: LTR Validation
D14: Level IDC
D15: SEI Message
D16: QP Range
D17: Priority ID
D18: Start or Stop Layer/View
D19: Error Resiliency
D20 - 23: Reserved; set to zero
10 bmControlsRuntime 3 Bitmap A bit set to 1 indicates that the
mentioned
control is supported during runtime:
D0: Select Layer
D1: Profile and Toolset.
D2: Video Resolution
D3: Minimum Frame Interval
D4: Slice Mode.
D5: Rate Control Mode
D6: Average Bit Rate
D7: CPB Size
D8: Peak Bit Rate.
D9: Quantization Parameter
D10: Synchronization and Long-Term
Reference Frame
D11: Long-Term Buffer
D12: Picture Long-Term Reference
D13: LTRValidation
D14: Level IDC
D15: SEI Message
D16: QP Range
D17: Priority ID.
D18: Start or Stop Layer/View
D19: Error Resiliency
D20 – D23: Reserved, set to zero.

The Encoding Unit Descriptor supports separate lists for bmControls and bmRuntimeControls.
The use here of two lists reflects the expectation that many UVC devices will not be able to
support the same features while streaming video as during initialization. This is partially due to
the inherent asynchronous nature of encoder control offered over USB.

Revision 1.5 August 9, 2012 57

 USB Device Class Definition for Video Devices

3.7.2.7 Extension Unit Descriptor

The Extension Unit is uniquely identified by the value in the bUnitID field of the Extension Unit
descriptor (XUD). No other Unit or Terminal within the same video function may have the same
ID. This value must be passed with each request that is directed to the Extension Unit.

The Extension Unit Descriptor allows the hardware designer to define any arbitrary set of
controls such that a class driver can act as an intermediary between vendor-supplied host
software and functionality of the device.

The guidExtensionCode field contains a vendor-specific code that further identifies the
Extension Unit.

The bNrInPins field contains the number of Input Pins (p) of the Extension Unit. The
connectivity of the Input Pins is described via the baSourceID() array that contains p elements.
The index i into the array is one-based and directly related to the Input Pin numbers.
baSourceID(i) contains the ID of the Unit or Terminal to which Input Pin i is connected.

The bmControls field is a bitmap, indicating the availability of certain video Controls in the
Extension Unit. For future expandability, the number of bytes occupied by the bmControls field
is indicated in the bControlSize field. All Controls are optional.

An index to a string descriptor is provided to further describe the Extension Unit.

The following table outlines the Extension Unit descriptor.

Table 3-10 Extension Unit Descriptor
Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 24+p+n
1 bDescriptorType 1 Constant CS_INTERFACE descriptor type
2 bDescriptorSubtype 1 Constant VC_EXTENSION_UNIT descriptor
subtype
3 bUnitID 1 Number A non-zero constant that uniquely
identifies the Unit within the video
function. This value is used in all requests
to address this Unit.
4 guidExtensionCode 16 GUID Vendor-specific code identifying the
Extension Unit
20 bNumControls 1 Number Number of controls in this extension unit
21 bNrInPins 1 Number Number of Input Pins of this Unit: p
22 baSourceID(1) 1 Number ID of the Unit or Terminal to which the
first Input Pin of this Extension Unit is
connected.
… … … … …

Revision 1.5 August 9, 2012 58

 USB Device Class Definition for Video Devices

22+(p-1) baSourceID(p) 1 Number ID of the Unit or Terminal to which the

last Input Pin of this Extension Unit is
connected.
22+p bControlSize 1 Number Size of the bmControls field, in bytes: n
23+p bmControls n Bitmap A bit set to 1 indicates that the mentioned
Control is supported:
D(n*8-1)..0: Vendor-specific
23+p+n iExtension 1 Index Index of a string descriptor that describes
this extension unit.

3.8 VideoControl Endpoint Descriptors

The following sections describe all possible endpoint-related descriptors for the VideoControl
interface.

3.8.1 VC Control Endpoint Descriptors

3.8.1.1 Standard VC Control Endpoint Descriptor

Because endpoint 0 is used as the VideoControl control endpoint, there is no dedicated standard
control endpoint descriptor.

3.8.1.2 Class-Specific VC Control Endpoint Descriptor

There is no dedicated class-specific control endpoint descriptor.

3.8.2 VC Interrupt Endpoint Descriptors

The standard and class-specific Interrupt Endpoint descriptors provide all information about the
device interrupt usage.

3.8.2.1 Standard VC Interrupt Endpoint Descriptor

The interrupt endpoint descriptor is identical to the standard endpoint descriptor defined in
section 9.6.6 "Endpoint" of USB Specification Revision 2.0 or USB Specification Revision 3.0. Its
fields are set to reflect the interrupt type of the endpoint. This endpoint is optional.

The following table outlines the standard VC Interrupt Endpoint descriptor.

Table 3-11 Standard VC Interrupt Endpoint Descriptor
Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 7
1 bDescriptorType 1 Constant ENDPOINT descriptor type
2 bEndpointAddress 1 Endpoint The address of the endpoint on the USB
device described by this descriptor. The
address is encoded as follows:
D7: Direction. 1 = IN endpoint

Revision 1.5 August 9, 2012 59

 USB Device Class Definition for Video Devices

D6..4: Reserved, reset to zero.

D3..0: The endpoint number, determined by
the designer.
3 bmAttributes 1 Bitmap D3..2: Synchronization type.
Must be set to 00 (None)
D1..0: Transfer type.
Must be set to 11 (Interrupt).
All other bits are reserved, and must be set to
zero.
4 wMaxPacketSize 2 Number Maximum packet size this endpoint is
capable of sending or receiving when this
configuration is selected.
6 bInterval 1 Number Interval for polling endpoint for data
transfers.
For full-speed interrupt endpoints, this value
is expressed in frames, and must range from
1 to 255.
For high-speed interrupt endpoints, the
bInterval value is used as the exponent for a
2bInterval-1 value; e.g., a bInterval of 4 means
a period of 8 (23). This value must be from 1
to 16.

3.8.2.2 Class-specific VC Interrupt Endpoint Descriptor

The class-specific interrupt endpoint descriptor provides information about the maximum
interrupt structure size that the device is capable of sending. The host driver will use this value to
allocate a buffer of sufficient size to receive the maximum interrupt structure size. This
descriptor is mandatory if the standard interrupt endpoint descriptor is defined.
Table 3-12 Class-specific VC Interrupt Endpoint Descriptor
Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 5
1 bDescriptorType 1 Constant CS_ENDPOINT descriptor type
2 bDescriptorSubType 1 Constant EP_INTERRUPT descriptor type
3 wMaxTransferSize 2 Number Maximum interrupt structure size this
endpoint is capable of sending.

3.9 VideoStreaming Interface Descriptors

The VideoStreaming (VS) interface descriptors contain all relevant information to characterize
the VideoStreaming interface in full.

Revision 1.5 August 9, 2012 60

 USB Device Class Definition for Video Devices

3.9.1 Standard VS Interface Descriptor

The standard VS interface descriptor is identical to the standard interface descriptor defined in
section 9.6.5 "Interface" of USB Specification Revision 2.0 or USB Specification Revision 3.0,
except that some fields now have dedicated values.
Table 3-13 Standard VS Interface Descriptor
Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 9
1 bDescriptorType 1 Constant INTERFACE descriptor type
2 bInterfaceNumber 1 Number Number of the interface. A zero-based value
identifying the index in the array of concurrent
interfaces supported by this configuration.
3 bAlternateSetting 1 Number Value used to select this alternate setting for
the interface identified in the prior field.
4 bNumEndpoints 1 Number Number of endpoints used by this interface
(excluding endpoint 0).
5 bInterfaceClass 1 Class CC_VIDEO. Video Interface Class code
(assigned by the USB). See section A.1,
"Video Interface Class Code".
6 bInterfaceSubClass 1 subclass SC_VIDEOSTREAMING. Video interface
subclass code (assigned by this specification).
See section A.2, "Video Interface Subclass
Codes".
7 bInterfaceProtocol 1 Protocol Must be set to PC_PROTOCOL_15.
8 iInterface 1 Index Index of a string descriptor that describes this
interface.

3.9.2 Class-Specific VS Interface Descriptors

The class-specific VS interface descriptors consist of Input Header, Output Header, Format and
Frame descriptors.
There is a single Input or Output Header descriptor for each VS interface, and a separate Format
descriptor for each supported video stream format and a separate list of Frame descriptors for
each Format descriptor (if the Format requires Frame descriptors). Header, Format and Frame
descriptors are only defined in alternate setting 0 of the relevant interface. They are not repeated
within subsequent alternate settings of the same interface.

3.9.2.1 Input Header Descriptor

The Input Header descriptor is used for VS interfaces that contain an IN endpoint for streaming
video data. It provides information on the number of different Format descriptors that will follow
it, as well as the total size of all class-specific descriptors in alternate setting zero of this
interface.

Revision 1.5 August 9, 2012 61

 USB Device Class Definition for Video Devices

The following table defines the class-specific VS interface Input Header descriptor.
Table 3-14 Class-specific VS Interface Input Header Descriptor
Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 13+(p*n).
1 bDescriptorType 1 Constant CS_INTERFACE descriptor type
2 bDescriptorSubtype 1 Constant VS_INPUT_HEADER descriptor subtype
3 bNumFormats 1 Number Number of video payload Format
descriptors following for this interface
(excluding video Frame descriptors): p
4 wTotalLength 2 Number Total number of bytes returned for the
class-specific VideoStreaming interface
descriptors including this header
descriptor.
6 bEndpointAddress 1 Endpoint The address of the isochronous or bulk
endpoint used for video data. The address
is encoded as follows:
D7: Direction
1 = IN endpoint
D6..4: Reserved, set to zero.
D3..0: The endpoint number, determined
by the designer.
7 bmInfo 1 Bitmap Indicates the capabilities of this
VideoStreaming interface:
D0: Dynamic Format Change supported
D7..1: Reserved, set to zero.
8 bTerminalLink 1 Constant The terminal ID of the Output Terminal to
which the video endpoint of this interface
is connected.
9 bStillCaptureMethod 1 Number Method of still image capture supported as
described in section 2.4.2.4, "Still Image
Capture":
0: None (Host software will not support
any form of still image capture)
1: Method 1
2: Method 2
3: Method 3
10 bTriggerSupport 1 Number Specifies if hardware triggering is
supported through this interface
0: Not supported
1: Supported
11 bTriggerUsage 1 Number Specifies how the host software shall
respond to a hardware trigger interrupt
event from this interface. This is ignored if

Revision 1.5 August 9, 2012 62

 USB Device Class Definition for Video Devices

the bTriggerSupport field is zero.

0: Initiate still image capture
1: General purpose button event. Host
driver will notify client application of
button press and button release events
12 bControlSize 1 Number Size of each bmaControls(x) field, in
bytes: n
13 bmaControls(1) n Bitmap For bits D3..0, a bit set to 1 indicates that
the named field is supported by the Video
Probe and Commit Control when
bFormatIndex is 1:
D0: wKeyFrameRate
D1: wPFrameRate
D2: wCompQuality
D3: wCompWindowSize

For bits D5..4, a bit set to 1 indicates that

the named control is supported by the
device when bFormatIndex is 1:
D4: Generate Key Frame
D5: Update Frame Segment
D6..(n*8-1): Reserved, set to zero

*Note* going forward from version 1.5 the

proper means to detect whether a field is
supported by Probe & Commit is to issue a
VS_PROBE_CONTROL(GET_CUR).
… … … … …
13+ bmaControls(p) n Bitmap For bits D3..0, a bit set to 1 indicates that
(p*n- the named field is supported by the Video
n) Probe and Commit Control when
bFormatIndex is p:
D0: wKeyFrameRate
D1: wPFrameRate
D2: wCompQuality
D3: wCompWindowSize

For bits D5..4, a bit set to 1 indicates that

the named control is supported by the
device when bFormatIndex is p:
D4: Generate Key Frame
D5: Update Frame Segment
D6..(n*8-1): Reserved, set to zero

Revision 1.5 August 9, 2012 63

 USB Device Class Definition for Video Devices

*Note* D0-D3 are deprecated. Going

forward from version 1.5 the proper means
to detect whether a field is supported by
Probe & Commit is to issue a
VS_PROBE_CONTROL(GET_CUR).

3.9.2.2 Output Header Descriptor

The Output Header descriptor is used for VS interfaces that contain an OUT endpoint for
streaming video data. It provides information on the number of different Format descriptors that
will follow it, as well as the total size of all class-specific descriptors in alternate setting zero of
this interface.

The following table defines the class-specific VS interface output header descriptor:
Table 3-15 Class-specific VS Interface Output Header Descriptor
Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 9+(p*n)
1 bDescriptorType 1 Constant CS_INTERFACE descriptor type
2 bDescriptorSubtype 1 Constant VS_OUTPUT_HEADER descriptor
subtype
3 bNumFormats 1 Number Number of video payload Format
descriptors following for this interface
(excluding video Frame descriptors): p
4 wTotalLength 2 Number Total number of bytes returned for the
class-specific VideoStreaming interface
descriptors including this header
descriptor.
6 bEndpointAddress 1 Endpoint The address of the isochronous or bulk
endpoint used for video data. The address
is encoded as follows:
D7: Direction
0 = OUT endpoint
D6..4: Reserved, set to zero
D3..0: The endpoint number, determined
by the designer.
7 bTerminalLink 1 Constant The terminal ID of the Input Terminal to
which the video endpoint of this interface
is connected.
8 bControlSize 1 Number Size of each bmaControls(x) field, in
bytes: n
9 bmaControls(1) n Bitmap For bits D3..0, a bit set to 1 indicates that
the named field is supported by the Video
Probe and Commit Control when

Revision 1.5 August 9, 2012 64

 USB Device Class Definition for Video Devices

bFormatIndex is 1:
D0: wKeyFrameRate
D1: wPFrameRate
D2: wCompQuality
D3: wCompWindowSize

D4..(n*8-1): Reserved, set to zero

*Note* D0-D3 are deprecated. Going

forward from version 1.5, the proper
means to detect whether a field is
supported by Probe & Commit is to issue
a VS_PROBE_CONTROL(GET_CUR).
… … … … …
9+(p*n- bmaControls(p) n Bitmap For bits D3..0, a bit set to 1 indicates that
n) the named field is supported by the Video
Probe and Commit Control when
bFormatIndex is p:
D0: wKeyFrameRate
D1: wPFrameRate
D2: wCompQuality
D3: wCompWindowSize

D4..(n*8-1): Reserved, set to zero

3.9.2.3 Payload Format Descriptors

A Payload Format descriptor defines the characteristics of a video stream with its specific
format. The following descriptors are defined in the separate Payload Specification documents
that accompany this document. For more information about a specific descriptor, see the
corresponding document.
Table 3-16 Payload Format Descriptor
Payload Format Descriptor Document
Uncompressed Video USB_Video_Payload_Uncompressed
MJPEG Video USB_Video_Payload_MJPEG
MPEG1-SS USB_Video_Payload_Stream_Based
MPEG2-PS USB_Video_Payload_Stream_Based
MPEG-2 TS USB_Video_Payload_MPEG2-TS
H.264 USB_Video_Payload_H264
VP8 USB_Video_Payload_VP8
SMTPE VC1 USB_Video_Payload_MPEG2-TS
MPEG-4 SL USB_Video_Payload_MPEG2-TS

Revision 1.5 August 9, 2012 65

 USB Device Class Definition for Video Devices

DV USB_Video_Payload_DV
Vendor Defined USB_Video_Payload_Stream_Based or
USB_Video_Payload_Frame_Based

3.9.2.4 Video Frame Descriptor

A Video Frame descriptor (or Frame descriptor for short) is used to describe the decoded video
and still image frame dimensions and other frame-specific characteristics supported by Frame-
based formats. Frame descriptors (if required) immediately follow the associated Format
descriptor.
The following Video Frame descriptors are defined in the separate Payload Specification
documents that accompany this document. For more information about a specific Frame
descriptor, see the corresponding document.
Table 3-17 Defined Video Frame Descriptor Resources
Video Frame Descriptor Document
Uncompressed USB_Video_Payload_Uncompressed
MJPEG USB_Video_Payload_MJPEG
Generic Frame-Based USB_Video_Payload_Frame_Based
H.264 USB_Video_Payload_H.264
VP8 USB_Video_Payload_VP8

3.9.2.5 Still Image Frame Descriptor

The Still Image Frame descriptor is only applicable for a VS interface that supports method 2 or
3 of still image capture in conjunction with frame-based Payload formats (e.g., MJPEG,
uncompressed, etc.). The Still Image Frame descriptor defines the characteristics of the still
image capture for these frame-based formats. A single still Image Frame descriptor follows the
Frame descriptor(s) for each Format descriptor group. If the Input Header descriptor’s
bStillCaptureMethod field is set to method 2 or 3, this Still Image Frame descriptor shall be
defined (see section 3.9.2.1, "Input Header Descriptor").

The Still Image Frame descriptor contains the range of image sizes available from the device,
which comprise the list of possible still image formats. To select a particular still image format,
host software sends control requests to the corresponding interface (see section 4.3.1.2, "Video
Still Probe Control and Still Commit Control").

The Still Image Frame descriptor is shown in Table 3-18 Still Image Frame Descriptor below.

The bEndpointAddress field contains the bulk endpoint address within the related VS interface
that is used for still image capture. The endpoint always functions as an IN-Endpoint.

The wWidth(x) and wHeight(x) fields form an array of image sizes supported by the device,
measured in pixels of an uncompressed image.

Revision 1.5 August 9, 2012 66

 USB Device Class Definition for Video Devices

The bNumImageSizePatterns represents the number of wWidth and wHeight pairs in the
array.

The bCompression field represents the image quality that would be generated by the device.
The range of compression values is from 0 to 255. A small value indicates a low compression
ratio and high quality image. The default setting of this value depends on device implementation.
The bCompression(x) fields form an array of compression ratios supported by the device for all
image sizes. The bNumCompressionPatterns field represents the number of bCompression
fields in this array.

A Still Image Frame descriptor identifies the following:

Table 3-18 Still Image Frame Descriptor
Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 10+(4*n)-
4+m
1 bDescriptorType 1 Constant CS_INTERFACE descriptor type
2 bDescriptorSubtype 1 Constant VS_STILL_IMAGE_FRAME descriptor
subtype
3 bEndpointAddress 1 Endpoint If method 3 of still image capture is used,
this contains the address of the bulk
endpoint used for still image capture. The
address is encoded as follows:
D7: Direction. (set to 1 = IN endpoint)
D6..4: Reserved, reset to zero
D3..0: The endpoint number, determined
by the designer
If method 2 of still image capture is used,
this field shall be set to zero.
4 bNumImageSizePat 1 Number Number of Image Size patterns of this
terns format: n
5 wWidth(1) 2 Number Width of the still image in pattern 1
7 wHeight(1) 2 Number Height of the still image in pattern 1
… … … … …
… … … … …
5+4*n- wWidth(n) 2 Number Width of the still image in pattern n
4
7+4*n- wHeight(n) 2 Number Height of the still image in pattern n
4
9+4*n- bNumCompression 1 Number Number of Compression pattern of this
4 Pattern format: m
10+4*n bCompression(1) 1 Number Compression of the still image in pattern 1
-4

Revision 1.5 August 9, 2012 67

 USB Device Class Definition for Video Devices

… … … … …
10+4*n bCompression(m) 1 Number Compression of the still image in pattern
-4+m-1 m

3.9.2.6 Color Matching Descriptor

The Color Matching descriptor is an optional descriptor used to describe the color profile of the
video data in an unambiguous way. Only one instance is allowed for a given format and if
present, the Color Matching descriptor shall be placed following the Video and Still Image
Frame descriptors for that format.

For example, this descriptor would be used with Uncompressed Video, MJPEG and MPEG-1
formats. It would not be used in the case MPEG-2, DV or MPEG-4 because the information is
already available implicitly (DV) or explicitly (MPEG-2, MPEG-4). If a format requires this
descriptor, the corresponding payload specification must enforce this requirement.
In the absence of this descriptor, or in the case of “Unspecified” values within the descriptor,
color matching defaults will be assumed. The color matching defaults are compliant with sRGB
since the BT.709 transfer function and the sRGB transfer function are very similar.

The viewing conditions and monitor setup are implicitly based on sRGB and the device should
compensate for them (D50 ambient white, dim viewing or 64 lux ambient illuminance, 2.2
gamma reference CRT, etc).
Table 3-19 Color Matching Descriptor
Offset Field Size Value Description
0 bLength 1 Constant 6
1 bDescriptorType 1 Number CS_INTERFACE type
2 bDescriptorSubtype 1 Number VS_COLORFORMAT
3 bColorPrimaries 1 Number This defines the color primaries
and the reference white.
0: Unspecified (Image
characteristics unknown)
1: BT.709, sRGB (default)
2: BT.470-2 (M)
3: BT.470-2 (B, G)
4: SMPTE 170M
5: SMPTE 240M
6-255: Reserved
4 bTransferCharacteristics 1 Number This field defines the opto-
electronic transfer characteristic of
the source picture also called the
gamma function.
0: Unspecified (Image
characteristics unknown)

Revision 1.5 August 9, 2012 68

 USB Device Class Definition for Video Devices

1: BT.709 (default)
2: BT.470-2 M
3: BT.470-2 B, G
4: SMPTE 170M
5: SMPTE 240M
6: Linear (V = Lc)
7: sRGB (very similar to BT.709)
8-255: Reserved
5 bMatrixCoefficients 1 Number Matrix used to compute luma and
chroma values from the color
primaries.
0: Unspecified (Image
characteristics unknown)
1: BT. 709
2: FCC
3: BT.470-2 B, G
4: SMPTE 170M (BT.601,
default)
5: SMPTE 240M
6-255: Reserved

3.10 VideoStreaming Endpoint Descriptors

The following sections describe all possible endpoint-related descriptors for the VideoStreaming
interface.

3.10.1 VS Video Data Endpoint Descriptors

The video data endpoint can be implemented as either an isochronous or bulk endpoint. The
standard isochronous or bulk endpoint descriptor provides pertinent information about how video
data streams are communicated to the video function. In addition, specific endpoint capabilities
and properties are reported.

3.10.1.1 Standard VS Isochronous Video Data Endpoint Descriptor

The standard VS isochronous video data endpoint descriptor is identical to the standard endpoint
descriptor defined in section 9.6.6 "Endpoint" of USB Specification Revision 2.0. D7 of the
bEndpointAddress field indicates whether the endpoint is a video source (D7 = 1) or a video
sink (D7 = 0). The bmAttributes field bits are set to reflect the isochronous type of the endpoint.
The synchronization type is indicated by D3..2 and must be set to Asynchronous. For further
details, refer to section 5.12.4.1 "Synchronization Type," of USB Specification Revision 2.0.

Revision 1.5 August 9, 2012 69

 USB Device Class Definition for Video Devices

Table 3-20 Standard VS Isochronous Video Data Endpoint Descriptor

Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 7
1 bDescriptorType 1 Constant ENDPOINT descriptor type
2 bEndpointAddress 1 Endpoint The address of the endpoint on the USB
device described by this descriptor. The
address is encoded as follows:
D7: Direction
0 = OUT endpoint
1 = IN endpoint
D6..4: Reserved, reset to zero
D3..0: The endpoint number, determined by
the designer
3 bmAttributes 1 Bitmap D3..2: Synchronization type
01 = Asynchronous
D1..0: Transfer type
01 = Isochronous
All other bits are reserved.
4 wMaxPacketSize 2 Number Maximum packet size this endpoint is
capable of sending or receiving when this
configuration is selected.
This is determined by the video bandwidth
constraints of the endpoint.
6 bInterval 1 Number Interval for polling endpoint for data
transfers.
This value is expressed as a period of frames
or microframes depending on device speed,
and must range from 1 to 16. The bInterval
value is used as the exponent for a 2bInterval-1
period.

3.10.1.2 Standard VS Bulk Video Data Endpoint Descriptor

The standard VS Bulk video data endpoint descriptor is identical to the standard endpoint
descriptor defined in section 9.6.6 "Endpoint" of USB Specification Revision 2.0. D7 of the
bEndpointAddress field indicates that this endpoint is a data source (D7 = 1) or a video sink
(D7 = 0). The bmAttributes field bits are set to reflect the bulk type of the endpoint.
Table 3-21 Standard VS Bulk Video Data Endpoint Descriptor
Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 7
1 bDescriptorType 1 Constant ENDPOINT descriptor type
2 bEndpointAddress 1 Endpoint The address of the endpoint on the USB

Revision 1.5 August 9, 2012 70

 USB Device Class Definition for Video Devices

device described by this descriptor. The

address is encoded as follows:
D7: Direction
0 = OUT endpoint
1 = IN endpoint
D6..4: Reserved, reset to zero
D3..0: The endpoint number, determined by
the designer
3 bmAttributes 1 Bitmap D1..0: Transfer type (set to 10 = Bulk)
All other bits are reserved.
4 wMaxPacketSize 2 Number Maximum packet size this endpoint is
capable of sending or receiving when this
configuration is selected.
6 bInterval 1 Number Interval for polling endpoint for data
transfers.
For high-speed bulk OUT endpoints, the
bInterval must specify the maximum NAK
rate of the endpoint. A value of 0 indicates
the endpoint never NAKs. Other values
indicate at most 1 NAK each bInterval
number of microframes. This value must be
in the range from 0 to 255.

3.10.2 VS Bulk Still Image Data Endpoint Descriptors

The standard bulk still image data endpoint descriptor provides pertinent information on how
still image data are communicated to the video function. In addition, specific endpoint
capabilities and properties are reported.

3.10.2.1 Standard VS Bulk Still Image Data Endpoint Descriptor

The standard VS Bulk still image data endpoint descriptor is identical to the standard endpoint
descriptor defined in section 9.6.6 "Endpoint" of USB Specification Revision 2.0. D7 of the
bEndpointAddress field indicates that this endpoint is a data source (D7 = 1). The
bmAttributes field bits are set to reflect the bulk type of the endpoint.

This optional endpoint is only implemented by the device if it supports method 3 of still image
capture. If implemented, it should always follow the Video Data endpoint (where available) in
descriptor ordering and endpoint addressing.
Table 3-22 Standard VS Bulk Still Image Data Endpoint Descriptor
Offset Field Size Value Description
0 bLength 1 Number Size of this descriptor, in bytes: 7

Revision 1.5 August 9, 2012 71

 USB Device Class Definition for Video Devices

1 bDescriptorType 1 Constant ENDPOINT descriptor type

2 bEndpointAddress 1 Endpoint The address of the endpoint on the USB
device described by this descriptor. The
address is encoded as follows:
D7: Direction (set to 1 = IN endpoint)
D6..4: Reserved, reset to zero
D3..0: The endpoint number, determined by
the designer
3 bmAttributes 1 Bitmap D1..0: Transfer type (set to 10 = Bulk)
All other bits are reserved.
4 wMaxPacketSize 2 Number Maximum packet size this endpoint is
capable of sending or receiving when this
configuration is selected.
6 bInterval 1 Number Not used, set to zero.

3.11 String Descriptors

The baseline requirement for devices in this class is for the device implementation to provide a
function name string descriptor in US English (LANGID = 0x0409). This will be referenced in
the iInterface field in the standard VideoControl interface descriptor. See section 3.7.1,
"Standard VC Interface Descriptor".

If the VideoControl interface is part of a Video Interface Collection, the iFunction field in the
IAD and the iInterface field in the Standard VC interface descriptor for this Video Interface
Collection must be equal. See section 3.5.

All other string descriptors are optional.

Since the device must implement the device name string descriptor, it must also support String
Descriptor Zero which contains the list of LANGID codes supported by the device. This
descriptor, as well as the layout of a standard UNICODE String Descriptor, is defined in section
9.6.7 "String" of the USB Specification Revision 2.0.

Revision 1.5 August 9, 2012 72

 USB Device Class Definition for Video Devices

4 Class-Specific Requests
Most class-specific requests are used to set and get video related Controls. These Controls fall
into two main groups: those that manipulate controls related to the video function, such as
brightness, exposure, selector position, etc. and those that influence data transfer over a video
data endpoint, such as the current frame rate.

 VideoControl Requests. Control of a video function is performed through the

manipulation of the attributes of individual Controls that are embedded in the Units and
Terminals of the video function. The class-specific VideoControl interface descriptor
contains a collection of Unit and Terminal descriptors, each indicating which Controls
are present in each entity. VideoControl requests are always directed to the single
VideoControl interface of the video function. The request contains enough information
(Unit ID, Control Selector) for the video function to route a specific request correctly.
 VideoStreaming Requests. Control of the class-specific behavior of a VideoStreaming
interface is performed through manipulation of Interface Controls. VideoStreaming
requests are directed to the interface where the Control resides.

Requests may be mandatory or optional and listed as such for every control. Where SET_CUR is
optional, its presence is determined via GET_INFO. If a video function does not support a
certain request, it must indicate this by stalling the control pipe when that request is issued to the
function.

4.1 Request Layout

The following paragraphs describe the general structure of the Set and Get requests. Subsequent
paragraphs detail the use of the Set/Get requests for the different request types.

4.1.1 Set Request

This request is used to set an attribute of a Control inside an entity of the video function.
Table 4-1 Set Request
bmRequestType bRequest wValue wIndex wLength Data
00100001 SET_CUR See Entity ID and Length of Parameter
following Interface. parameter block.
00100010 paragraphs. Endpoint. block.

The bmRequestType field specifies that this is a SET request (D7=0). It is a class-specific
request (D6..5=01), directed to either the VideoControl interface, or a VideoStreaming interface
of the video function (D4..0=00001), or the video data endpoint of a VideoStreaming interface
(D4..0=00010).

Revision 1.5 August 9, 2012 73

 USB Device Class Definition for Video Devices

The bRequest field contains a constant that identifies which attribute of the addressed Control is
to be modified. Possible attributes for a Control are:
Current setting attribute (SET_CUR)

If the addressed Control or entity does not support modification of a certain attribute, the control
pipe must indicate a stall when an attempt is made to modify that attribute. Only the CUR
attribute is supported for the Set request. For the list of Request constants, refer to section A.8,
"Video Class-Specific Request Codes"

The wValue field interpretation is qualified by the value in the wIndex field. Depending on what
entity is addressed, the layout of the wValue field changes. The following paragraphs describe
the contents of the wValue field for each entity separately. In most cases, the wValue field
contains the Control Selector (CS) in the high byte. It is used to address a particular Control
within entities that can contain multiple Controls. If the entity only contains a single Control,
there is no need to specify a Control Selector and the wValue field can be used to pass additional
parameters.

The wIndex field specifies the interface or endpoint to be addressed in the low byte, and the
entity ID or zero in the high byte. In case an interface is addressed, the virtual entity "interface"
can be addressed by specifying zero in the high byte. The values in wIndex must be appropriate
to the recipient. Only existing entities in the video function can be addressed, and only
appropriate interface or endpoint numbers may be used. If the request specifies an unknown or
non-entity ID or an unknown interface or endpoint number, the control pipe must indicate a stall.

The actual parameter(s) for the Set request are passed in the data stage of the control transfer.
The length of the parameter block is indicated in the wLength field of the request. The layout of
the parameter block is qualified by both the bRequest and wIndex fields. Refer to the following
sections for a detailed description of the parameter block layout for all possible entities.

4.1.2 Get Request

This request returns the attribute setting of a specific Control inside an entity of the video
function.
Table 4-2 Get Request
bmRequestType bRequest wValue wIndex wLength Data
10100001 GET_CUR See following Entity ID and Length of Parameter
GET_MIN paragraphs. Interface parameter block
10100010 GET_MAX Endpoint. block
GET_RES
GET_LEN
GET_INFO
GET_DEF

Revision 1.5 August 9, 2012 74

 USB Device Class Definition for Video Devices

The bmRequestType field specifies that this is a GET request (D7=1). It is a class-specific
request (D6..5=01), directed to either the VideoControl interface or a VideoStreaming interface
of the video function (D4..0=00001), or the video data endpoint of a VideoStreaming interface
(D4..0=00010).

The bRequest field contains a constant that identifies which attribute of the addressed Control or
entity is to be returned. Possible attributes for a Control are:
Current setting attribute (GET_CUR)
Minimum setting attribute (GET_MIN)
Maximum setting attribute (GET_MAX)
Default setting attribute (GET_DEF)
Resolution attribute (GET_RES)
Data length attribute (GET_LEN)
Information attribute (GET_INFO)

The GET_INFO request queries the capabilities and status of the specified control. When issuing
this request, the wLength field shall always be set to a value of 1 byte. The result returned is a
bit mask reporting the capabilities of the control. The bits are defined as:
Table 4-3 Defined Bits Containing Capabilities of the Control
Bit field Description Bit State
D0 1=Supports GET value requests Capability
D1 1=Supports SET value requests Capability
D2 1=Disabled due to automatic mode (under State
device control)
D3 1= Autoupdate Control (see section 2.4.2.2 Capability
"Status Interrupt Endpoint")
D4 1= Asynchronous Control (see sections Capability
2.4.2.2 "Status Interrupt Endpoint" and 2.4.4,
“Control Transfer and Request Processing”)
D5 1= Disabled due to incompatibility with State
Commit state.
D7..D6 Reserved (Set to 0) --

The two bits in GET_INFO that reflect the state of the control are D2 (Disabled due to
Automatic Mode) and D5 (Disabled due to incompatibility with Commit state). The other bits
are capability bits. Capability bits should not change when state bits change. For example, when
a control is set in Automatic Mode (D2 set), the bit D1 must not be updated in GET_INFO.
If a control is implemented such that D2 can be set, the device needs to have the capability of
sending Control Change Interrupts, thus D3 (Autoupdate Control) must be set. If a control is
implemented such that D5 can be set, the device should to have the capability of sending Control
Change Interrupts.

Revision 1.5 August 9, 2012 75

 USB Device Class Definition for Video Devices

If an Encoding Unit control is implemented such that the device may initiate a change in the
minimum and/or maximum setting attribute for that control, then the device should have the
capability of sending Control Change Interrupts to notify the host of the new GET_MIN and/or
GET_MAX settings, thus the D3 (Autoupdate Control) must be set.

The device indicates hardware default values for Unit, Terminal and Interface Controls through
their GET_DEF values. These values may be used by the host to restore a control to its default
setting.

If the addressed Control or entity does not support readout of a certain attribute, the control pipe
must indicate a stall when an attempt is made to read that attribute. For the list of Request
constants, refer to section A.8, "Video Class-Specific Request Codes".

The wValue field interpretation is qualified by the value in the wIndex field. Depending on what
entity is addressed, the layout of the wValue field changes. The following paragraphs describe
the contents of the wValue field for each entity separately. In most cases, the wValue field
contains the Control Selector (CS) in the high byte. It is used to address a particular Control
within entities that can contain multiple Controls. If the entity only contains a single Control,
there is no need to specify a Control Selector and the wValue field can be used to pass additional
parameters.

The wIndex field specifies the interface or endpoint to be addressed in the low byte, and the
entity ID or zero in the high byte. In case an interface is addressed, the virtual entity "interface"
can be addressed by specifying zero in the high byte. The values in wIndex must be appropriate
to the recipient. Only existing entities in the video function can be addressed, and only
appropriate interface or endpoint numbers may be used. If the request specifies an unknown or
non-entity ID, or an unknown interface or endpoint number, the control pipe must indicate a
stall.

The actual parameter(s) for the Get request are returned in the data stage of the control transfer.
The length of the parameter block to return is indicated in the wLength field of the request. If
the parameter block is longer than is indicated in the wLength field, only the initial bytes of the
parameter block are returned. If the parameter block is shorter than is indicated in the wLength
field, the device indicates the end of the control transfer by sending a short packet when further
data is requested. The layout of the parameter block is qualified by both the bRequest and
wIndex fields. Refer to the following sections for a detailed description of the parameter block
layout for all possible entities.

4.2 VideoControl Requests

The following sections describe the possible requests that can be used to manipulate the video
Controls that a video function exposes through its VideoControl interface and Units contained
within it. The same layout of the parameter blocks is used for both the Set and Get requests.

Revision 1.5 August 9, 2012 76

 USB Device Class Definition for Video Devices

Each of the following control definitions specifies whether requests are mandatory or optional
for that control. Any implemented request must comply with the definition for that control. The
device manufacturer is free to implement any other requests, but the definition of those
unspecified requests shall be ignored by host implementations, with the exception of the
GET_LEN request. If the GET_LEN request is implemented, the host software will use the result
to determine the correct buffer length for Set and Get requests.

4.2.1 Interface Control Requests

These requests are used to set or read an attribute of an interface Control inside the VideoControl
interface of the video function.
Table 4-4 Interface Control Requests
bmRequestType bRequest wValue wIndex wLength Data
00100001 SET_CUR CS Zero and Length of Parameter
Interface. parameter block.
10100001 GET_CUR block.
GET_MIN
GET_MAX
GET_RES
GET_INFO

The bRequest field indicates which attribute the request is manipulating. The MIN, MAX, and
RES attributes are not supported for the Set request.

The wValue field specifies the Control Selector (CS) in the high byte, and the low byte must be
set to zero. The Control Selector indicates which type of Control this request is manipulating. If
the request specifies an unknown CS to that endpoint, the control pipe must indicate a stall.

4.2.1.1 Power Mode Control

This control sets the device power mode. Power modes are defined in the following table.
Table 4-5 Power Mode Control
Device power mode Description
Full power mode Device operates at full functionality in this mode. For example,
the device can stream video data via USB, and can execute all
requests that are supported by the device. This mode is
mandatory, even if the device doesn’t support VIDEO POWER
MODE CONTROL.

Revision 1.5 August 9, 2012 77

 USB Device Class Definition for Video Devices

Vendor-dependent Device operates in low power mode. In this mode, the device
power mode continues to operate, although not at full functionality.
For example, as the result of setting the device to this power
mode, the device will stop the Zoom function. To avoid
confusing the user, the device should issue an interrupt
(GET_INFO) to notify the user that the Zoom function is
disabled.
In this mode, the device can stream video data, the functionality
of USB is not affected, and the device can execute all requests
that it supports.
This mode is optional.

The power mode that is supported by the device must be passed to the host, as well as the power
source, since if the device is working with battery power, the host can change the device power
mode to “vendor-dependent power mode” to reduce power consumption.
Information regarding power modes and power sources is communicated through the following
bit fields. D7..D5 indicates which power source is currently used in the device. The D4 indicates
that the device supports “vendor-dependent power mode”. Bits D7..D4 are set by the device and
are read-only. The host can change the device power mode by setting a combination of D3..D0.

The host can update the power mode during video streaming.

The D3..D0 value of 0000B indicates that the device is in, or should transition to, full power
mode. The D3..D0 value of 0001B indicates that the device is in, or should transition to, vendor-
dependent power mode.
The host must specify D3..D0 only when the power mode is required to switch, and the other
fields must be set to 0.
Table 4-6 Device Power Mode
Control selector VC_VIDEO_POWER_MODE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO
wLength 1
Offset Field Size Value Description
0 bDevicePowerMode 1 Bitmap

Bit Description R W
D3..0 Power Mode setting o o
0000B:Full power mode
0001B:device dependent
power mode (opt.)
All other bits are reserved.
D4 Device dependent power o x
mode supported.

Revision 1.5 August 9, 2012 78

 USB Device Class Definition for Video Devices

D5 Device uses power o x

supplied by USB.
D6 Device uses power o x
supplied by Battery.
D7 Device uses power o x
supplied by A.C.

4.2.1.2 Request Error Code Control

This read-only control indicates the status of each host-initiated request to a Terminal, Unit,
interface or endpoint of the video function. If the device is unable to fulfill the request, it will
indicate a stall on the control pipe and update this control with the appropriate code to indicate
the cause. This control will be reset to 0 (No error) upon the successful completion of any control
request (including requests to this control). The table below specifies the bRequestErrorCode
error codes that the device must return from a VC_REQUEST_ERROR_CODE_CONTROL
request. Asynchronous control requests are a special case, where the initial request will update
this control, but the final result is delivered via the Status Interrupt Endpoint (see sections
2.4.2.2, "Status Interrupt Endpoint" and 2.4.4, "Control Transfer and Request Processing").
Table 4-7 Request Error Code Control
Control Selector VC_REQUEST_ERROR_CODE_CONTROL
Mandatory Requests GET_CUR, GET_INFO
wLength 1
Offset Field Size Value Description
0 bRequestErrorCode 1 Number 0x00: No error
0x01: Not ready
0x02: Wrong state
0x03: Power
0x04: Out of range
0x05: Invalid unit
0x06: Invalid control
0x07: Invalid Request
0x08: Invalid value within range
0x09-0xFE: Reserved for future use
0xFF: Unknown

No error: The request succeeded.

Not ready: The device has not completed a previous operation. The device will recover from this
state as soon as the previous operation has completed.

Wrong State: The device is in a state that disallows the specific request. The device will remain
in this state until a specific action from the host or the user is completed.

Revision 1.5 August 9, 2012 79

 USB Device Class Definition for Video Devices

Power: The actual Power Mode of the device is not sufficient to complete the Request.

Out of Range: Result of a SET_CUR Request when attempting to set a value outside of the MIN
and MAX range, or a value that does not satisfy the constraint on resolution (see section 4.2.2,
“Unit and Terminal Control Requests”).

Invalid Unit: The Unit ID addressed in this Request is not assigned.

Invalid Control: The Control addressed by this Request is not supported.

Invalid Request: This Request is not supported by the Control.

Invalid value with range: Results of a SET_CUR Request when attempting to set a value that is
inside the MIN and MAX range but is not supported.

4.2.2 Unit and Terminal Control Requests

The requests below are used to set or read an attribute of a Control inside of a Unit or Terminal
of the video function.
Table 4-8 Unit and Terminal Control Requests

bmRequestType bRequest wValue wIndex wLength Data

00100001 SET_CUR CS Unit or Length of Parameter
Terminal ID parameter block
10100001 GET_CUR and block
GET_MIN Interface
GET_MAX
GET_RES
GET_INFO
GET_DEF
00100001 SET_CUR_ALL Set to 0 Unit or
Terminal ID
10100001 GET_CUR_ALL
GET_MIN_ALL
GET_MAX_ALL
GET_RES_ALL
GET_DEF_ALL

The bRequest field indicates which attribute the request is manipulating. The MIN, MAX and
RES attributes are not supported for the Set request.

The wValue field specifies the Control Selector (CS) in the high byte, and zero in the low byte.
The Control Selector indicates which type of Control this request is manipulating. When

Revision 1.5 August 9, 2012 80

 USB Device Class Definition for Video Devices

processing all Controls as part of a batch request (GET_###_ALL), wValue is not needed and
must be set to 0. If the request specifies an unknown or unsupported CS to that Unit or Terminal,
the control pipe must indicate a protocol STALL.

The value of wLength must be calculated as follows. Use wIndex to determine the Unit or
Terminal of interest. For that Unit or Terminal, establish which Controls are supported using the
bmControls field of the associated Unit or Terminal Descriptor. wLength is the sum of the
length of all supported Controls for the target Unit or Terminal. The Data must be ordered
according to the order of the Controls listed in the bmControls field of the target Unit or
Terminal descriptor. If the Unit or Terminal supports batch requests, then each Control in the
Unit or Terminal must contribute to the Data field, even if it does not support the associated
single operation request.

If a Control supports GET_MIN, GET_MAX and GET_RES requests, the values of MAX, MIN
and RES shall be constrained such that (MAX-MIN)/RES is an integral number. Furthermore, the
CUR value (returned by GET_CUR, or set via SET_CUR) shall be constrained such that (CUR-
MIN)/RES is an integral number. The device shall indicate protocol STALL and update the
Request Error Code Control with 0x04 “Out of Range” if an invalid CUR value is provided in a
SET_CUR operation (see section 2.4.4, “Control Transfer and Request Processing”).

There are special Terminal types (such as the Camera Terminal and Media Transport Terminal)
that have type-specific Terminal Controls defined. The controls for the Media Transport
Terminal are defined in a companion specification (see the USB Device Class Definition for
Video Media Transport Terminal specification). The controls for the Camera Terminal are
defined in the following sections.

As this specification evolves, new controls in the Camera Terminal, Processing Unit, and
Encoding Unit are added to the list of associated Control Selectors at the end (Tables A-12
through A-14). However, in the sections below, the description of the functionality is placed next
to controls with associated functionality.

4.2.2.1 Camera Terminal Control Requests

The following paragraphs present a detailed description of all possible Controls a Camera
Terminal can incorporate. For each Control, the layout of the parameter block together with the
appropriate Control Selector is listed for all forms of the Get/Set Camera Terminal Control
request. All values are interpreted as absolute (fixed-origin), and not relative unless otherwise
specified. They are also assumed to be unsigned unless otherwise specified.

4.2.2.1.1 Scanning Mode Control

The Scanning Mode Control setting is used to control the scanning mode of the camera sensor. A
value of 0 indicates that the interlace mode is enabled, and a value of 1 indicates that the
progressive or the non-interlace mode is enabled.

Revision 1.5 August 9, 2012 81

 USB Device Class Definition for Video Devices

Table 4-9 Scanning Mode Control

Control Selector CT_SCANNING_MODE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO
wLength 1
Offset Field Size Value Description
0 bScanningMode 1 Boolean The setting for the attribute of the
addressed Scanning Mode Control:
0: Interlaced
1: Progressive

4.2.2.1.2 Auto-Exposure Mode Control

The Auto-Exposure Mode Control determines whether the device will provide automatic
adjustment of the Exposure Time and Iris controls. Attempts to programmatically set the auto-
adjusted controls shall result in a protocol STALL and an error code of bRequestErrorCode =
“Wrong state”. A GET_RES request issued to this control will return a bitmap of the modes
supported by this control. A valid request to this control would have only one bit set (a single
mode selected). This control must accept the GET_DEF request and return its default value.
Table 4-10 Auto-Exposure Mode Control
Control Selector CT_AE_MODE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_RES, GET_INFO,
GET_DEF
wLength 1
Offset Field Size Value Description
0 bAutoExposureMode 1 Bitmap The setting for the attribute of the
addressed Auto-Exposure Mode
Control:
D0: Manual Mode – manual
Exposure Time, manual Iris
D1: Auto Mode – auto Exposure
Time, auto Iris
D2: Shutter Priority Mode –
manual Exposure Time, auto Iris
D3: Aperture Priority Mode – auto
Exposure Time, manual Iris
D4..D7: Reserved, set to zero.

4.2.2.1.3 Auto-Exposure Priority Control

The Auto-Exposure Priority Control is used to specify constraints on the Exposure Time Control
when the Auto-Exposure Mode Control is set to Auto Mode or Shutter Priority Mode. A value of
zero indicates that the frame rate must remain constant. A value of 1 indicates that the frame rate
may be dynamically varied by the device. The default value is zero (0).

Revision 1.5 August 9, 2012 82

 USB Device Class Definition for Video Devices

Table 4-11 Auto-Exposure Priority Control

Control Selector CT_AE_PRIORITY_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO
wLength 1
Offset Field Size Value Description
0 bAutoExposurePriority 1 Number The setting for the attribute of the
addressed AutoExposure Priority
control.

4.2.2.1.4 Exposure Time (Absolute) Control

The Exposure Time (Absolute) Control is used to specify the length of exposure. This value is
expressed in 100µs units, where 1 is 1/10,000th of a second, 10,000 is 1 second, and 100,000 is
10 seconds. A value of zero (0) is undefined. Note that the manual exposure control is further
limited by the frame interval, which always has higher precedence. If the frame interval is
changed to a value below the current value of the Exposure Control, the Exposure Control value
will automatically be changed. The default Exposure Control value will be the current frame
interval until an explicit exposure value is chosen. When the Auto-Exposure Mode control is in
Auto mode or Aperture Priority mode attempts to programmatically set this control shall result in
a protocol STALL and an error code of bRequestErrorCode = “Wrong state”. This control
must accept the GET_DEF request and return its default value.
Table 4-12 Exposure Time (Absolute) Control
Control Selector CT_EXPOSURE_TIME_ABSOLUTE_CONTROL
Mandatory Requests GET_CUR, GET_MIN, GET_MAX, GET_RES,
GET_INFO, GET_DEF
Optional Requests SET_CUR
wLength 4
Offset Field Size Value Description
0 dwExposureTimeAbsolute 4 Number The setting for the attribute of the
addressed Exposure Time
(Absolute) Control:
0: Reserved
1: 0.0001 sec
…
100000: 10 sec
…

4.2.2.1.5 Exposure Time (Relative) Control

The Exposure Time (Relative) Control is used to specify the electronic shutter speed. This value
is expressed in number of steps of exposure time that is incremented or decremented. A value of
one (1) indicates that the exposure time is incremented one step further, and a value 0xFF
indicates that the exposure time is decremented one step further. This step is implementation

Revision 1.5 August 9, 2012 83

 USB Device Class Definition for Video Devices

specific. A value of zero (0) indicates that the exposure time is set to the default value for
implementation. The default values are implementation specific. . When the Auto-Exposure
Mode control is in Auto mode or Aperture Priority mode attempts to programmatically set this
control shall result in a protocol STALL and an error code of bRequestErrorCode = “Wrong
state”.

If both Relative and Absolute Controls are supported, a SET_CUR to the Relative Control with a
value other than 0x00 shall result in a Control Change interrupt for the Absolute Control (see
section 2.4.2.2, “Status Interrupt Endpoint”).
Table 4-13 Exposure Time (Relative) Control
Control Selector CT_EXPOSURE_TIME_RELATIVE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO
wLength 1
Offset Field Size Value Description
0 bExposureTimeRelative 1 Signed The setting for the attribute of the
Number addressed Exposure Time (Relative)
Control:
0: default
1: incremented by 1 step
0xFF: decremented by 1 step

4.2.2.1.6 Focus (Absolute) Control

The Focus (Absolute) Control is used to specify the distance to the optimally focused target. This
value is expressed in millimeters. The default value is implementation-specific. This control
must accept the GET_DEF request and return its default value. When the Auto-Focus Mode
control is enabled, attempts to programmatically set this control shall result in a protocol STALL
and an error code of bRequestErrorCode = “Wrong state”.
Table 4-14 Focus (Absolute) Control
Control Selector CT_FOCUS_ABSOLUTE_CONTROL
Mandatory Requests GET_CUR, GET_MIN, GET_MAX, GET_RES, GET_INFO,
GET_DEF
Optional Requests SET_CUR
wLength 2
Offset Field Size Value Description
0 wFocusAbsolute 2 Number The setting for the attribute of the
addressed Focus (Absolute) Control.

4.2.2.1.7 Focus (Relative) Control

The Focus (Relative) Control is used to move the focus lens group to specify the distance to the
optimally focused target.

Revision 1.5 August 9, 2012 84

 USB Device Class Definition for Video Devices

The bFocusRelative field indicates whether the focus lens group is stopped or is moving for
near or for infinity direction. A value of 1 indicates that the focus lens group is moved for near
direction. A value of 0 indicates that the focus lens group is stopped. And a value of 0xFF
indicates that the lens group is moved for infinity direction. The GET_MIN, GET_MAX,
GET_RES and GET_DEF requests will return zero for this field.

The bSpeed field indicates the speed of the lens group movement. A low number indicates a
slow speed and a high number indicates a high speed. The GET_MIN, GET_MAX and
GET_RES requests are used to retrieve the range and resolution for this field. The GET_DEF
request is used to retrieve the default value for this field. If the control does not support speed
control, it will return the value 1 in this field for all these requests.

If both Relative and Absolute Controls are supported, a SET_CUR to the Relative Control with a
value other than 0x00 shall result in a Control Change interrupt for the Absolute Control at the
end of the movement (see section 2.4.2.2, “Status Interrupt Endpoint”). The end of movement
can be due to physical device limits, or due to an explicit request by the host to stop the
movement. If the end of movement is due to physical device limits (such as a limit in range of
motion), a Control Change interrupt shall be generated for this Relative Control. If there is no
limit in range of motion, a Control Change interrupt is not required.
When the Auto-Focus Mode control is enabled, attempts to programmatically set this control
shall result in a protocol STALL and an error code of bRequestErrorCode = “Wrong state”.

Table 4-15 Focus (Relative) Control

Control Selector CT_FOCUS_RELATIVE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO, GET_DEF, GET_MIN,
GET_MAX, GET_RES
wLength 2
Offset Field Size Value Description
0 bFocusRelative 1 Signed The setting for the attribute of the
number addressed Focus (Relative) Control:
0: Stop
1: Focus Near direction
0xFF: Focus Infinite direction
1 bSpeed 1 Number Speed for the control change

4.2.2.1.8 Focus, Simple Range

The Focus, Simple Control setting determines the absolute focus of the lens at a very granular
level: Macro, People, and Scene. This control may only be used when the camera is in manual or
auto focus mode. This control must accept the GET_DEF request and return its default value.
When the Auto-Focus Mode control is enabled, attempts to programmatically set this control
shall result in a protocol STALL and an error code of bRequestErrorCode = “Wrong state”.

Revision 1.5 August 9, 2012 85

 USB Device Class Definition for Video Devices

Table 4-16 Focus, Simple Range Control

Control Selector CT_FOCUS_SIMPLE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO, GET_DEF
wLength 1
Offset Field Size Value Description
0 bFocus 1 Number 0x00 – full range
0x01 – macro. Less than 0.3
meters.
0x02 – people. 0.3m to 3m
0x03 – scene. 3m to infinity
0x04 to 0xFF - reserved

4.2.2.1.9 Focus, Auto Control

The Focus, Auto Control setting determines whether the device will provide automatic
adjustment of the Focus Absolute and/or Relative Controls. A value of 1 indicates that automatic
adjustment is enabled. Attempts to programmatically set the related controls shall result in a
protocol STALL and an error code bRequestErrorCode = “Wrong state”. This control must
accept the GET_DEF request and return its default value.
Table 4-17 Focus, Auto Control
Control Selector CT_FOCUS_AUTO_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO, GET_DEF
wLength 1
Offset Field Size Value Description
0 bFocusAuto 1 Boolean The setting for the attribute of the
addressed Focus Auto control.

4.2.2.1.10 Iris (Absolute) Control

The Iris (Absolute) Control is used to specify the camera's aperture setting. This value is
expressed in units of fstop * 100. The default value is implementation-specific.
This control will not accept SET requests when the Auto-Exposure Mode control is in Auto
mode or Shutter Priority mode, and the control pipe shall indicate a stall in this case. This control
must accept the GET_DEF request and return its default value.
Table 4-18 Iris (Absolute) Control
Control Selector CT_IRIS_ABSOLUTE_CONTROL
Mandatory Requests GET_CUR, GET_MIN, GET_MAX, GET_RES, GET_INFO,
GET_DEF
Optional Requests SET_CUR
wLength 2
Offset Field Size Value Description
0 wIrisAbsolute 2 Number The setting for the attribute of the

Revision 1.5 August 9, 2012 86

 USB Device Class Definition for Video Devices

addressed Iris (Absolute) Control.

4.2.2.1.11 Iris (Relative) Control

The Iris (Relative) Control is used to specify the camera's aperture setting. This value is a signed
integer and indicates the number of steps to open or close the iris. A value of 1 indicates that the
iris is opened 1 step further. A value of 0xFF indicates that the iris is closed 1 step further. This
step of iris is implementation specific. A value of zero (0) indicates that the iris is set to the
default value for the implementation. The default value is implementation specific. This control
will not accept SET requests when the Auto-Exposure Mode control is in Auto mode or Shutter
Priority mode, and the control pipe shall indicate a stall in this case.

If both Relative and Absolute Controls are supported, a SET_CUR to the Relative Control with a
value other than 0x00 shall result in a Control Change interrupt for the Absolute Control (see
section 2.4.2.2, “Status Interrupt Endpoint”).
Table 4-19 Iris (Relative) Control
Control Selector CT_IRIS_RELATIVE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO
wLength 1
Offset Field Size Value Description
0 bIrisRelative 1 Number The setting for the attribute of the
addressed Iris (Relative) Control:
0: Default
1: Iris is opened by 1 step.
0xFF: Iris is closed by 1 step.

4.2.2.1.12 Zoom (Absolute) Control

The Zoom (Absolute) Control is used to specify or determine the Objective lens focal length.
This control is used in combination with the wObjectiveFocalLengthMin and
wObjectiveFocalLengthMax fields in the Camera Terminal descriptor to describe and control
the Objective lens focal length of the device. The MIN and MAX values are sufficient to imply
the resolution, so the RES value must always be 1. The MIN, MAX and default values are
implementation dependent. This control must accept the GET_DEF request and return its default
value.
Table 4-20 Zoom (Absolute) Control
Control Selector CT_ZOOM_ABSOLUTE_CONTROL
Mandatory Requests GET_CUR, GET_MIN, GET_MAX, GET_RES,
GET_INFO, GET_DEF
Optional Requests SET_CUR
wLength 2
Offset Field Size Value Description

Revision 1.5 August 9, 2012 87

 USB Device Class Definition for Video Devices

0 wObjectiveFocalLength 2 Number The value of Zcur

4.2.2.1.13 Zoom (Relative) Control

The Zoom (Relative) Control is used to specify the zoom focal length relatively as powered
zoom.

The bZoom field indicates whether the zoom lens group is stopped or the direction of the zoom
lens. A value of 1 indicates that the zoom lens is moved towards the telephoto direction. A value
of zero indicates that the zoom lens is stopped, and a value of 0xFF indicates that the zoom lens
is moved towards the wide-angle direction. The GET_MIN, GET_MAX, GET_RES and
GET_DEF requests will return zero for this field.

The bDigitalZoom field specifies whether digital zoom is enabled or disabled. If the device only
supports digital zoom, this field would be ignored. The GET_DEF request will return the default
value for this field. The GET_MIN, GET_MAX and GET_RES requests will return zero for this
field.

The bSpeed field indicates the speed of the control change. A low number indicates a slow speed
and a high number indicates a higher speed. The GET_MIN, GET_MAX and GET_RES
requests are used to retrieve the range and resolution for this field. The GET_DEF request is
used to retrieve the default value for this field. If the control does not support speed control, it
will return the value 1 in this field for all these requests.

If both Relative and Absolute Controls are supported, a SET_CUR to the Relative Control with a
value other than 0x00 shall result in a Control Change interrupt for the Absolute Control at the
end of the movement (see section 2.4.2.2, “Status Interrupt Endpoint”). The end of movement
can be due to physical device limits, or due to an explicit request by the host to stop the
movement. If the end of movement is due to physical device limits (such as a limit in range of
motion), a Control Change interrupt shall be generated for this Relative Control.
Table 4-21 Zoom (Relative) Control
Control Selector CT_ZOOM_RELATIVE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO, GET_DEF, GET_MIN,
GET_MAX, GET_RES
wLength 3
Offset Field Size Value Description
0 bZoom 1 Signed The setting for the attribute of the
number addressed Zoom Control:
0: Stop
1: moving to telephoto direction
0xFF: moving to wide-angle direction
1 bDigitalZoom 1 Boolean 0: Digital Zoom OFF
1: Digital Zoom On

Revision 1.5 August 9, 2012 88

 USB Device Class Definition for Video Devices

2 bSpeed 1 Number Speed for the control change

4.2.2.1.14 PanTilt (Absolute) Control

The PanTilt (Absolute) Control is used to specify the pan and tilt settings.
The dwPanAbsolute is used to specify the pan setting in arc second units. 1 arc second is 1/3600
of a degree. Values range from –180*3600 arc second to +180*3600 arc second, or a subset
thereof, with the default set to zero. Positive values are clockwise from the origin (the camera
rotates clockwise when viewed from above), and negative values are counterclockwise from the
origin. This control must accept the GET_DEF request and return its default value.
The dwTiltAbsolute Control is used to specify the tilt setting in arc second units. 1 arc second is
1/3600 of a degree. Values range from –180*3600 arc second to +180*3600 arc second, or a
subset thereof, with the default set to zero. Positive values point the imaging plane up, and
negative values point the imaging plane down. This control must accept the GET_DEF request
and return its default value.
Table 4-22 PanTilt (Absolute) Control
Control Selector CT_PANTILT_ABSOLUTE_CONTROL
Mandatory Requests GET_CUR, GET_MIN, GET_MAX, GET_RES, GET_INFO,
GET_DEF
Optional Requests SET_CUR
wLength 8
Offset Field Size Value Description
0 dwPanAbsolute 4 Signed The setting for the attribute of the
number addressed Pan (Absolute) Control.
4 dwTiltAbsolute 4 Signed The setting for the attribute of the
number addressed Tilt (Absolute) Control.

4.2.2.1.15 PanTilt (Relative) Control

The PanTilt (Relative) Control is used to specify the pan and tilt direction to move.

The bPanRelative field is used to specify the pan direction to move. A value of 0 indicates to
stop the pan, a value of 1 indicates to start moving clockwise direction, and a value of 0xFF
indicates to start moving counterclockwise direction. The GET_DEF, GET_MIN, GET_MAX
and GET_RES requests will return zero for this field.

The bPanSpeed field is used to specify the speed of the movement for the Pan direction. A low
number indicates a slow speed and a high number indicates a higher speed. The GET_MIN,
GET_MAX and GET_RES requests are used to retrieve the range and resolution for this field.
The GET_DEF request is used to retrieve the default value for this field. If the control does not
support speed control for the Pan control, it will return the value 1 in this field for all these
requests.

Revision 1.5 August 9, 2012 89

 USB Device Class Definition for Video Devices

The bTiltRelative field is used to specify the tilt direction to move. A value of zero indicates to
stop the tilt, a value of 1 indicates that the camera point the imaging plane up, and a value of
0xFF indicates that the camera point the imaging plane down. The GET_DEF, GET_MIN,
GET_MAX and GET_RES requests will return zero for this field.

The bTiltSpeed field is used to specify the speed of the movement for the Tilt direction. A low
number indicates a slow speed and a high number indicates a higher speed. The GET_MIN,
GET_MAX and GET_RES requests are used to retrieve the range and resolution for this field.
The GET_DEF request is used to retrieve the default value for this field. If the control does not
support speed control for the Tilt control, it will return the value 1 in this field for all these
requests.

If both Relative and Absolute Controls are supported, a SET_CUR to the Relative Control with a
value other than 0x00 shall result in a Control Change interrupt for the Absolute Control at the
end of the movement (see section 2.4.2.2, “Status Interrupt Endpoint”). The end of movement
can be due to physical device limits, or due to an explicit request by the host to stop the
movement. If the end of movement is due to physical device limits (such as a limit in range of
motion), a Control Change interrupt shall be generated for this Relative Control. If there is no
limit in range of motion, a Control Change interrupt is not required.
Table 4-23 PanTilt (Relative) Control
Control Selector CT_PANTILT_RELATIVE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO, GET_DEF, GET_MIN,
GET_MAX, GET_RES
wLength 4
Offset Field Size Value Description
0 bPanRelative 1 Signed The setting for the attribute of the
number addressed Pan (Relative) Control:
0: Stop
1: moving to clockwise direction
0xFF: moving to counter clockwise
direction
1 bPanSpeed 1 Number Speed of the Pan movement
2 bTiltRelative 1 Signed The setting for the attribute of the
number addressed Tilt (Relative) Control:
0: Stop
1: point the imaging plane up
0xFF: point the imaging plane down
3 bTiltSpeed 1 Number Speed for the Tilt movement

Revision 1.5 August 9, 2012 90

 USB Device Class Definition for Video Devices

4.2.2.1.16 Roll (Absolute) Control

The Roll (Absolute) Control is used to specify the roll setting in degrees. Values range from –
180 to +180, or a subset thereof, with the default being set to zero. Positive values cause a
clockwise rotation of the camera along the image viewing axis, and negative values cause a
counterclockwise rotation of the camera. This control must accept the GET_DEF request and
return its default value.
Table 4-24 Roll (Absolute) Control
Control Selector CT_ROLL_ABSOLUTE_CONTROL
Mandatory Requests GET_CUR, GET_MIN, GET_MAX, GET_RES, GET_INFO,
GET_DEF
Optional Requests SET_CUR
wLength 2
Offset Field Size Value Description
0 wRollAbsolute 2 Signed The setting for the attribute of the
number addressed Roll (Absolute) Control.

4.2.2.1.17 Roll (Relative) Control

The Roll (Relative) Control is used to specify the roll direction to move.

The bRollRelative field is used to specify the roll direction to move. A value of 0 indicates to
stop the roll, a value of 1 indicates to start moving in a clockwise rotation of the camera along
the image viewing axis, and a value of 0xFF indicates to start moving in a counterclockwise
direction. The GET_DEF, GET_MIN, GET_MAX and GET_RES requests will return zero for
this field.

The bSpeed is used to specify the speed of the roll movement. A low number indicates a slow
speed and a high number indicates a higher speed. The GET_MIN, GET_MAX and GET_RES
requests are used to retrieve the range and resolution for this field. The GET_DEF request is
used to retrieve the default value for this field. If the control does not support speed control, it
will return the value 1 in this field for all these requests.

If both Relative and Absolute Controls are supported, a SET_CUR to the Relative Control with a
value other than 0x00 shall result in a Control Change interrupt for the Absolute Control at the
end of the movement (see section 2.4.2.2, “Status Interrupt Endpoint”). The end of movement
can be due to physical device limits, or due to an explicit request by the host to stop the
movement. If the end of movement is due to physical device limits (such as a limit in range of
motion), a Control Change interrupt shall be generated for this Relative Control. If there is no
limit in range of motion, a Control Change interrupt is not required.
Table 4-25 Roll (Relative) Control
Control Selector CT_ROLL_RELATIVE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO, GET_DEF, GET_MIN,

Revision 1.5 August 9, 2012 91

 USB Device Class Definition for Video Devices

GET_MAX, GET_RES
wLength 2
Offset Field Size Value Description
0 bRollRelative 1 Signed The setting for the attribute of the
number addressed Roll (Relative) Control:
0: Stop
1: moving clockwise rotation
0xFF: moving counter clockwise rotation
1 bSpeed 1 Number Speed for the Roll movement

4.2.2.1.18 Privacy Control

The Privacy Control setting is used to prevent video from being acquired by the camera sensor.
A value of 0 indicates that the camera sensor is able to capture video images, and a value of 1
indicates that the camera sensor is prevented from capturing video images.
This control shall be reported as an AutoUpdate control.
Table 4-26 Privacy Shutter Control
Control Selector CT_PRIVACY_CONTROL
Mandatory Requests GET_CUR, GET_INFO
Optional Requests SET_CUR
wLength 1
Offset Field Size Value Description
0 bPrivacy 1 Boolean The setting for the attribute of the
addressed Privacy Control:
0: Open
1: Close

4.2.2.1.19 Digital Window Control

The windowing API is based on “pixel” coordinates where each row and column of pixels on the
sensor can be referred to by integers between zero - (height-1) and zero- (width-1). The point at
0,0 is the top, left of the coordinate system and (height-1), (width-1) is the bottom, right of the
coordinates system.
Table 4- 27 Digital Window Control
Control Selector CT_DIGITAL_WINDOW_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_DEF
wLength 12
Offse Field Size Value Description
t
0 wWindow_Top 2 Number Top coordinate
In global sensor coordinates (pixels)
2 wWindow_Left 2 Number Left coordinate
In global sensor coordinates (pixels)

Revision 1.5 August 9, 2012 92

 USB Device Class Definition for Video Devices

4 wWindow_Bottom 2 Number Bottom coordinate

In global sensor coordinates (pixels)
6 wWindow_Right 2 Number Right coordinate
In global sensor coordinates (pixels)
8 wNumSteps 2 Number Number of steps to move from current
window to window specified by rectangle
above.
10 bmNumStepsUnits 2 Bitmap Units of number of steps specified above.
D0: video frames
D1: milliseconds
D2-D15: Reserved

CT_DIGITAL_WINDOW_CONTROL is used to specify a target window for viewing, and the

number of steps to take in moving from the current window to the new window the control
specifies. To prevent invalid windows being specified:
wWindow_Bottom ≥ wWindow_Top
and
wWindow_Right ≥ wWindow_Left

GET_MAX should return the sensor size as well as maximum number of supported steps in the
units indicated by bmNumStepsUnits. If the bmNumStepsUnits has not been set, the default
value should be used. GET_CUR returns the current coordinates of the digitial window used for
capture. If the device is moving between settings (e.g. wNumSteps > 1), GET_CUR references
the digital window of the current step.

4.2.2.1.20 Digital Region of Interest (ROI) Control

The rectangle specified by CT_REGION_OF_INTEREST_CONTROL will be in global sensor
coordinates. The units are in Pixels and independent of the field of view. They are not impacted
by any cropping or scaling that is currently being used. The ROI must be within the current
Digital Window as specified by the CT_WINDOW control.

The bmAutoControls bitmask determines which, if any, on board features should track to the
region of interest. To detect if a device supports a particular Auto Control, use GET_MAX
which returns a mask indicating all supported Auto Controls.

Table 4- 28 Region of Interest Control

Control Selector CT_REGION_OF_INTEREST_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_DEF
wLength 10
Offset Field Size Value Description
0 wROI_Top 2 Number Top coordinate
In global sensor coordinates (pixels)
2 wROI_Left 2 Number Left coordinate

Revision 1.5 August 9, 2012 93

 USB Device Class Definition for Video Devices

In global sensor coordinates (pixels)

4 wROI_Bottom 2 Number Bottom coordinate
In global sensor coordinates (pixels)
6 wROI_Right 2 Number Right coordinate
In global sensor coordinates (pixels)
8 bmAutoControls 2 Bitmap D0: Auto Exposure
D1: Auto Iris
D2: Auto White Balance
D3: Auto Focus
D4: Auto Face Detect
D5: Auto Detect and Track
D6: Image Stabilization
D7: Higher Quality
D8 – D15: Reserved, set to zero

GET_MAX shall return the current Window as specified by

CT_DIGITAL_WINDOW_CONTROL.

GET_CUR returns the current Region of Interest (RoI) being employed by the device. This RoI
should be the same as specified in most recent SET_CUR except in the case where the ‘Auto
Detect and Track’ and/or ‘Image Stabilization’ bit have been set.

4.2.2.2 Selector Unit Control Requests

These requests are used to set or read an attribute of a Selector Control inside a Selector Unit of
the video function.

A Selector Unit represents a video stream source selector. The valid range for the CUR, MIN,
and MAX attributes is from one up to the number of Input Pins of the Selector Unit. This value
can be found in the bNrInPins field of the Selector Unit descriptor. The RES attribute can only
have a value of one.
Table 4-29 Selector Unit Control Requests
Control Selector SU_INPUT_SELECT_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX,GET_RES,
GET_INFO
wLength 1
Offset Field Size Value Description
0 bSelector 1 Number The setting for the attribute of the
Selector Control.

4.2.2.3 Processing Unit Control Requests

These requests are used to set or read an attribute of a video Control inside a Processing Unit of

Revision 1.5 August 9, 2012 94

 USB Device Class Definition for Video Devices

the video function.

The following paragraphs present a detailed description of all possible Controls a Processing
Unit can incorporate. For each Control, the layout of the parameter block together with the
appropriate Control Selector is listed for all forms of the Get/Set Processing Unit Control
request. All values are interpreted as unsigned unless otherwise specified.

Revision 1.5 August 9, 2012 95

 USB Device Class Definition for Video Devices

4.2.2.3.1 Backlight Compensation Control

The Backlight Compensation Control is used to specify the backlight compensation. A value of
zero indicates that the backlight compensation is disabled. A non-zero value indicates that the
backlight compensation is enabled. The device may support a range of values, or simply a binary
switch. If a range is supported, a low number indicates the least amount of backlight
compensation. The default value is implementation-specific, but enabling backlight
compensation is recommended. This control must accept the GET_DEF request and return its
default value.
Table 4-30 Backlight Compensation Control
Control Selector PU_BACKLIGHT_COMPENSATION_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX,
GET_RES, GET_INFO, GET_DEF
wLength 2
Offset Field Size Value Description
0 wBacklightCompensation 2 Number The setting for the attribute of the
addressed Backlight Compensation
control.

4.2.2.3.2 Brightness Control

This is used to specify the brightness. This is a relative value where increasing values indicate
increasing brightness. The MIN and MAX values are sufficient to imply the resolution, so the
RES value must always be 1. The MIN, MAX and default values are implementation dependent.
This control must accept the GET_DEF request and return its default value.
Table 4-31 Brightness Control
Control Selector PU_BRIGHTNESS_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_RES,
GET_INFO, GET_DEF
wLength 2
Offset Field Size Value Description
0 wBrightness 2 Signed The setting for the attribute of the
number addressed Brightness control.

4.2.2.3.3 Contrast Control

This is used to specify the contrast value. This is a relative value where increasing values
indicate increasing contrast. The MIN and MAX values are sufficient to imply the resolution, so
the RES value must always be 1. The MIN, MAX and default values are implementation
dependent. This control must accept the GET_DEF request and return its default value.
Table 4-32 Contrast Control
Control Selector PU_CONTRAST_CONTROL

Revision 1.5 August 9, 2012 96

 USB Device Class Definition for Video Devices

Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_RES,

GET_INFO, GET_DEF
wLength 2
Offset Field Size Value Description
0 wContrast 2 Number The setting for the attribute of the
addressed Contrast control.

4.2.2.3.4 Contrast, Auto Control

The Contrast Auto Control setting determines whether the device will provide automatic
adjustment of the related control. A value of 1 indicates that automatic adjustment is enabled.
Attempts to set the related Contrast control shall result in a protocol STALL and an error code of
bRequestErrorCode = ‘Wrong state’. This control must accept the GET_DEF request and return
its default value.
Table 4-33 Contrast, Auto Control
Control Selector PU_CONTRAST_AUTO_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO, GET_DEF
wLength 1
Offset Field Size Value Description
0 bContrastAuto 1 Number The setting for the
attribute of the
addressed Contrast,
Auto control.

4.2.2.3.5 Gain Control

This is used to specify the gain setting. This is a relative value where increasing values indicate
increasing gain. The MIN and MAX values are sufficient to imply the resolution, so the RES
value must always be 1. The MIN, MAX and default values are implementation dependent. This
control must accept the GET_DEF request and return its default value.
Table 4-34 Gain Control
Control Selector PU_GAIN_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_RES,
GET_INFO, GET_DEF
wLength 2
Offset Field Size Value Description
0 wGain 2 Number The setting for the attribute of the
addressed Gain control.

Revision 1.5 August 9, 2012 97

 USB Device Class Definition for Video Devices

4.2.2.3.6 Power Line Frequency Control

This control allows the host software to specify the local power line frequency, in order for the
device to properly implement anti-flicker processing, if supported. The default is
implementation-specific. This control must accept the GET_DEF request and return its default
value.
Table 4-35 Power Line Frequency Control
Control Selector PU_POWER_LINE_FREQUENCY_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO, GET_DEF
wLength 1
Offset Field Size Value Description
0 bPowerLineFrequency 1 Number The setting for the attribute of the
addressed Power Line Frequency
control:
0: Disabled
1: 50 Hz
2: 60 Hz
3: Auto

4.2.2.3.7 Hue Control

This is used to specify the hue setting. The value of the hue setting is expressed in degrees
multiplied by 100. The required range must be a subset of -18000 to 18000 (-180 to +180
degrees). The default value must be zero. This control must accept the GET_DEF request and
return its default value.
Table 4-36 Hue Control
Control Selector PU_HUE_CONTROL
Mandatory Requests GET_CUR, GET_MIN, GET_MAX, GET_RES, GET_INFO,
GET_DEF
Optional Requests SET_CUR
wLength 2
Offset Field Size Value Description
0 wHue 2 Signed The setting for the attribute of the
number addressed Hue control.

4.2.2.3.8 Hue, Auto Control

The Hue Auto Control setting determines whether the device will provide automatic adjustment
of the related control. A value of 1 indicates that automatic adjustment is enabled. Attempts to
programmatically set the related control shall result in a protocol STALL with an error code of
bRequestErrorCode = “Wrong state”. This control must accept the GET_DEF request and
return its default value.

Revision 1.5 August 9, 2012 98

 USB Device Class Definition for Video Devices

Table 4-37 Hue, Auto Control

Control Selector PU_HUE_AUTO_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO, GET_DEF
wLength 1
Offset Field Size Value Description
0 bHueAuto 1 Number The setting for the attribute of the
addressed Hue, Auto control.

4.2.2.3.9 Saturation Control

This is used to specify the saturation setting. This is a relative value where increasing values
indicate increasing saturation. A Saturation value of 0 indicates grayscale. The MIN and MAX
values are sufficient to imply the resolution, so the RES value must always be 1. The MIN, MAX
and default values are implementation-dependent. This control must accept the GET_DEF
request and return its default value.
Table 4-38 Saturation Control
Control Selector PU_SATURATION_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_RES,
GET_INFO, GET_DEF
wLength 2
Offset Field Size Value Description
0 wSaturation 2 Number The setting for the attribute of the
addressed Saturation control.

4.2.2.3.10 Sharpness Control

This is used to specify the sharpness setting. This is a relative value where increasing values
indicate increasing sharpness, and the MIN value always implies "no sharpness processing",
where the device will not process the video image to sharpen edges. The MIN and MAX values
are sufficient to imply the resolution, so the RES value must always be 1. The MIN, MAX and
default values are implementation-dependent. This control must accept the GET_DEF request
and return its default value.
Table 4-39 Sharpness Control
Control Selector PU_SHARPNESS_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_RES,
GET_INFO, GET_DEF
wLength 2
Offset Field Size Value Description
0 wSharpness 2 Number The setting for the attribute of the
addressed Sharpness control.

Revision 1.5 August 9, 2012 99

 USB Device Class Definition for Video Devices

4.2.2.3.11 Gamma Control

This is used to specify the gamma setting. The value of the gamma setting is expressed in gamma
multiplied by 100. The required range must be a subset of 1 to 500, and the default values are
typically 100 (gamma = 1) or 220 (gamma = 2.2). This control must accept the GET_DEF
request and return its default value.
Table 4-40 Gamma Control
Control Selector PU_GAMMA_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_RES,
GET_INFO, GET_DEF
wLength 2
Offset Field Size Value Description
0 wGamma 2 Number The setting for the attribute of the
addressed Gamma control.

4.2.2.3.12 White Balance Temperature Control

This is used to specify the white balance setting as a color temperature in degrees Kelvin. This is
offered as an alternative to the White Balance Component control. Minimum range should be
2800 (incandescent) to 6500 (daylight) for webcams and dual-mode cameras. The supported
range and default value for white balance temperature is implementation-dependent. This control
must accept the GET_DEF request and return its default value.
Table 4-41 White Balance Temperature Control
Control Selector PU_WHITE_BALANCE_TEMPERATURE_CONTROL
Mandatory Requests GET_CUR, GET_MIN, GET_MAX, GET_RES, GET_INFO,
GET_DEF
Optional Requests SET_CUR
wLength 2
Offset Field Size Value Description
0 wWhiteBalanceTe 2 Number The setting for the attribute of the addressed
mperature White Balance Temperature control.

4.2.2.3.13 White Balance Temperature, Auto Control

The White Balance Temperature Auto Control setting determines whether the device will
provide automatic adjustment of the related control. A value of 1 indicates that automatic
adjustment is enabled. Attempts to programmatically set the related control shall result in a
protocol STALL with an error code of bRequestErrorCode = “Wrong state”. This control must
accept the GET_DEF request and return its default value.
Table 4-42 White Balance Temperature, Auto Control
Control Selector PU_WHITE_BALANCE_TEMPERATURE_AUTO_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO, GET_DEF
wLength 1

Revision 1.5 August 9, 2012 100

 USB Device Class Definition for Video Devices

Offset Field Size Value Description

0 bWhiteBalanceTe 1 Number The setting for the attribute of the addressed
mperatureAuto White Balance Temperature, Auto control.

4.2.2.3.14 White Balance Component Control

This is used to specify the white balance setting as Blue and Red values for video formats. This
is offered as an alternative to the White Balance Temperature control. The supported range and
default value for white balance components is implementation-dependent. The device shall
interpret the controls as blue and red pairs. This control must accept the GET_DEF request and
return its default value.
Table 4-43 White Balance Component Control
Control Selector PU_WHITE_BALANCE_COMPONENT_CONTROL
Mandatory Requests GET_CUR, GET_MIN, GET_MAX, GET_RES,
GET_INFO, GET_DEF
Optional Requests SET_CUR
wLength 4
Offset Field Size Value Description
0 wWhiteBalanceBlue 2 Number The setting for the blue component of
the addressed White Balance
Component control.
1 wWhiteBalanceRed 2 Number The setting for the red component of
the addressed White Balance
Component control.

Revision 1.5 August 9, 2012 101

 USB Device Class Definition for Video Devices

4.2.2.3.15 White Balance Component, Auto Control

The White Balance Component Auto Control setting determines whether the device will provide
automatic adjustment of the related control. A value of 1 indicates that automatic adjustment is
enabled. Attempts to programmatically set the related control shall result in a protocol STALL
with an error code of bRequestErrorCode = “Wrong state”. This control must accept the
GET_DEF request and return its default value.
Table 4-44 White Balance Component, Auto Control
Control Selector PU_WHITE_BALANCE_COMPONENT_AUTO_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO, GET_DEF
wLength 1
Offset Field Size Value Description
0 bWhiteBalanceCo 1 Number The setting for the attribute of the
mponentAuto addressed White Balance Component,
Auto control.

4.2.2.3.16 Digital Multiplier Control

Use of this control has been deprecated and will be removed in the next revision of this
specification.

This is used to specify the amount of Digital Zoom applied to the optical image. This is the
position within the range of possible values of multiplier m, allowing the multiplier resolution to
be described by the device implementation. The MIN and MAX values are sufficient to imply
the resolution, so the RES value must always be 1. The MIN, MAX and default values are
implementation dependent. If the Digital Multiplier Limit Control is supported, the MIN and
MAX values shall match the MIN and MAX values of the Digital Multiplier Control. The Digital
Multiplier Limit Control allows either the Device or the Host to establish a temporary upper limit
for the Zcur value, thus reducing dynamically the range of the Digital Multiplier Control. If
Digital Multiplier Limit is used to decrease the Limit below the current Zcur value, the Zcur value
will be adjusted to match the new limit and the Digital Multiplier Control shall send a Control
Change Event to notify the host of the adjustment.
Table 4-45 Digital Multiplier Control
Control Selector PU_DIGITAL_MULTIPLIER_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_RES,
GET_INFO, GET_DEF
wLength 2
Offset Field Size Value Description
0 wMultiplierStep 2 Number The value Zcur

Revision 1.5 August 9, 2012 102

 USB Device Class Definition for Video Devices

4.2.2.3.17 Digital Multiplier Limit Control

This is used to specify an upper limit for the amount of Digital Zoom applied to the optical
image. This is the maximum position within the range of possible values of multiplier m. The
MIN and MAX values are sufficient to imply the resolution, so the RES value must always be 1.
The MIN, MAX and default values are implementation dependent.
Table 4-46 Digital Multiplier Limit Control
Control Selector PU_DIGITAL_MULTIPLIER_LIMIT_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_RES,
GET_INFO, GET_DEF
wLength 2
Offset Field Size Value Description
0 wMultiplierLimit 2 Number A value specifying the upper bound for
Zcur

4.2.2.3.18 Analog Video Standard Control

This is used to report the current Video Standard of the stream captured by the Processing Unit.
Table 4-47 Analog Video Standard Control
Control Selector PU_ANALOG_VIDEO_STANDARD_CONTROL
Mandatory Requests GET_CUR, GET_INFO
wLength 1
Offset Field Size Value Description
0 bVideoStandard 1 Number The Analog Video Standard of the input
video signal.
0: None
1: NTSC – 525/60
2: PAL – 625/50
3: SECAM – 625/50
4: NTSC – 625/50
5: PAL – 525/60
6-255: Reserved. Do not use.

4.2.2.3.19 Analog Video Lock Status Control

This is used to report whether the video decoder has achieved horizontal lock of the analog input
signal. If the decoder is locked, it is assumed that a valid video stream is being generated. This
control is to be supported only for analog video decoder functionality.

Table 4-48 Analog Video Lock Status Control

Control Selector PU_ANALOG_LOCK_STATUS_CONTROL
Mandatory Requests GET_CUR, GET_INFO
wLength 1

Revision 1.5 August 9, 2012 103

 USB Device Class Definition for Video Devices

Offset Field Size Value Description

0 bStatus 1 Number 0: Video Decoder is locked.
1: Video Decoder is not locked.
2-255: Reserved. Do not use.

4.2.2.4 Encoding Units

Encoding Unit requests are used to set or read the attributes of a video Control inside an
Encoding Unit of the video function. The following paragraphs present a detailed description of
all possible Controls an Encoding Unit can incorporate. For each Control, the Control data
structure together with the appropriate Control Selector is listed for all forms of the Get/Set
Encoding Unit Control requests. All values are interpreted as unsigned unless otherwise
specified.
Encoding Units, when considered as a whole, define the Encoding Unit State of the device. The
Encoding Unit State is per Encoding Unit and is established with a successful
VS_COMMIT_CONTROL(SET_CUR) to the associated VSInterface. The Encoding Unit State
establishes values for all fields in all supported Encoding Units.
Encoding Units can be called before streaming starts as well as during streaming. When called
before streaming, Encoding Units must not be called until a successful
VS_COMMIT_CONTROL(SET_CUR) request has been accepted by the device. Encoding
Units should use the Commit state, as returned by VS_COMMIT_CONTROL(GET_CUR), to
understand the boundaries of what is currently possible. A request issued to an Encoding Unit
applies to the layers/views specified by wLayerOrViewID. When the stream consists of a single
layer, the Encoding Unit is applied to the stream given by the stream_id subfield in
wLayerOrViewID, all other subfields in wLayerOrViewID shall be ignored by the device.

4.2.2.4.1 Encoding Units Operational Model

Device support of Encoding Unit controls is optional. Encoding Units can be used to configure
the codec before streaming or while streaming. All requests to Encoding Unit controls before a
successful VS_COMMIT_CONTROL(SET_CUR), except GET_LEN, shall result in protocol
STALL. Use VS_COMMIT_CONTROL(GET_CUR) to get the Commit state at any time. The
device shall use the Commit state together with the current Encoding Unit state to validate any
subsequent Encoding Unit control request. When a GET_CUR request is issued to the EU
control before any SET_CUR, the GET_CUR state shall be the same as the GET_DEF state.
Upon a successful VS_COMMIT_CONTROL(SET_CUR), the device must have a valid
GET_DEF state for each Encoding Unit control supported. Upon a successful
VS_COMMIT_CONTROL(SET_CUR), there may be fields in the Encoding Unit controls data
structures that have not yet been initialized such as parameters for rate control or resolutions for
a simulcast payload. For these unspecified fields, the device must establish default values such
that the device can successfully stream with no further configuration from the host. The rules the
device must follow to establish defaults are defined in Table 4-49 . After a successful
VS_COMMIT_CONTROL(SET_CUR) request, the Encoding Unit state can only be changed by
either issuing a successful SET_CUR request to an Encoding Unit Control, or by setting the
device to a new Commit state that is different from the previous Commit state. When the device

Revision 1.5 August 9, 2012 104

 USB Device Class Definition for Video Devices

enters a Commit state that is different from the previous Commit state, the device must update
the Encoding Unit state using the rules defined in Table 4-49 together with the new Commit data
structure.
Table 4-49 Default Encoding Unit State after VS_COMMIT_CONTROL(SET_CUR)
Request.

Method of Configuration

EU Control First stream Additional streams in

(stream_id = 0) simulcast payload
(stream_id = 1…n)
wLayerOrViewID 0 0

Width, height COMMIT Device determined.

(bFrameIndex) Must be less than or
equal to resolution of
first stream.
Minimum frame interval COMMIT Device
(dwFrameInterval) determined.Must be
greater than or equal to
minimum frame
interval of first stream.
Average bit rate Device determined Device determined
CPB size Device determined Device determined
Rate control mode COMMIT COMMIT
(bmRateControlModes) (bmRateControlModes)
Quantization parameter Device determined Device determined
Priority ID COMMIT (bUsage & COMMIT (bUsage &
bmLayoutPerStream) bmLayoutPerStream)
Slice mode Device determined Device determined
Profile COMMIT Same as first stream
(bFrameIndex)
bmSettings COMMIT Same as first stream.
(bFrameIndex)
Peak bit rate Peak bit rate in Rate Peak bit rate in Rate
Control Parameter EU. Control Parameter EU.
Level IDC COMMIT Device determined.
(bFrameIndex) Must be less than or
equal to LevelIDC of
first stream.
bNumHostControlLTRBuffers Device determined Device determined
bLTRMode Device determined Device determined
bmValidLTRs Device determined Device determined
bmErrorResiliencyFeatures Device determined Device determined

 COMMIT(x): Indicates that the parameter is given by the VS_COMMIT_CONTROL(GET_CUR) structure field indicated
within parentheses.

Revision 1.5 August 9, 2012 105

 USB Device Class Definition for Video Devices

In this specification, all Encoding Unit Controls except EU_SELECT_LAYER_CONTROL- and

EU_START_OR_STOP_LAYER_CONTROL are encoder configuration controls. The host may
issue SET_CUR requests to encoder configuration controls to configure encoder attributes for
one or multiple layers.
Once a Commit state has been established, the host may issue one or more Encoder
Configuration Control requests before or during streaming. From the perspective of this
specification, each Encoder Configuration Control operation is atomic.

4.2.2.4.1.1 Device States for Probe, Commit, and Encoding Units

Figure 4-1 illustrates three high level device states and the USB requests that can trigger
transitions between these device states when the requests succeed. USB requests not shown in
the Figure below, such as GET_CUR, GET_MIN, GET_MAX, GET_DEF among others, do not
trigger a state transition. State 0 represents the device in the USB Configured State (see Section
9.1.1.5 “Configured” of USB Specification Revision 2.0 and USB Specification Revision 3.0). At
this point, the Encoding Unit State in undefined. State 1 represents the active device state after a
successful VS_COMMIT_CONTROL(SET_CUR). Additional configuration of the encoder
through Encoding Units may occur before or during streaming. When in state 1 or 2, a
VS_COMMIT_CONTROL(SET_CUR) equal to the current
VS_COMMIT_CONTROL(GET_CUR) will have no effect on the device, that means that the
Encoding Unit state will also remain the same. When in state 1 or 2, a
VS_COMMIT_CONTROL(SET_CUR) different than the current
VS_COMMIT_CONTROL(GET_CUR) will update both the Commit and the Encoding Unit
State according to the new Commit state and the rules defined in Table 4-49; this transition is
not shown in the state diagram.
Table 4-50 describes how Encoding Units interact with each these states. This description
includes possible errors that may be logged by the device in response to USB Requests issued to
Encoding Unit controls. Errors may be retrieved by the host using the Request Error Code
Control (see section 4.2.1.2 "Request Error Code Control").

Revision 1.5 August 9, 2012 106

 USB Device Class Definition for Video Devices

NOT STREAMING STREAMING

EU_*_CONTROL(SET_CUR) EU_*_CONTROL(SET_CUR)
SET_ INTERFACE (> 0)

0 1 2

VS_COMMIT_CONTROL(SET_CUR)

SET_ INTERFACE(0)

Figure 4-1 Device State Transitions Diagram

When in state 2, streaming, the only valid parameter value in a SET_INTERFACE request is
alternate setting 0.

Table 4-50 Encoding Units, Devices States and Error Code Control Responses
State Streaming State Encoding Unit Error Code Control Response to USB
State Requests issued to Encoding Unit Controls
0 not streaming Undefined, except Protocol STALL with error code:
GET_LEN  “Invalid Control” if EU Control is not
supported. Else,
 “Invalid Request” if USB request is not
supported for this Control. Else,
 “Wrong State” if EU Control is supported
after initial
VS_COMMIT_CONTROL(SET_CUR)
1 not streaming defined by the Protocol STALL with error code:
default values as  “Invalid Control” if EU Control is not
given in Table 4-49 supported. Else,
together with the  “Invalid Request” if USB request is not
values set in a supported for this Control. Else,
successful  “Wrong State” if EU is supported only
SET_CUR request while streaming or if the active
to the Encoding wLayerOrView is not valid. Else,
Unit  “Out of Range” if any of the input

Revision 1.5 August 9, 2012 107

 USB Device Class Definition for Video Devices

arguments to the EU is invalid

2 streaming defined by the Protocol STALL with error code:
default values as  “Invalid Control” if EU Control is not
given in Table 4-49 supported. Else,
together with the  “Invalid Request” if USB request is not
values set in a supported for this Control. Else,
successful  “Wrong State” if EU Control is supported
SET_CUR request only before streaming or if the active
to the Encoding wLayerOrView is not valid. Else,
Unit  “Out of Range” if any of the input
arguments to the EU is invalid

4.2.2.4.1.2 Encoding Unit Requests

Table 4-51describes Encoding Unit request attributes. Each Encoding Unit supports one or more
of these attributes as specified in sections 4.2.2.4.2 through 4.2.2.4.19.
Table 4-51 Encoding Unit Requests
Attribute Description
GET_CUR Returns the current values of all fields. Prior to the initial
VS_COMMIT_CONTROL(SET_CUR), the GET_CUR state
for all Encoding Units is undefined. After the the initial
VS_COMMIT_CONTROL(SET_CUR) and prior to the initial
SET_CUR request to the Encoding Unit control, the
GET_CUR state must be the same as the GET_DEF state.
GET_MIN Returns the minimum value for all fields.
GET_MAX Returns the maximum value for all fields.
GET_RES Returns the resolution value for all fields.
GET_DEF Returns the default value for all fields.
GET_INFO Returns the capabilities and status of the specified control
GET_LEN Returns the data length of the specified control
SET_CUR Sets the current Encoding Unit state. For Encoder
Configuration Controls, SET_CUR is used for layer
configuration and sets all the fields.
This request shall stall in case an unsupported state or field is
specified. See the Request Error Code Control section for the
definition of specific request error codes.

4.2.2.4.1.3 Stream Negotiation Examples

The example below shows a successful configuration and streaming of a temporally encoded
video stream.

Revision 1.5 August 9, 2012 108

 USB Device Class Definition for Video Devices

Host Device

PROBE_CONTROL(SET_CUR)

PROBE_CONTROL(GET_CUR)

Configure device for up to COMMIT_CONTROL(SET_CUR)

720p @ 30 fps streaming
Commit to 720p @ 30 fps

COMMIT_CONTROL(GET_CUR)

EU_RESOLUTION_CONTROL(SET_CUR)

Change the resolution to 360p

EU_AVERAGE_BITRATE_CONTROL(SET_CUR)

Change the bitrate to 400 kbps

Scale back to 360p EU_CPB_SIZE_CONTROL(SET_CUR)
streaming before starting
Change the CPB size to 12500 =
(400 kbps * 500ms) / 16
SELECT_ALTERNATE_INTERFACE(1)

Start streaming

Dynamically scale up to 720p EU_CPB_SIZE_CONTROL(SET_CUR)

streaming
Change the CPB size to 406250 =
EU_AVERAGE_BITRATE_CONTROL(SET_CUR) (1300 kbps * 500ms) / 16

Change the bitrate to 1300 kbps

EU_RESOLUTION_CONTROL(SET_CUR)

Change the resolution to 720p

Figure 4-2 Successful USB Isochronous Negotiation Example including Dynamic

Resolution Changes
Figure 4-2 illustrates stream configuration and successful USB isochronous bandwidth
negotiation in a single stream encoded video scenario. In this example, during Probe/Commit the
host negotiates a video stream to allow a resolution and frame rate up to 720p at 30 fps. Then,
before streaming, the host configures the device to stream 360p at 30 fps, including the

Revision 1.5 August 9, 2012 109

 USB Device Class Definition for Video Devices

appropriate bitrate and CPB size. At this point the host starts the stream selecting alternate
interface number 1. At a later time, while still streaming, the host increases the resolution to
720p. To do this, the host first increases the CPB size and then increases the bit rate. It is
important to keep within the limits of all bitrate and buffer size limits when making these
changes. Finally, the host changes the resolution to 720p.

The example below shows a failed Encoding Unit request to set the frame rate to a higher value
than was negotiated.

Host Device

PROBE_CONTROL(SET_CUR)

PROBE_CONTROL(GET_CUR)

Configure device for up to COMMIT_CONTROL(SET_CUR)

VGA @ 15 fps streaming
Commit to VGA @ 15 fps

COMMIT_CONTROL(GET_CUR)

SELECT_ALTERNATE_INTERFACE(1)

Start streaming

Increase the frame rate to EU_MIN_FRAME_INTERVAL_CONTROL (SET_CUR)

30 fps
Set the frame interval to 33 ms
(30 fps)

Try a smaller resolution at (Fail)

30 fps EU_RESOLUTION_CONTROL(SET_CUR)

Change the resolution to QVGA

EU_MIN_FRAME_INTERVAL_CONTROL (SET_CUR)

Set the frame interval to 33 ms

(30 fps)

(Fail)

Revision 1.5 August 9, 2012 110

 USB Device Class Definition for Video Devices

Figure 4-3 Failed Request to an Encoding Unit Control Example

Figure 4-3 demonstrates a failed Encoding Unit request. In this example, the host attempts to set
the frame interval to a shorter duration than negotiated during Probe and Commit. After failure,
the host attempts to negotiate a shorter frame interval by reducing the resolution first. This
attempt also fails since the limits on frame interval are independent of the resolution or any other
stream parameter. This example illustrates the importance of negotiating the expected maximum
for each parameter (in this case minimum frame interval) as part of Probe and Commit.

4.2.2.4.1.4 Setting Resolution and Frame Interval

When changing resolution and / or minimum frame interval, whether during initialization or
streaming, the following restrictions apply.
 The resolution must be equal to or less than the value established during Probe and
Commit
 The frame interval must be equal to or greater than the frame interval established during
Probe and Commit
 Care should be taken to avoid exceeding the negotiated bandwidth and buffer size limits
when changing resolution, bitrate, or frame interval.

4.2.2.4.1.5 Wildcard Masks

Wildcard masks may be used in association with the wLayerOrViewID field of the Select Layer
control. Use of a wildcard mask selects all layers whose type is determined by the associated sub-
field(s), for example, using a wildcard mask in the temporal_id subfield results in the selection of all
temporal layers. Wildcard masks must address one or more subfield. Any Encoding Unit called
subsequent to a Select Layer request using wildcard masks must respect the resulting layer selection
unless otherwise noted in the control. Note that using wildcards implies a transaction where the
Request must succeed or fail completely, not partially. Except for the Select Layer control,
GET_MIN, GET_MAX, GET_DEF, GET_RES and GET_CUR requests shall protocol STALL if
wLayerOrViewID is SET to a wildcard mask. The example table below shows which bits are set in
wLayerOrViewID to create a wildcard mask for the four subfields.

Table 4-52 Bit Layout of wLayerOrViewID for SVC Wildcard Masks

wLayerOrViewID reserved stream_id temporal_id quality_id dependency_id
bits 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
All Dependency 1 1 1
Layers
All Quality 1 1 1 1
Layers
All Temporal 1 1 1
Layers
All Streams 1 1 1

Revision 1.5 August 9, 2012 111

 USB Device Class Definition for Video Devices

4.2.2.4.2 Select Layer Control

The Select Layer control is used to specify the active layer/view for subsequent EU control
operations. This includes the stream_id for simulcast streams. The default value for
wLayerOrViewID is 0. This value selects the base layer of the first stream. Multiple
layers/views can be specified in wLayerOrViewID by using a wildcard mask for SET_CUR
requests. For the Select Layer control wildcard masks are also allowed for GET_CUR requests.
Table 4-53 Select Layer Control
Control Selector EU_SELECT_LAYER_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO, GET_LEN
wLength 2
Offset Field Size Value Description

0 wLayerOrViewID 2 Number For multi-layer streams, a combination of

dependency_id, quality_id, temporal_id and
stream_id. Bits:
0-2: dependency_id
3-6: quality_id
7-9: temporal_id
10-12: stream_id
13-15: Reserved, set to 0

For multi-view streams, this value contains a

combination of view_id, temporal_id,
stream_id, and interface number. Bits:
0-6: view_id
7-9: temporal_id
10-12: stream_id
13-15: Reserved, set to 0

When the stream does not support layering or

multiples view, bits 0-9 are zero.

The dependency_id, quality_id, temporal_id, and view_id of each layer of a stream are determined
by the scaling capability mode negotiated in Probe and Commit. The stream_id parameter is used to
differentiate between streams when simulcast of two or more streams is enabled. In the case of a
single stream, stream_id is always zero. Each additional stream is given a unique stream_id by
incrementing the stream_id by 1. If the stream supports simulcast transport and the stream_id does
not exist, the device should protocol STALL with error code “Out of Range”. If wLayerOrViewID
specifies a layer or view that is not defined by the bmLayoutPerStream established during
VS_COMMIT the device shall protocol STALL with error code “Out of Range”. If the stream does
not support simulcast transport and the stream_id is not zero, the device should ignore the stream_id
and proceed with the Select Layer request for the first stream.

Revision 1.5 August 9, 2012 112

 USB Device Class Definition for Video Devices

4.2.2.4.2.1 Sub-bitstream Definition

Several Encoding Units apply to sub-bitstreams instead of individual layers. The sub-bitstream is
determined as follows. Let wLayerOrViewID be the result from a GET_CUR request issued to
the EU_SELECT_LAYER_CONTROL control. The sub-bitstream is given by all the layers that
meet all the following conditions:
 stream_id is equal to the stream_id indicated by wLayerOrViewID.
 temporal_id is less or equal than temporal_id indicated by wLayerOrViewID.
 dependency_id is less than dependency_id indicated by wLayerOrViewID, or
dependency_id is equal to dependency_id indicated by wLayerOrViewID and quality_id
is less or equal than quality_id indicated by wLayerOrViewID.

4.2.2.4.2.2 Multicast
This specification supports multicast, e.g. multiple encoded video streams from a single video
function. The device must offer a separate Encoding Unit for each Video Streaming interface that
delivers encoded video.

Video Control Interface

Video Streaming
Video Function Interface 1

Sensor CT PU OT USB IN Endpoint 1

Video Streaming
Interface 2
EU OT USB IN Endpoint 2

Video Streaming
Interface 3
EU OT USB IN Endpoint 3

Figure 4-4 Example of Multicast Topology

By using the bSourceID of the Encoding Units, Output Terminals, and VS Interfaces, the device
topology above can be understood by the host.
The device must support a complete Probe and Commit operation for each streaming interface, one at
a time. The expected behavior is that Probe and Commit is complete on one interface before
beginning with another. Additionally, if a device supports multicast, GET_MAX requests to the

Revision 1.5 August 9, 2012 113

 USB Device Class Definition for Video Devices

Probe control must accurately reflect the capabilities of the device, including what is already been
negotiated using VS_COMMIT_CONTROL on other streaming interfaces.

4.2.2.4.3 Video Resolution Control

The Video Resolution control is used to independently initialize the resolution of each layer in a
multi-layer stream. The wWidth and wHeight fields are used to set the width and height of the
decoded bitmap frame of a stream.
When using this control to change video resolution, the new resolution is restricted to ratios as
described in the bResolutionScaling field in the video Format descriptor.
The bit rate limits described by dwMinBitRate and dwMaxBitRate may change when video
resolution is changed. The application should use the GET_MIN and GET_MAX request on the
Rate Control Parameter control to determine the new values for minimum bit rate and maximum
bit rate.
If the stream does not support spatial scalability, then this control applies to the stream_id of the
active wLayerOrViewID and the device shall ignore all other subfields in
wLayerOrViewID.GET_RES applies to encoders that support arbitrary resolutions and shall
reflect any limitations on the step size between resolutions.
Table 4-54 Video Resolution Control
Control Selector EU_VIDEO_RESOLUTION_CONTROL
Mandatory SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_DEF,
Requests GET_INFO, GET_LEN
wLength 4
Offset Field Size Value Description
0 wWidth 2 Number The width, in pixels, of pictures output from the
decoding process.
2 wHeight 2 Number The height, in pixels, of pictures output from the
decoding process.

4.2.2.4.4 Profile and Toolset Control

The Profile and Toolset control is used to specify the profile_idc and constraint flags for the
current layer(s) and to further constrain features within the profile. There are restrictions on
using this control to change profile. Specifically, this control cannot be used to increase the
number of layers beyond the number negotiated in Probe & Commit. Also, this control cannot be
used to move the codec between different bUsage values. The wConstrainedToolset field
contains reserved bits for future extensibility.
Table 4-55 Profile Toolset Control
Control Selector EU_PROFILE_TOOLSET_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_DEF, GET_INFO,
GET_LEN
wLength 5
Offset Field Size Value Description
0 wProfile 2 Number wProfile indicates the profile and

Revision 1.5 August 9, 2012 114

 USB Device Class Definition for Video Devices

applicable constraints for the current

format. This field is used with video
formats that support multiple profiles. For
example, for H.264:
0x4240: Constrained Baseline Profile
0x4200: Baseline Profile
0x4D00: Main Profile
0x640C: Constrained High Profile
0x6400: High Profile
0x5304: Scalable Constrained Baseline Profile
0x5300: Scalable Baseline Profile
0x5604: Scalable Constrained High Profile
0x5600: Scalable High Profile
0x7600: Multiview High Profile
0x8000: Stereo High Profile
2 wConstrainedToolset 2 Number Reserved. Set to 0.
4 bmSettings 1 Bitmap This bitmap enables features reported by
the bmCapabilities field of the Video
Format Descriptor.

Refer to payload specification for details.

4.2.2.4.5 Minimum Frame Interval Control

The Minimum Frame Interval control is used to specify the minimum frame interval of the
stream associated with the current wLayerOrViewID. This value reflects the maximum frame
rate when combining the base layer and all temporal enhancement layers. The new
dwFrameInterval must be a frame interval that is advertised in the currently negotiated Frame
descriptor and greater than or equal to the frame interval negotiated in probe and commit. The
minimum frame interval is given per stream where the stream is indicated by the stream_id
subfield in wLayerOrViewID. All other subfields in wLayerOrViewID shall be ignored by the
device.
For SET_CUR requests, this control can be used to configure multiple streams simultaneously by
using a wildcard mask-e.g. stream_id=7 (‘111’). For streams with multiple temporal layers each
layer shall have a frame interval that is dwFrameInterval * 2(N-1-layer_number), where N is the
number of temporal layers and layer_number is an incremental identifier for each layer starting
at 0 for the base layer and incrementing by 1 for each successive temporal layer. For example, a
stream with three temporal layers may have aggregate frame intervals of 333333, 666666, and
1333332.
Table 4-56 Minimum Frame Interval Control
Control Selector EU_MIN_FRAME_INTERVAL_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_DEF,
GET_INFO, GET_LEN
wLength 4

Revision 1.5 August 9, 2012 115

 USB Device Class Definition for Video Devices

Offset Field Size Value Description

0 dwFrameInterval 4 Number Frame interval in 100-ns units. Cannot go below
the value of dwFrameInterval set in
Probe/Commit. The value must be an interval
advertised in the associated Frame Descriptor that
has the same resolution, profile, and toolset of the
current stream.
This control is applied per stream and ignores any
layer information in wLayerOrViewID.
Note that changing the frame interval using this control may also change the bit rate limits
described by dwMinBitRate and dwMaxBitRate in the negotiated video Frame descriptor. Use
GET_MIN and GET_MAX with the Average Bitrate control to determine the new values for
minimum bit rate and maximum bit rate.

4.2.2.4.6 Slice Mode Control

The Slice Mode control is used to specify the slice mode for the current wLayerOrViewID. This
control is only supported if the codec supports using more than one slice per frame. If the device
supports this control it must support one or more of the slice modes described below.
Table 4-57 Slice Mode Control
Control Selector EU_SLICE_MODE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_DEF, GET_MIN, GET_MAX,
GET_INFO, GET_LEN
wLength 4
Offset Field Size Value Description
0 wSliceMode 2 Number Most significant byte: Mode.
0: Maximum number of MBs per slice mode
1: Target compressed size per slice mode
2: Number of slices per frame mode
3: Number of Macroblock rows per slice mode
4-255: Reserved

2 wSliceConfig The meaning of this field depends on wSliceMode

Number
Setting mode:
Mode 0: Maximum number of MBs per slice.
Mode 1: Target size for each slice NALU in bytes.
Mode 2: Number of slices per frame.
Mode 3: Number of macroblock rows per slice.
GET_MIN, GET_MAX, and GET_RES return values for wSliceConfigSetting based on the
current wSliceMode

4.2.2.4.7 Rate Control Mode Control

The Rate Control Mode control is used to specify the rate control mode of the current layer.
Once the mode is established, rate control parameters are set using the Peak Bit Rate Control,
Average Rate Control Parameter Control, or Quantization Control.

Revision 1.5 August 9, 2012 116

 USB Device Class Definition for Video Devices

For variable and constant bit rate buffer modeling, this document specifies rate control operation
in terms of a leaky bucket model. The bits used to encode each picture are analogous to cups of
water being dumped into the top of the bucket when each picture is encoded (after re-ordering
the pictures as necessary for bitstream orders that differ from display order); the level of water in
the bucket indicates the number of bits waiting to be sent to the decoder; and the water leaking
out of a hole in the bottom of the bucket corresponds to bits flowing into the decoder through a
transmission channel. The leaky bucket is a traffic meter that contains two parameters:
 RP = dwPeakBitRate (bits per second), which is the peak bit rate at which bit can flow
out from the bottom of the bucket
 B = dwCPBsize (bits), which is the coded picture buffer (CPB) capacity.

Picture

Picture

Picture

Buffer (B)
Bits waiting
to be sent

Peak Bitrate (Rp)

The buffer serves to smooth out local bit rate fluctuations while limiting the total bit usage that is
possible over longer durations and limiting the buffering capacity necessary for a decoder to be
able to decode the video content.
The leaky bucket model at the encoder has a corresponding mirror-image model that operates
from the decoder perspective. As bits leak out of the encoder buffer, they conceptually enter into

Revision 1.5 August 9, 2012 117

 USB Device Class Definition for Video Devices

a corresponding decoder input buffer, which continues to fill up until the decoding time of a
picture arrives – at which time the bits for that picture are removed from the decoder’s CPB.
If too many bits are dumped into the bucket too quickly, the buffer capacity B would be
exceeded before enough bits have time to drain out of the hole in the bottom of the bucket, and
the buffer is said to “overflow” from the encoder perspective. From the decoder perspective, an
overflow could occur if the removal of pictures from the decoder CPB at the decoding times of
those pictures is not fast enough to keep up with the amount of bits that have been flowing into it
from the encoder.
The encoder shall ensure that the leaky bucket never overflows.
Table 4-58 Rate Control Mode Control
Control Selector EU_RATE_CONTROL_MODE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_DEF, GET_INFO, GET_LEN
wLength 1
Offset Field Size Value Description
0 bRateControlMode 1 Number 0: Reserved
1: Variable Bit Rate low delay (VBR)
2: Constant bit rate (CBR)
3: Constant QP
4: Global VBR low delay (GVBR)
5: Variable bit rate non-low delay (VBRN)
6: Global VBR non-low delay (GVBRN)
7-255: Reserved

Note that bRateControlMode selected must be marked as supported by

bmSupportedRateControlModes in the Video Format Descriptor.

4.2.2.4.7.1 Variable Bit Rate (VBR)

The VBR control mode limits the long-term maximum bit rate of transmission, but allows the bit
rate to vary substantially on a short-term basis. In particular, in the VBR mode the encoder is not
required to continue to produce a significant bit rate in cases where there is little or no motion
activity in the coded video scene content.
In the variable bit rate (VBR) mode of bit rate control operation, it is allowed for the encoder’s
leaky bucket to sometimes “run dry” – i.e., for all bits to leave the encoder CPB and for there to
be some period of time during which no bits are flowing. The presence of such periods of time
during which no bits are flowing implies that the average bit rate over a long duration of video
content can be less than the peak bit rate RP.
The average bit rate parameter dwAverageBitRate serves only as a guideline, as follows:
The average bit rate parameter dwAverageBitRate represents only a “target” or
“guideline”, indicating the average bit rate expected to be produced by the video encoder
when operating under normal lighting conditions with a normal degree of video scene
activity. To calculate dwAvergeBitRate the encoder should sample over the period T =
CPB size / Peak bit rate.

Revision 1.5 August 9, 2012 118

 USB Device Class Definition for Video Devices

It is allowed for the bit rate produced by the video encoder to exceed dwAverageBitRate
(e.g., when there is an exceptionally high degree of activity in the video scene).
It is also allowed for the bit rate produced by the video encoder to be less than
dwAverageBitRate (e.g., when there is very little activity in the video scene or when
lighting conditions are poor).
In contrast, the peak bit rate RP = dwPeakBitRate bps and the total buffer capacity B,
which are the operating parameters of the leaky bucket model, correspond to a mandatory
maximum not to be exceeded by the encoder on a long-term basis (i.e., the leaky bucket model
shall not overflow).
The average bit rate parameter dwAverageBitRate shall be set to a value less than or equal to peak
bit rate = dwPeakBitRate x 64.
For multi-layer bitstreams, the VBR-control model applies to the currently selected sub-bitstream
as defined in section 4.2.2.4.2.1 For single layer bitstreams, this rate control model applies to the
entire bitstream, because there is only one layer in the bitstream.

4.2.2.4.7.2 Constant Bit Rate (CBR)

CBR control mode is similar to the VBR control mode, except that dummy bits are generated
when necessary in order to assure an exactly constant bit rate flowing from the encoder leaky
bucket (there are always bits flowing to the decoder). For CBR operation the average bit rate is
equal to the peak bit rate, as there is no variability in the bit rate in this case. In the CBR mode, it
is required that the average bit rate parameter dwAverageBitRate shall be set equal to the peak
bit rate = dwPeakBitRate x 64.
For multi-layer bitstreams, this rate-control model applies to the currently selected sub-bitstream
as defined in section 4.2.2.4.2.1. For single layer bitstreams, this rate-control model applies to
the entire bitstream, because there is only one layer in the bitstream.
The encoder shall ensure that the leaky bucket never overflows nor underflows in the CBR case.

4.2.2.4.7.3 Constant QP Mode

In Constant QP mode, the encoder shall use the dwQpPrime fields to derive a constant QP for I,
P, and B slices (or EI, EP, and EB slices if quality or spatial scalability is employed).

4.2.2.4.7.4 Global VBR

In Global VBR mode, the rate-control model applies to the entire bitstream. For single-layer
bitstreams, this mode is identical to VBR mode. For multi-layer bitstreams, this mode implies
that no explicit rate control is required for the sub-bitstreams, as long as the entire bitstream
obeys the rate-control model specified in the Rate Control Mode EU.

4.2.2.4.7.5 Low Delay and Non-Low Delay Modes

See individual payload specifications for details on how these modes are supported.

Revision 1.5 August 9, 2012 119

 USB Device Class Definition for Video Devices

4.2.2.4.8 Average Bit Rate Control

The Average bit rate control is used to specify the average bit rate of the current selected
substream. This control is used in all bRateControlMode modes except Constant QP. SET_CUR
shall protocol STALL with an error code of bRequestErrorCode = “Wrong state” if the
specified stream is not set to CBR, VBR, VBRN, Global VBR or Global VBRN.
For single bitstreams or when in Global VBR rate-control mode, this EU applies to the entire
bitstream. For multi-layer streams, dwAverageBitRate is set per sub-bitstream, as defined in section
4.2.2.4.2.1.
This EU has no effect when the device operates in Constant QP rate-control mode.

The dwAverageBitRate value returned by the device upon a GET_MIN request specifies the
minimum average bit rate for the sub-bitstream at the sub-bitstream frame interval. For
GET_MAX, the device returns the maximum capability for the overall stream at the current
frame interval for that stream, summing up bit rates across all layers in the sub-bitstream. These
values are for the active resolution as specified by
EU_VIDEO_RESOLUTION_CONTROL(GET_CUR).
Table 4-59: Average Bitrate Control
Control Selector EU_AVERAGE_BITRATE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_DEF,
GET_INFO, GET_LEN, GET_RES
wLength 4
Offset Field Size Value Description
0 dwAverageBitRate 4 Number Average bit rate, in bits per second. Must be
less than or equal to dwPeakBitRate x 64.
Applies for all rate control modes except
Constant QP.

This control should be reported as an AutoUpdate control. When this control is reported as
AutoUpdate, any changes to the encoder that alter the values of GET_MIN and GET_MAX must
be reported via AutoUpdate.

4.2.2.4.9 CPB Size Control

The CPB Size control is used to specify the CPB size of the current selected substream. This
control is used in all bRateControlMode modes except Constant QP. SET_CUR shall protocol
STALL with an error code of bRequestErrorCode = “Wrong state” if the specified stream is
not set to CBR, VBR, VBRN, Global VBR or Global VBRN.
For single bitstreams or when in Global VBR rate-control mode, this EU applies to the entire
bitstream. For multi-layer streams, dwCPBsize is set per sub-bitstream, as defined in section 9.3.4.
EU has no effect when the device operates in Constant QP rate-control mode.
The dwCPBsize value returned by the device upon a GET_MIN request specifies the minimum
CPB size for use with the average bit rate for the current sub-bitstream. For GET_MAX the
device shall return the maximum CPB size for the current sub-bitstream supported by the device

Revision 1.5 August 9, 2012 120

 USB Device Class Definition for Video Devices

for the given bLevelIDC specified in the Frame descriptor. The host should set the current
dwCPBsize appropriately for the dwPeakBitRate.

Table 4-60: CPB Size Control

Control Selector EU_CPB_SIZE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_DEF,
GET_INFO, GET_LEN, GET_RES
wLength 4
Offset Field Size Value Description
0 dwCPBsize 4 Number CPB size, in units of 16 bits, corresponding to
the CPB size. Must be a size supported by the
profile and level combination, if that applies.
This control applies for all rate control modes
except Constant QP.
This control should be reported as an AutoUpdate control. When this control is reported as
AutoUpdate, any changes to the encoder that alter the values of GET_MIN and GET_MAX must
be reported via AutoUpdate.

4.2.2.4.10 Peak Bit Rate Control

The Peak Bit Rate control is used to specify the peak bit rate of the current layer. This control
only applies when bRateControlMode is set to VBR, VBRN, Global VBR or Global VBRN.
SET_CUR shall protocol STALL with an error code of bRequestErrorCode = “Wrong state” if
the specified stream is not set to VBR, VBRN, Global VBR or Global VBRN.
Table 4-61: Peak Bit Rate Control
Control Selector EU_PEAK_BIT_RATE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_DEF,
GET_INFO, GET_LEN, GET_RES
wLength 4
Offset Field Size Value Description
0 dwPeakBitRate 4 Number Peak bit rate, in units of 64 bits per second. Must
be a rate supported by the profile and level
combination. Used only for VBR, VBRN, Global
VBR and Global VBRN mode.
This control should be reported as an AutoUpdate control. When this control is reported as
AutoUpdate, any changes to the encoder that alter the values of GET_MIN and GET_MAX must
be reported via AutoUpdate.
.

4.2.2.4.11 Quantization Parameter Control

The Quantization Parameter control is used to specify quantization parameters for the current
layer. This control only applies when bRateControlMode is set to Constant QP. SET_CUR shall
stall if the bRateControlMode mode for the active layer is not set to Constant QP. The value of
the quantization parameters set by this control are valid until modified by this control, or until

Revision 1.5 August 9, 2012 121

 USB Device Class Definition for Video Devices

the rate control mode changes to something other than Constant QP. GET_MIN, GET_MAX,
and GET_RES shall return values for each field.

Table 4-62: Quantization Parameter Control

Control Selector EU_QUANTIZATION_PARAMS_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MAX, GET_MIN, GET_DEF,
GET_INFO, GET_LEN, GET_RES
wLength 6
Offset Field Size Value Description
0 wQpPrime_I 2 Number Only applicable in constant QP rate-control mode.
Use this parameter to set/get QP for I frames. The
details on how to interpret this number is located in
the appropriate payload specification.

2 wQpPrime_P 2 Number Only applicable in constant QP rate-control mode.

Use this parameter to set/get QP for P frames. The
details on how to interpret this number is located in
the appropriate payload specification.

4 wQpPrime_B 2 Number Only applicable in constant QP rate control mode.

Use this parameter to set/get QP for B frames. The
details on how to interpret this number is located in
the appropriate payload specification.

If a slice type (I, P, B) is not supported by the format, device or if it is not supported for the
active wLayerOrViewID, then
 For a SET_CUR request to the EU_QUANTIZATION_PARAMS_CONTROL the
device will ignore the wQpPrime field for that slice type.

 For a GET_CUR request to EU_QUANTIZATION_PARAMS_CONTROL the device

will return 0xFFFF for the wQpPrime field for that slice type.

4.2.2.4.12 Quantization Parameter Range Control

The QP Range control is used to specify the allowed quantization parameters when encoding to
target bitrate. This control can be used in all bRateControlMode modes except Constant QP.
SET_CUR shall protocol STALL with an error code of bRequestErrorCode = “Wrong state” if
bRateControlMode is Constant QP.
For single layer bitstreams, this EU applies to the entire bitstream. For multi-layer streams, bMinQP
and bMaxQP are set per sub-bitstream, as defined in section 9.3.4.
Table 4-63: Quantization Parameter Range Control
Control Selector EU_QP_RANGE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_RES
GET_INFO, GET_DEF, GET_LEN

Revision 1.5 August 9, 2012 122

 USB Device Class Definition for Video Devices

wLength 2
Offset Field Size Value Description
0 bMinQp 1 Number Minimum quantization parameter to use for the frame
(both luma and chroma).
1 bMaxQp 1 Number Maximum quantization parameter to use for the frame
(both luma and chroma).

This control can be used to narrow the range of possible quantization parameter values (for both
luma and chroma) to use in the encoding process. GET_MIN returns the highest bMinQp and
lowest bMaxQp for the given rate control mode and negotiated bit rate. GET_MAX returns the
lowest bMinQp and highest bMaxQp for the given rate control mode and negotiated bit rate.
The bMinQp and bMaxQp values reflect the global minimum and maximum Qp, including all
offsets.
This control should be reported as an AutoUpdate control. When this control is reported as
AutoUpdate, any changes to the encoder that alter the values of GET_MIN and GET_MAX must
be reported via AutoUpdate.

4.2.2.4.13 Synchronization and Long Term Reference Frame Control

This control is used to manage insertion of synchronization frames and long-term reference
frames into the current layer. For options that insert an IDR or GDR, this control applies to the
entire stream. When the host requests the generation of a sync frame, the encoder shall insert the
specified bSyncFrameType into all the dependency layers associated with the current stream as
identified by the current wLayerOrViewID. This control can be used to generate long term
reference frames only if bit D6 in bmCapabilities in the negotiated Frame Descriptor is set to 1.

Table 4-64: Synchronization and Long Term Reference Frame Control

Control Selector EU_SYNC_REF_FRAME_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_INFO,
GET_LEN
wLength 4
Offset Field Size Value Description
0 bSyncFrameType 1 Number 0: Reset. Allow the encoder to determine the
timing and type of synchronization frames.
1: Generate an IDR frame for all of the
dependency layers of the current stream, if all
of the layers of the current stream have
quality_id equal to 0. Otherwise, generate key
frames for all the dependency layers.
2: Generate an IDR frame that is a long-term
reference frame for all of the dependency layers
of the current stream, if all of the layers of the
current stream have quality_id equal to 0.
Otherwise, generate a key frame for the
associated dependency layers of the current

Revision 1.5 August 9, 2012 123

 USB Device Class Definition for Video Devices

wLayerOrViewID.
3: Generate a non-IDR random-access I frame
for the associated dependency layers of the
current wLayerOrViewID.
4: Generate a non-IDR random-access I frame
that is a long-term reference frame for the
associated dependency layers of the current
wLayerOrViewID.
5: Generate a P frame that is a-long term
reference frame for the associated dependency
layers of the current wLayerOrViewID.
6: Gradual Decoder Refresh (GDR)
7-255: Reserved.
1 wSyncFrameInte 2 Number In milliseconds. This field indicates the periodic
rval recurrences of the selected bSyncFrameType.
A value of wSyncFrameInterval = 0 indicates
a single bSyncFrameType with no requirement
for periodic recurrence.
3 bGradualDecode 1 Number Indicates a count of frames over which the
rRefresh gradual decoder refresh occurs. Only valid
when bSyncFrameType = 6 (GDR). When
bSyncFrameType is not 6, this field must be 0.
From a recovery point of view,
bGradualDecoderRefresh + 1 represents the
number of frames required to completely
refresh the picture.

Bits:
0-6: recovery_frame_cnt
7: Reserved

Use wSyncFrameInterval to establish the

interval between Gradual Decoder Refresh
(GDR) periods.
When using GDR, the encoder should use
recovery point SEI messages to indicate gradual
decoder refresh access points.

GET_MIN and GET_MAX can be used to determine the minimum and maximum
recovery_frame_cnt over which the encoder can implement GDR. GET_MIN and GET_MAX
can also be used to determine the minimum and maximum wSynchFrameInterval of the current
bSyncFrameType.
GET_MIN and GET_MAX may be used to determine if the device supports changes to
wSyncFrameInterval. If GET_MIN and GET_MAX return the same wSyncFrameInterval
value as GET_CUR, then the device does not support changes to this value.

Revision 1.5 August 9, 2012 124

 USB Device Class Definition for Video Devices

4.2.2.4.14 Long-Term Buffer Control

The EU_LTR_BUFFER_CONTROL provides control of encoder’s long term reference buffers.
When combined with the EU_LTR_PICTURE_CONTROL, this control enables outside
management over one or more of the encoder long term reference buffers. The host should check
the device’s long term buffer availability using this control before using the
EU_LTR_PICTURE_CONTROL.

The Long Term Reference controls defined in this specification support two different trust
models:
Trust Until (only notify when something has gone wrong)
Don’t Trust Until (notify all results, success or failure)
The host can decide which Trust Model to use by setting bTrustMode to the desired mode.
To implement a “Don’t Trust Until” solution, set bTrustMode to 0. If an LTR has been
validated by the host, then the associated position in the list of LTRs can be set to 1 using the
EU_LTR_VALIDATION_CONTROL. To implement a “Trust Until” model, set bTrustMode
to 1. If an LTR is confirmed as no longer valid by the host, then the associated position in the list
of LTRs can be set to 0 using the EU_LTR_VALIDATION_CONTROL.

Table 4-65: Long-Term Buffer Control

Control Selector EU_LTR_BUFFER_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_DEF, GET_INFO, GET_LEN,
GET_MAX
wLength 2
Offset Field Size Value Description

0 bNumHostContro 1 Number Number of Long-Term Reference Frames the host

lLTRBuffers can control.

1 bTrustMode 1 Number Trust mode for the LTR feature.

0 – For each inserted LTR, device sets associated
bit in bmValidLTRs to 0 (Don’t Trust Until)
1 – For each inserted LTR, device sets associated
bit in bmValidLTRs to 1 (Trust Until)

The EU_LTR_BUFFER_CONTROL allows for discovery and allocation of long term reference
(LTR) frames on the device. It also allows the host to set the device behavior when inserting new
LTR frames using EU_LTR_PICTURE_CONTROL. If the encoder on the device does not have
enough memory to enable long term reference frames at the current resolution, then the
GET_MAX shall return bNumHostControlLTRBuffers equal to 0. If the encoder allows the
host to manage the LTR buffers, it shall assign continuous index space starting from index 1.

Revision 1.5 August 9, 2012 125

 USB Device Class Definition for Video Devices

GET_MAX returns the maximum number of LTR buffers available for host control.

4.2.2.4.15 Long-Term Reference Picture Control

The EU_LTR_PICTURE_CONTROL tells the encoder to generate an LTR and place it at a
specific index in the LTR buffer.
If bTrustMode is 0, the device shall consider the LTR inserted at bPutAtPositionInLTRBuffer
as invalid until validated using EU_LTR_VALIDATION_CONTROL. This means that index
bPutAtPositionInLTRBuffer in bmValidLTRs must be 0 as soon as the new LTR is available.
If bTrustMode is 1, the device shall consider the LTR inserted at bPutAtPositionInLTRBuffer
as valid. This means that index bPutAtPositionInLTRBuffer in bmValidLTRs must be 1 as soon
as the new LTR is available.

Table 4-66: Long-Term Reference Picture Control

Control Selector EU_LTR_PICTURE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_DEF, GET_INFO, GET_LEN
wLength 2
Offset Field Size Value Description
0 bPutAtPositionIn 1 Number Next frame should be put at certain position in Long
LTRBuffer Term Reference Buffer (LTRB). This frame must be
in the temporal base layer. It must not be one of the
scheduled sync frames schedule by the
EU_SYNC_REF_FRAME_CONTROL control.

0 - Encoder is free to choose where to save the frame

inside the encoder controlled buffers. It cannot be
saved in any of the host controlled LTR Buffers.

1 : index 1 in the host controlled buffers

2 : index 2 in the host controlled buffers
…
M : position M (maximum) in the host controlled
buffers
Where M = bNumAppControlLTRBuffers
1 bLTRMode 1 Number Determines which frames can be used as references
when creating the new LTR.

0 - Encoder is restricted to valid host managed LTR

frames

1 - Encoder may use any valid host or encoder

controlled LTR frames.

Revision 1.5 August 9, 2012 126

 USB Device Class Definition for Video Devices

2 – Encoder may use any valid reference in the DPB

When bLTRMode is 0, the new LTR generated by this control must only reference host
controlled LTRs that have been validated. When bLTRMode is 1, the new LTR generated by
this control must only reference valid host or encoder controlled LTRs. When bLTRMode is 2,
the encoder is free to use any long or short term references it wishes when creating the new
LTR.. When a request for a new LTR frame is still pending, the device shall protocol STALL
any new requests to this control for new LTR frames with bRequestErrorCode = “Not Ready”,

4.2.2.4.16 Long-Term Reference Validation Control

The EU_LTR_VALIDATION_CONTROL instructs the encoder to update the list of valid LTRs
which may be used for subsequent LTR frames. For each bit in bValidLTRs, a value of 1 means
the LTR is valid and a value of 0 means the LTR is invalid. There are two different lists of LTR
buffers, those managed by the host and those managed by the encoder. This control validates or
invalidates both. The bmValidLTRs bitmask in this control can be used to indicate which host
controlled LTRs can be used by EU_LTR_PICTURE_CONTROL. The bmValidLTRs bitmask
may also be used to limit which LTRs should be used as references.
bmValidLTRs should be initialized to all zeros.

Table 4-67: Long-Term Reference Validation Control

Control Selector EU_LTR_VALIDATION_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_DEF, GET_INFO, GET_LEN
wLength 2
Offset Field Size Value Description
0 bmValidLTRs 2 Bitmap A list of reference LTRs that may be used to generate
subsequent frames.

D0: LTR frame with index = 0

D1 - LTR frame with index = 1
D2 - LTR frame with index = 2
D3 - LTR frame with index = 3
Etc.

4.2.2.4.17 SEI Messages Control

The control allows discovery and control of the H.264 SEI_payloads, as specified in Annex D of
the H.264 specification.

Revision 1.5 August 9, 2012 127

 USB Device Class Definition for Video Devices

Table 4-68: SEI Message Control

Control Selector EU_SEI_PAYLOADTYPE_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO, GET_DEF, GET_MAX,
GET_MIN
wLength 8
Offset Field Size Value Description
0 bmSEIMessages 8 Bitmap A bit set indicates the associated SEI type is
enabled.
D0: buffering_period
D1: pic_timing
D2: pan_scan_rect
D3: filler_payload
D4: user_data_registered_itu_t_t35
D5: user_data_unregistered
D6: recovery_point
D7: dec_ref_pic_marking_repetition
D8: spare_pic
D9: scene_info
D10: sub_seq_info
D11: sub_seq_layer_characteristics
D12: sub_seq_characteristics
D13: full_frame_freeze
D14: full_frame_freeze_release
D15: full_frame_snapshot
D16: progressive_refinement_segment_start
D17: progressive_refinement_segment_end
D18: motion_constrained_slice_group_set
D19: film_grain_characteristics
D20: deblocking_filter_display_preference
D21: stereo_video_info
D22: post_filter_hint
D23: tone_mapping_info
D24: scalability_info
D25: sub_pic_scalable_layer
D26: non_required_layer_rep
D27: priority_layer_info
D28: layers_not_present
D29: layer_dependency_change
D30: scalable_nesting
D31: base_layer_temporal_hrd
D32: quality_layer_integrity_check
D33: redundant_pic_property
D34: tl0_dep_rep_index
D35: tl_switching_point

Revision 1.5 August 9, 2012 128

 USB Device Class Definition for Video Devices

D36: parallel_decoding_info
D37: mvc_scalable_nesting
D38: view_scalability_info
D39: multiview_scene_info
D40: multiview_acquisition_info
D41: non_required_view_component
D42: view_dependency_change
D43: operation_points_not_present
D44: base_view_temporal_hrd
D45: frame_packing_arrangement
D63..D46: Reserved. Set to 0
Each bit in bmSEIMessages represents a different SEI message and when the associated bit is 1
the SEI message is enabled. Multiple types of SEI messages can be enabled/disabled
simultaneously with this control. Bits set in the GET_CUR response will indicate the SEI
messages that are currently enabled. Bits set in the GET_MAX response will indicate the SEI
messages that the device supports. Bits set in the GET_MIN response will indicate which SEI
messages are enabled and cannot be disabled by the host.

4.2.2.4.18 Priority Control

This control can be used to set priority for video payload layers.

Table 4-69: Priority Control

Control Selector EU_PRIORITY_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO, GET_LEN
wLength 1
Offset Field Size Value Description
0 bPriority 1 Number The value of priority for the scalable layer specified by
the current wLayerOrViewID.

On a GET_DEF request, the device shall return values based on bUsage.

4.2.2.4.19 Start or Stop Layer Control

This control is used to start or stop streaming of the current layer(s). By default all layers will be
streamed. The host may stop individual layers (and their dependents) from streaming by using
this control. Once stopped, this control may be used to restart them.

Table 4-70: Start or Stop Layer Control

Control Selector EU_START_OR_STOP_LAYER_CONTROL
Mandatory SET_CUR, GET_CUR, GET_INFO, GET_LEN
Requests
wLength 1
Offset Field Size Value Description
0 bUpdate 1 Number 0: Stop streaming the current layer and all layers

Revision 1.5 August 9, 2012 129

 USB Device Class Definition for Video Devices

depending on it.
1: Start streaming the current layer (and all layers it
depends on) once the device starts streaming. If this
control is not issued before streaming is enabled
(before SET_INTERFACE), the encoder shall
stream every layer when streaming is enabled
GET_CUR returns the current state of the layer (0
if stopped, or 1 if started).

4.2.2.4.20 Level IDC Control

This control is used to set the level_idc of the video stream. The default value of bLevelIDC in
this control is initialized to the value of bLevelIDC from the video Frame descriptor selected as
part of probe and commit. The host can issue SET_CUR requests to this control to reduce the
maximum level_idc the encoder shall use, i.e. to reduce the minimum level_idc that the decoder
will need to support.
Table 4-71: Level IDC Control
Control Selector EU_LEVEL_IDC_LIMIT_CONTROL
Mandatory SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_DEF
Requests
wLength 1
Offset Field Size Value Description
0 bLevelIDC 1 Numbe The level, as specified by the level_idc flag (9, 10, 11, 12,
r 13, 20, 21, 22, 30, 31, 32, 40, 41, 42, etc). For example:
0x1F: Level 3.1.
0x28: Level 4.0.

Note that this should indicate the minimum level that

supports the resolution, maximum bit rate and CPB size
set for the stream. For a multi-layer stream this refers to
the highest enhancement layer. For multi-view streams
this refers to all views.

Note that the host must adjust the current average bit rate and current CBP size prior to changing
the level_idc to guarantee those do not violate the new level_idc .
When this value changes, the device must adjust the values returned by
EU_AVERAGE_BITRATE_CONTROL(GET_MAX) and
EU_CPB_SIZE_CONTROL(GET_MAX) to satisfy the new level_idc.

4.2.2.4.21 Error Resiliency Control

The Error Resiliency control is used to set error resiliency features on the video encoder.
Because error resiliency features are specific to each encoder, the only field in this control is a 16
bit bitmap that allows up to 16 different features to be turned on or off per video payload.

Revision 1.5 August 9, 2012 130

 USB Device Class Definition for Video Devices

Table 4-72: Error Resiliency Control

Control Selector EU_ERROR_RESILIENCY_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_DEF, GET_RES
wLength 2
Offset Field Size Value Description
0 bmErrorResiliencyFeatures 2 Bitmap Payload format specific bitmap of
available error resiliency features.

In response to a GET_RES response, device shall set the bits for supported error resiliency
features to 1. All other bits should be set to 0. In response to a GET_DEF request device shall set
the bits to 1 for the tools that are enabled in the device default configuration. All other bits shall
be set to 0. While the exact meaning of bmErrorResiliencyFeatures is established in each
payload specification, several possible examples are given below.

For VP8, bmErrorResiliencyFeatures could have the following assignment:

D0: Random Macroblock Intra Refresh.
D1: Frame-level error resiliency.
D2: Partition-level error resiliency.
D15-D3: Reserved. Set to 0.
For H.264, bmErrorResiliencyFeatures could have the following assignment:
D0: Random Macroblock Intra Refresh.
D1: Flexible Macroblock Ordering.
D2: Arbitrary Slice Ordering.
D3: Redundant Slices.
D4: Data Partitioning.
D15-D5: Reserved. Set to 0.

4.2.2.5 Extension Unit Control Requests

These requests are used to set or read a video Control within an Extension Unit.
Table 4-73 Extension Unit Control Requests
bmRequestType bRequest wValue wIndex wLength Data
00100001 SET_CUR CS Extension Length of Parameter
Unit ID and parameter block
10100001 GET_CUR Interface block
GET_MIN
GET_MAX
GET_RES
GET_DEF
GET_LEN
GET_INFO

Revision 1.5 August 9, 2012 131

 USB Device Class Definition for Video Devices

The bRequest field indicates which attribute the request is manipulating. The MIN, MAX, and
RES attributes are not supported for the Set request.

The wValue field specifies the Control Selector (CS) in the high byte and zero in the low byte.
The Control Selector indicates which vendor-defined control within the Extension Unit that this
request is manipulating. If the request specifies an unknown or unsupported CS to that Unit, the
control pipe must indicate a stall. However, if the request specifies an available control, the
request should succeed.

The range of CS values supported by the Extension Unit is dictated by the number of controls
specified by the bNumControls field in the Extension Unit descriptor. See section 3.7.2.7,
"Extension Unit Descriptor". The range shall be [1..bNumControls].

The GET_LEN request queries for the length of the parameter block of the specified control.
When issuing the GET_LEN request, the wLength field shall always be set to a value of 2 bytes.
The result returned shall be the length specified for all other requests on the same control.

All controls supported by the Extension Unit must support the following requests:
GET_CUR, GET_MIN, GET_MAX, GET_RES, GET_INFO, GET_DEF, GET_LEN.

The following request(s) are optional, depending on the control usage and behavior:
SET_CUR

All Extension Unit controls are vendor-defined. The vendor must provide the relevant host
software to program these controls. The generic host driver will not have knowledge of the
control semantics, but acts as a control transport between the vendor-provided host software and
the device.

However, by using the GET_LEN request, the host driver would be able to query the length and
raw data stored in the vendor-defined controls. While it would not be able to interpret this data, it
would be capable of saving and restoring these control settings if required.

4.3 VideoStreaming Requests

VideoStreaming requests can be directed either to the VideoStreaming interface or to the
associated video-data endpoint, depending on the location of the Control to be manipulated.

4.3.1 Interface Control Requests

These requests are used to set or read an attribute of an interface Control inside a particular
VideoStreaming interface of the video function.

Revision 1.5 August 9, 2012 132

 USB Device Class Definition for Video Devices

Table 4-74 Interface Control Requests inside a Particular VideoStreaming Interface

bmRequestType bRequest wValue wIndex wLength Data
00100001 SET_CUR CS Zero and Length of Parameter
Interface parameter block
10100001 GET_CUR block
GET_MIN
GET_MAX
GET_RES
GET_DEF
GET_LEN
GET_INFO

The bRequest field indicates which attribute the request is manipulating.

The wValue field specifies the Control Selector (CS) in the high byte, and the low byte must be
set to zero. The CS indicates the type of Control that this request is manipulating. If the request
specifies an unknown CS to that endpoint, the control pipe must indicate a stall.

The VideoStreaming interface controls allow the host software to query and set parameters
related to the video stream format and the video stream encoder. These parameters include the
format, frame size and frame rate of the video stream, as well as the format and frame size of still
images captured by the device that are associated with the video stream. For devices that support
host-adjustable video stream encoder parameters, controls allowing the adjustment of the key
frame rate and compression quality, among other parameters, are also supported. Only Stream
Error Code Control supports interrupt with VideoStreaming interface.

4.3.1.1 Video Probe and Commit Controls

The streaming parameters selection process is based on a shared negotiation model between the
host and the video streaming interface, taking into account the following features:
 shared nature of the USB
 interdependency of streaming parameters
 payload independence
 modification of streaming parameters during streaming

This negotiation model is supported by the Video Probe and Commit controls. The Probe control
allows retrieval and negotiation of streaming parameters. When an acceptable combination of
streaming parameters has been obtained, the Commit control is used to configure the hardware
with the negotiated parameters from the Probe control.

Additional Encoding Units may used to finalize the configuration of the video streaming
interface after Probe and Commit but before streaming starts. This hybrid model of Descriptor

Revision 1.5 August 9, 2012 133

 USB Device Class Definition for Video Devices

plus Encoding Unit was chosen as the best model to navigate the complex space of encoder
configuration.
Table 4-75 Video Probe and Commit Controls
Control Selector VS_PROBE_CONTROL
VS_COMMIT_CONTROL
Mandatory Requests See tables below
wLength 48
Offse Field Size Value Description
t
0 bmHint 2 Bitmap Bitfield control indicating to the function
what fields shall be kept fixed (indicative
only):
D0: dwFrameInterval
D1: wKeyFrameRate
D2: wPFrameRate
D3: wCompQuality
D4: wCompWindowSize
D15..5: Reserved (0)

The hint bitmap indicates to the video

streaming interface which fields shall be
kept constant during stream parameter
negotiation. For example, if the selection
wants to favor frame rate over quality,
the dwFrameInterval bit will be set (1).

This field is set by the host, and is read-

only for the video streaming interface.
2 bFormatIndex 1 Number Video format index from a Format
descriptor for this video interface.

Select a specific video stream format by

setting this field to the one-based index
of the associated Format descriptor. To
select the first format defined by a
device, a value one (1) is written to this
field. This field must be supported even
if only one video format is supported by
the device.

This field is set by the host.

Revision 1.5 August 9, 2012 134

 USB Device Class Definition for Video Devices

3 bFrameIndex 1 Number Video frame index from a Frame

descriptor.

This field selects the video frame

resolution from the array of resolutions
supported by the selected stream. The
index value ranges from 1 to the number
of Frame descriptors following a
particular Format descriptor. For frame-
based formats, this field must be
supported even if only one video frame
index is supported by the device.

For video payloads with no defined

Frame descriptor, this field shall be set to
zero (0).

This field is set by the host.

4 dwFrameInterval 4 Number Frame interval in 100 ns units.

This field sets the desired video frame

interval for the selected video stream and
frame index. The frame interval value is
specified in 100 ns units. The device
shall support the setting of all frame
intervals reported in the Frame
Descriptor corresponding to the selected
Video Frame Index. For frame-based
formats, this field must be implemented
even if only one video frame interval is
supported by the device.

When used in conjunction with an IN

endpoint, the host shall indicate its
preference during the Probe phase. The
value must be from the range of values
supported by the device.

When used in conjunction with an OUT

endpoint, the host shall accept the value
indicated by the device.
8 wKeyFrameRate 2 Number Key frame rate in key-frame per video-
frame units.

Revision 1.5 August 9, 2012 135

 USB Device Class Definition for Video Devices

This field is only applicable to sources

(and formats) capable of streaming video
with adjustable compression parameters.
Use of this control is at the discretion of
the device, and is indicated in the VS
Input or Output Header descriptor.

The Key Frame Rate field is used to

specify the compressor’s key frame rate.
For example, if one of every ten encoded
frames in a video stream sequence is a
key frame, this control would report a
value of 10. A value of 0 indicates that
only the first frame is a key frame.

When used in conjunction with an IN

endpoint, the host shall indicate its
preference during the Probe phase. The
value must be from the range of values
supported by the device.

When used in conjunction with an OUT

endpoint, the host shall accept the value
indicated by the device.
10 wPFrameRate 2 Number PFrame rate in PFrame/key frame units.

This field is only applicable to sources

(and formats) capable of streaming video
with adjustable compression parameters.
Use of this control is at the discretion of
the device, and is indicated in the VS
Input or Output Header descriptor.

The P Frame Rate Control is used to

specify the number of P frames per key
frame. As an example of the relationship
between the types of encoded frames,
suppose a key frame occurs once in
every 10 frames, and there are 3 P frames
per key frame. The P frames will be
spaced evenly between the key frames.
The other 6 frames, which occur between
the key frames and the P frames, will be
bi-directional (B) frames.

Revision 1.5 August 9, 2012 136

 USB Device Class Definition for Video Devices

When used in conjunction with an IN

endpoint, the host shall indicate its
preference during the Probe phase. The
value must be from the range of values
supported by the device.

When used in conjunction with an OUT

endpoint, the host shall accept the value
indicated by the device.
12 wCompQuality 2 Number Compression quality control in abstract
units 1 (lowest) to 10000 (highest).

This field is only applicable to sources

(and formats) capable of streaming video
with adjustable compression parameters.
Use of this field is at the discretion of the
device, and is indicated in the VS Input
or Output Header descriptor.

This field is used to specify the quality of

the video compression. Values for this
property range from 1 to 10000 (1
indicates the lowest quality, 10000 the
highest). The resolution reported by this
control will determine the number of
discrete quality settings that it can
support.

When used in conjunction with an IN

endpoint, the host shall indicate its
preference during the Probe phase. The
value must be from the range of values
supported by the device.

When used in conjunction with an OUT

endpoint, the host shall accept the value
indicated by the device.
14 wCompWindowSize 2 Number Window size for average bit rate control.

This field is only applicable to sources

(and formats) capable of streaming video
with adjustable compression parameters.
Use of this control is at the discretion of

Revision 1.5 August 9, 2012 137

 USB Device Class Definition for Video Devices

the device, and is indicated in the VS

Input or Output Header descriptor.

The Compression Window Size Control

is used to specify the number of encoded
video frames over which the average size
cannot exceed the specified data rate. For
a window of size n, the average frame
size of any consecutive n frames will not
exceed the stream's specified data rate.
Individual frames can be larger or
smaller.

For example, if the data rate has been set

to 100 kilobytes per second (KBps) on a
10 frames per second (fps) movie with a
compression window size of 10, the
individual frames can be any size, as
long as the average size of a frame in any
10-frame sequence is less than or equal
to 10 kilobytes.

When used in conjunction with an IN

endpoint, the host shall indicate its
preference during the Probe phase. The
value must be from the range of values
supported by the device.

When used in conjunction with an OUT

endpoint, the host shall accept the value
indicated by the device.
16 wDelay 2 Number Internal video streaming interface latency
in ms from video data capture to
presentation on the USB.

When used in conjunction with an IN

endpoint, this field is set by the device
and read only from the host.

When used in conjunction with an OUT

endpoint, this field is set by the host and
read only from the device.

18 dwMaxVideoFrameSize 4 Number Maximum video frame or codec-specific

Revision 1.5 August 9, 2012 138

 USB Device Class Definition for Video Devices

segment size in bytes.

For frame-based formats, this field

indicates the maximum size of a single
video frame. When streaming simulcast
this number reflects the maximum video
frame size of the negotiated Frame
descriptor. For frame-based formats, this
field must be supported.

For stream-based formats, and when this

behavior is enabled via the
bmFramingInfo field (below), this field
indicates the maximum size of a single
codec-specific segment. The sender is
required to indicate a segment boundary
via the FID bit in the payload header.
This field is ignored (for stream-based
formats) if the bmFramingInfo bits are
not enabled.

When used in conjunction with an IN

endpoint, this field is set by the device
and read only from the host.

When used in conjunction with an OUT

endpoint, this field is set by the host and
read only from the device.
22 dwMaxPayloadTransfer 4 Number Specifies the maximum number of bytes
Size that the device can transmit or receive in
a single payload transfer. This field must
be supported.

This field is set by the device and read

only from the host. Some host
implementations restrict the maximum
value permitted for this field. The host
shall avoid overshoot of single payload
transfer size by reconfiguring the device.
(e.g. by updating bitrates, resolutions
etc.)
26 dwClockFrequency 4 Number The device clock frequency in Hz for the
specified format. This will specify the
units used for the time information fields

Revision 1.5 August 9, 2012 139

 USB Device Class Definition for Video Devices

in the Video Payload Headers in the data

stream.

This parameter is set by the device and

read only from the host.
30 bmFramingInfo 1 Bitmap Bitfield control supporting the following
values:

D0: If set to 1, the Frame ID (FID)

field is required in the Payload
Header (see description of D0
in section 2.4.3.3, “Video and
Still Image Payload Headers”).
The sender is required to toggle
the Frame ID at least every
dwMaxVideoFrameSize bytes
(see above).
D1: If set to 1, indicates that the End
of Frame (EOF) field may be
present in the Payload Header
(see description of D1 in section
2.4.3.3, “Video and Still Image
Payload Headers”). It is an error
to specify this bit without also
specifying D0.
D2: If set to 1, indicates that the End
of Slice (EOS) field may be
present in the Payload Header.
It is an error to specify this bit
without also specifying D0.
D7..3: Reserved (0)

This control indicates to the function

whether payload transfers will contain
out-of-band framing information in the
Video Payload Header (see section
2.4.3.3, “Video and Still Image Payload
Headers”).

For known frame-based formats (e.g.,

MJPEG, Uncompressed, DV), this field
is ignored.

For known stream-based formats, this

Revision 1.5 August 9, 2012 140

 USB Device Class Definition for Video Devices

field allows the sender to indicate that it

will identify segment boundaries in the
stream, enabling low-latency buffer
handling by the receiver without the
overhead of parsing the stream itself.

When used in conjunction with an IN

endpoint, this control is set by the device,
and is read-only from the host.

When used in conjunction with an OUT

endpoint, this parameter is set by the
host, and is read-only from the device.
31 bPreferedVersion 1 Number The preferred payload format version
supported by the host or device for the
specified bFormatIndex value.

This parameter allows the host and

device to negotiate a mutually agreed
version of the payload format associated
with the bFormatIndex field. The host
initializes this and the following
bMinVersion and bMaxVersion fields
to zero on the first Probe Set. Upon
Probe Get, the device shall return its
preferred version, plus the minimum and
maximum versions supported by the
device (see bMinVersion and
bMaxVersion below). The host may
issue a subsequent Probe Set/Get
sequence to specify its preferred version
(within the ranges returned in
bMinVersion and bMaxVersion from
the initial Probe Set/Get sequence). The
host is not permitted to alter the
bMinVersion and bMaxVersion values.

This field will support up to 256 (1-255)

versions of a single payload format. The
version number is drawn from the minor
version of the Payload Format
specification. For example, version 1.2 of
a Payload Format specification would
result in a value of 2 for this parameter.

Revision 1.5 August 9, 2012 141

 USB Device Class Definition for Video Devices

32 bMinVersion 1 Number The minimum payload format version

supported by the device for the specified
bFormatIndex value.

This value is initialized to zero by the

host and reset to a value in the range of 1
to 255 by the device. The host is not
permitted to modify this value (other
than to restart the negotiation by setting
bPreferredVersion, bMinVersion and
bMaxVersion to zero).
33 bMaxVersion 1 Number The maximum payload format version
supported by the device for the specified
bFormatIndex value.

This value is initialized to zero by the

host and reset to a value in the range of 1
to 255 by the device. The host is not
permitted to modify this value (other
than to restart the negotiation by setting
bPreferredVersion, bMinVersion and
bMaxVersion to zero).
34 bUsage 1 Number Current bUsage:
1-8: Real-time modes
9-16: Broadcast modes
17-24: File storage modes
25 – 31: Multiview modes
32-255: Reserved

This bitmap enables features reported by

the bmUsages field of the Video Frame
Descriptor.

For temporally encoded video formats,

this field must be supported, even if the
device only supports a single value for
bUsage.
35 bBitDepthLuma 1 Number Represents bit_depth_luma_minus8 + 8,
which must be the same as
bit_depth_chroma_minus8 + 8.
36 bmSettings 1 Bitmap A bitmap of flags that is used to discover
and control specific features of a
temporally encoded video stream. When
it is supported, it is defined in the

Revision 1.5 August 9, 2012 142

 USB Device Class Definition for Video Devices

associated Payload specification. This

bitmap enables features reported by the
bmCapabilities field of the Video Frame
Descriptor.

For temporally encoded video formats,

this field must be supported.
37 bMaxNumberOfRefFra 1 Number Host indicates the maximum number of
mesPlus1 frames stored for use as references.

38 bmRateControlModes 2 Number This field contains 4 subfields, each of

which is a 4 bit number.

It enables features reported by the

bmSupportedRateControlModes field
of the Video Format Descriptor.

Each 4 bit number indicates the rate-

control mode for a stream of encoded
video. If the video payload does not
support rate control, this entire field
should be set to 0.
bmRateControlModes supports up to
four simulcast streams. For simulcast
transport the number of streams is
inferred from the bmLayoutPerStream
field. Otherwise, the number of streams
is 1.
D3-D0: Rate-control mode for the first
simulcast stream (with stream_id=0.)
D7-D4: Rate-control mode for the
second simulcast stream (with
stream_id=1).
D11-D8: Rate control mode for the third
simulcast stream (with stream_id=2).
D15-D12: Rate control mode for the
fourth simulcast stream (with
stream_id=3.)
When bmRateControlModes is non-
zero, each 4-bit subfield can take one of
the following values:

Revision 1.5 August 9, 2012 143

 USB Device Class Definition for Video Devices

0: Not applicable, because this stream is

non-existent.
1: VBR with underflow allowed
2: CBR
3: Constant QP
4: Global VBR, underflow allowed
5: VBR without underflow
6: Global VBR without underflow
7-15: Reserved

For temporally encoded video formats,

this field must be supported, even if the
device only supports a single value for
bmRateControlModes.
40 bmLayoutPerStream 8 Number This field contains 4 subfields, each of
which is a 2 byte number.

For simulcast transport, this field

indicates the specific layering structure
for each stream, up to four simulcast
streams. For a single, multi-layer stream,
only the first two bytes are used. For a
single stream with no enhancement
layers, this field shall be set to 0. See
individual payload specification for how
to interpret each 2 byte sub-field.

For temporally encoded video formats,

this field must be supported.

4.3.1.1.1 Probe and Commit Operational Model

Unsupported fields shall be set to zero by the device. Fields left for streaming parameters
negotiation shall be set to zero by the host. For example, after a SET_CUR request initializing
the FormatIndex and FrameIndex, the device will return the new negotiated field values for the
supported fields when retrieving the Probe control GET_CUR attribute.
In order to avoid negotiation loops, the device shall always return streaming parameters with
decreasing data rate requirements. Unsupported streaming parameters shall be reset by the
streaming interface to supported values according to the negotiation loop avoidance rules. This
convention allows the host to cycle through supported values of a field.

During Probe and Commit, the following fields, if supported, shall be negotiated in order of
decreasing priority:
 bFormatIndex

Revision 1.5 August 9, 2012 144

 USB Device Class Definition for Video Devices

 bFrameIndex
 dwMaxPayloadTransferSize
 bUsage
 bmLayoutPerStream
 Fields set to zero by the host with their associated bmHint bit set to 1
 All the remaining fields set to zero by the host
For simplicity when streaming temporally encoded video, the required bandwidth for each
streaming interface shall be estimated using the maximum bit rate for the selected
profile/resolution and the number of simulcast streams. The USB bandwidth reserved shall be the
calculated by the host as the advertised dwMaxBitRate from the selected Frame Descriptor
multiplied times the number of simulcast streams as defined in the bmLayoutPerStream field.
The interface descriptor for the video function should have multiple alternate settings that
support the required bandwidths calculated in the manner above.

The following table describes VS_PROBE_CONTROL request attributes.

Table 4-76 VS_PROBE_CONTROL Requests
Attribute Description
GET_CUR Returns the current state of the streaming interface. All supported fields
set to zero will be returned with an acceptable negotiated value.
Prior to the initial SET_CUR operation, the GET_CUR state is
undefined. This request shall stall in case of negotiation failure.
GET_MIN Returns the minimum value for negotiated fields.
GET_MAX Returns the maximum value for negotiated fields.
GET_RES Return the resolution of each supported field in the Probe/Commit data
structure.
GET_DEF Returns the default value for the negotiated fields.
GET_LEN Returns the length of the Probe data structure.
GET_INFO Queries the capabilities and status of the Control. The value returned for
this request shall have bits D0 and D1 each set to one (1), and the
remaining bits set to zero (0) (see section 4.1.2, “Get Request”).
SET_CUR Sets the streaming interface Probe state. This is the attribute used for
stream parameter negotiation.

This request shall protocol STALL in the case where the device would
be place into an unsupported state or the case where value for a
negotiated field is out of range. For exact errors to register, see section
4.2.1.2 "Request Error Code Control".

The following table describes VS_COMMIT_CONTROL request attributes.

Revision 1.5 August 9, 2012 145

 USB Device Class Definition for Video Devices

Table 4-77 VS_COMMIT_CONTROL Requests

Attribute Description
GET_CUR Returns the current state of the streaming interface. Prior to the initial
SET_CUR operation, the GET_CUR state is undefined.
GET_MIN Not specified.
GET_MAX Not specified.
GET_RES Not specified.
GET_DEF Not specified.
GET_LEN Returns the length of the Commit data structure.
GET_INFO Queries the capabilities and status of the Control. The value returned for
this request shall have bits D0 and D1 each set to one (1), and the
remaining bits set to zero (0) (see section 4.1.2, “Get Request”).
SET_CUR Sets the device state. This sets the active device state. The field values
must be the result of a successful VS_PROBE_CONTROL(GET_CUR)
request. This request shall protocol STALL in case an unsupported
state is specified.

This request shall protocol STALL in the case where the device would
be place into an unsupported state or the case where value for a
negotiated field is out of range. For exact errors to register, see section
4.2.1.2 "Request Error Code Control".

4.3.1.1.2 Stream Negotiation Examples

Successful USB isochronous bandwidth negotiation
Host issues a VS_PROBE_CONTROL(SET_CUR) which is accepted and populated by the
device in a VS_PROBE_CONTROL(GET_CUR). After this single round of negotiation the host
sends a VS_COMMIT_CONTROL(SET_CUR) using the field values of VS
_PROBE_CONTROL returned by the device in VS_PROBE_CONTROL(GET_CUR). After the
successful Commit, the host starts streaming by selecting the first alternate interface.

Revision 1.5 August 9, 2012 146

 USB Device Class Definition for Video Devices

Host Device

VS_PROBE_CONTROL(SET_CUR)

VS_PROBE_CONTROL(GET_CUR)

VS_COMMIT_CONTROL(SET_CUR)

SET_INTERFACE(1)

Figure 4-5 Successful USB Isochronous Bandwidth Negotiation

Revision 1.5 August 9, 2012 147

 USB Device Class Definition for Video Devices

4.3.1.1.2.1 USB isochronous bandwidth negotiation failure

In this case the host requests video settings and an alternate interface that are incompatible. This
could happen if the alternate interface does not support the data throughput required by the video
settings. This request shall protocol STALL with “Wrong State”. The host tries again with a set
of features that demand less USB bandwidth, and this time succeeds.

Host Device

VS_PROBE_CONTROL(SET_CUR)

VS_PROBE_CONTROL(GET_CUR)

VS_COMMIT_CONTROL(SET_CUR)

SET_INTERFACE(1)

(Fail)
VS_PROBE_CONTROL(SET_CUR)

VS_PROBE_CONTROL(GET_CUR)

VS_COMMIT_CONTROL(SET_CUR)

SET_INTERFACE (1)

Figure 4-6 Failed USB Isochronous Bandwidth Negotiation

Revision 1.5 August 9, 2012 148

 USB Device Class Definition for Video Devices

Dynamic stream settings modification while streaming

In the example below, the host changes the streaming parameters bCompQuality and
bCompWindowSize after streaming has started by issuing an updated
VS_COMMIT_CONTROL(SET_CUR). This operation was successful because the new settings
did not exceed the negotiated USB bandwidth.

Host Device

VS_PROBE_CONTROL(SET_CUR)

VS_PROBE_CONTROL(GET_CUR)

VS_COMMIT_CONTROL(SET_CUR)

SET_INTERFACE(1)

Streaming

VS_COMMIT_CONTROL(SET_CUR)

Figure 4-7 Dynamic Stream Settings Modification while Streaming

Revision 1.5 August 9, 2012 149

 USB Device Class Definition for Video Devices

4.3.1.2 Video Still Probe Control and Still Commit Control

These still-image controls are required for video functions supporting method 2 or 3 for still-
image retrieval.
Table 4-78 Video Still Probe Control and Still Commit Control
Control Selector VS_STILL_PROBE_CONTROL
VS_STILL_COMMIT_CONTROL
Mandatory Requests See tables below
wLength 11
Offset Field Size Value Description
0 bFormatIndex 1 Number Video format index from a Format
descriptor.
A specific still-image format is
selected by setting this field to the
one-based index for the associated
Format descriptor. To select the first
format defined by a device, a value
of 1 is written to this control.
1 bFrameIndex 1 Number Video frame index from a Frame
descriptor.
This field selects the still-image
frame resolution from the array of
resolutions supported by the selected
still-image format. The index value
ranges from one to the number of
Still Image Size Patterns reported by
the selected Still Image Frame
descriptor.
2 bCompressionIndex 1 Number Compression index from a Frame
descriptor.
This field selects the still-image
frame compression from the array of
compression patterns supported by
the selected still-image format. The
index value ranges from one to the
number of Still Image Compression
Patterns reported by the selected Still
Image Frame descriptor.
3 dwMaxVideoFrameSize 4 Number Maximum still image size in bytes.

This field indicates the maximum

size of a single still image.
This parameter is set by the device

Revision 1.5 August 9, 2012 150

 USB Device Class Definition for Video Devices

and read only from the host.

7 dwMaxPayloadTransferSize 4 Number Specifies the maximum number of
bytes that the device can transmit or
receive in a single payload transfer.

The following table describes VS_STILL_PROBE_CONTROL request attributes:

Table 4-79 VS_STILL_PROBE_CONTROL Requests
Attribute Description
GET_CUR Returns the current state of the device. Prior to the initial SET_CUR
operation, the GET_CUR state is undefined.
GET_MIN Returns the minimum value for negotiated fields.
GET_MAX Returns the maximum value for negotiated fields.
GET_RES Not specified.
GET_DEF Returns the default value for negotiated fields.
GET_LEN Returns the length of the Probe data structure.
GET_INFO Queries the capabilities and status of the Control. The value returned for
this request shall have bits D0 and D1 each set to one (1), and the
remaining bits set to zero (0) (see section 4.1.2, “Get Request”).
SET_CUR Sets the streaming interface state. This is the state used for stream
parameters negotiations.

The following table describes VS_STILL_COMMIT_CONTROL request attributes:

Table 4-80 VS_STILL_COMMIT_CONTROL Requests
Attribute Description
GET_CUR Returns the current state of the device. After device configuration, this
state is undefined.
GET_MIN Not specified.
GET_MAX Not specified.
GET_RES Not specified.
GET_DEF Not specified.
GET_LEN Returns the length of the Commit data structure.
GET_INFO Queries the capabilities and status of the Control. The value returned for
this request shall have bits D0 and D1 each set to one (1), and the
remaining bits set to zero (0) (see section 4.1.2, “Get Request”).
SET_CUR Sets the device state. This sets the active device state; the field values
must be the result of a VS_STILL_PROBE_CONTROL(GET_CUR)
request. When the associated still-image pipe is active, this attribute
cannot be used and the control pipe shall indicate a stall for this request.

Revision 1.5 August 9, 2012 151

 USB Device Class Definition for Video Devices

4.3.1.3 Synch Delay Control

The purpose for Synch Delay Control is to dynamically synchronize multiple video streams from
a device to host, or from multiple devices to the host to compensate for differing latencies among
multiple streams. Latency is the internal delay of the source from acquisition to data delivery on
the bus.

Only those devices that are capable of video streaming with adjustable delay latency parameters
support this control.

The Control is used to notify the video application buffer memory manager on the device to
control an internal latency by controlling output timing of the video data to its endpoint.

It is the responsibility of the host (video sink) to synchronize streams by scheduling the
rendering of samples at the correct moment, taking into account the internal delays of all media
streams being rendered.
Table 4-81 Synch Delay Control
Control Selector VS_SYNCH_DELAY_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_RES,
GET_INFO, GET_DEF
wLength 2
Offset Field Size Value Description
0 wDelay 2 Number Delay from the time that the packet
should be sent. wDelay is expressed in
microsecond units.

4.3.1.4 Still Image Trigger Control

This control notifies the device to begin sending still-image data over the relevant isochronous or
bulk pipe. A dedicated still-image bulk pipe is only used for method 3 of still image capture.
This control shall only be set while streaming is occurring, and the hardware shall reset it to the
“Normal Operation” mode after the still image has been sent. This control is only required if the
device supports method 2 or method 3 of still-image retrieval. See section 2.4.2.4 “Still Image
Capture”.
Table 4-82 Still Image Trigger Control
Control Selector VS_STILL_IMAGE_TRIGGER_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO
wLength 1
Offset Field Size Value Description
0 bTrigger 1 Number The setting for the Still Image Trigger
Control:
0: Normal operation.
1: Transmit still image.

Revision 1.5 August 9, 2012 152

 USB Device Class Definition for Video Devices

2: Transmit still image via dedicated bulk

pipe.
3: Abort still image transmission.

4.3.1.5 Generate Key Frame Control

This control is only supported by devices capable of streaming video with adjustable
compression parameters, and support for this control is indicated in the VideoStreaming Header
descriptor.

The Generate Key Frame Control is used to notify the video encoder on the device to generate a
key frame in the device stream at its earliest opportunity. After the key frame has been generated,
the device shall reset the control to the “Normal Operation” mode. This control is only applicable
to video formats that support temporal compression (such as MPEG-2 Video), and while
streaming is occurring. In all other cases, the device shall respond to requests by indicating a stall
on the control pipe.
Table 4-83 Generate Key Frame Control
Control Selector VS_GENERATE_KEY_FRAME_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_INFO
wLength 1
Offset Field Size Value Description
0 bGenerateKeyFrame 1 Number The setting for the attribute of the
addressed Generate Key Frame control:
0: Normal operation
1: Generate Key Frame

4.3.1.6 Update Frame Segment Control

This control is only supported by devices capable of streaming video with adjustable
compression parameters, and support for this control is indicated in the VideoStreaming Header
descriptor.

The Update Frame Segment Control is used to notify the video encoder on the device to encode
the specified range of video frame segments with intra coding (no dependency on surrounding
frames) at its earliest opportunity. A video frame segment corresponds to a group of macroblocks
that can be decoded independently, such as a slice in MPEG Video, or a Group of Blocks in
H.26x Video. This control is only applicable to video formats that support the concept of a video
frame segment, and while streaming is occurring. In all other cases, the device shall respond to
requests by indicating a stall on the control pipe.

The device will indicate the number of frame segments that it supports through the GET_MAX
request, for which the device will indicate the maximum frame segment index supported in both

Revision 1.5 August 9, 2012 153

 USB Device Class Definition for Video Devices

the bStartFrameSegment and bEndFrameSegment fields. The minimum value for these fields
shall always be zero. The resolution for this control shall always be set to 1.
Table 4-84 Update Frame Segment Control
Control Selector VS_UPDATE_FRAME_SEGMENT_CONTROL
Mandatory Requests SET_CUR, GET_CUR, GET_MIN, GET_MAX, GET_RES,
GET_INFO, GET_DEF
wLength 2
Offset Field Size Value Description
0 bStartFrameSegment 1 Number The zero-based index of the first frame
segment in the range to update
1 bEndFrameSegment 1 Number The zero-based index of the last frame
segment in the range to update

4.3.1.7 Stream Error Code Control

This read-only control indicates the cause of a stream error that may arise during video or still-
image transfer. In such cases, the device will update this control with the appropriate code to
indicate the cause of the error.

The host software should send a GET_CUR request to this control to determine the error when
one of the following events occurs:
- The Error bit in the video or still image payload header is set by the device (see section
2.4.3.2.2, “Sample Isochronous Transfers”).
- The device issues a “Stream Error” interrupt to the host, with the source being the Stream
Error Code Control (see section 2.4.2.2, “Status Interrupt Endpoint”).
- A bulk video endpoint returns a STALL packet to the host in the data or handshake stage
of the transaction.

For scenarios where the host is transmitting video data to the device, the host can not use the
Error bit in the payload header to detect a device error. Therefore, in order to determine when a
streaming error occurs, the host must rely on either a Control Change interrupt from the device
or a bulk endpoint stall.
Table 4-85 Stream Error Code Control
Control Selector VS_STREAM_ERROR_CODE_CONTROL
Mandatory Requests GET_CUR, GET_INFO
wLength 1
Offset Field Size Value Description
0 bStreamErrorCode 1 Number 0: No Error.

1: Protected content – This situation

occurs if the data source device detects
that the video or still-image data is

Revision 1.5 August 9, 2012 154

 USB Device Class Definition for Video Devices

protected and cannot be transmitted. In

this case, empty packets containing only
headers will be sent for the duration of
the protected content.

2: Input buffer underrun – If the data

source device is not able to supply data at
the requested rate, it will transmit empty
packets containing only headers for the
duration of the buffer underrun.

3. Data discontinuity – Indicates a data

discontinuity (arising from bad media,
encoder errors, etc.) preceding the data
payload in the current transfer.

4: Output buffer underrun – The data sink

device is not being supplied with data at a
sufficient rate.

5: Output buffer overrun – The data sink

device is being supplied with data at a
rate exceeding its buffering capabilities.

6: Format change – A dynamic format

change event occurred. See section
2.4.3.6, "Device Initiated Dynamic
Format Change Support".

7: Still image capture error - An error

occurred during still-image capture.

Revision 1.5 August 9, 2012 155

 USB Device Class Definition for Video Devices

Appendix A. Video Device Class Codes

A.1. Video Interface Class Code

Table A- 1 Video Interface Class Code
Video Interface Class Code Value
CC_VIDEO 0x0E

A.2. Video Interface Subclass Codes

Table A- 2 Video Interface Subclass Codes
Video Subclass Code Value
SC_UNDEFINED 0x00
SC_VIDEOCONTROL 0x01
SC_VIDEOSTREAMING 0x02
SC_VIDEO_INTERFACE_COLLECTION 0x03

A.3. Video Interface Protocol Codes

Table A- 3 Video Interface Protocol Codes
Video Protocol Code Value
PC_PROTOCOL_UNDEFINED 0x00
PC_PROTOCOL_15 0x01

A.4. Video Class-Specific Descriptor Types

Table A- 4 Video Class-Specific Descriptor Types
Descriptor Type Value
CS_UNDEFINED 0x20
CS_DEVICE 0x21
CS_CONFIGURATION 0x22
CS_STRING 0x23
CS_INTERFACE 0x24
CS_ENDPOINT 0x25

Revision 1.5 August 9, 2012 156

 USB Device Class Definition for Video Devices

A.5. Video Class-Specific VC Interface Descriptor Subtypes

Table A- 5 Video Class-Specific VC Interface Descriptor Subtypes
Descriptor Subtype Value
VC_DESCRIPTOR_UNDEFINED 0x00
VC_HEADER 0x01
VC_INPUT_TERMINAL 0x02
VC_OUTPUT_TERMINAL 0x03
VC_SELECTOR_UNIT 0x04
VC_PROCESSING_UNIT 0x05
VC_EXTENSION_UNIT 0x06
VC_ENCODING_UNIT 0x07

A.6. Video Class-Specific VS Interface Descriptor Subtypes

Table A- 6 Video Class-Specific VS Interface Descriptor Subtypes
Descriptor Subtype Value
VS_ UNDEFINED 0x00
VS_INPUT_HEADER 0x01
VS_OUTPUT_HEADER 0x02
VS_STILL_IMAGE_FRAME 0x03
VS_FORMAT_UNCOMPRESSED 0x04
VS_FRAME_UNCOMPRESSED 0x05
VS_FORMAT_MJPEG 0x06
VS_FRAME_MJPEG 0x07
Reserved 0x08
Reserved 0x09
VS_FORMAT_MPEG2TS 0x0A
Reserved 0x0B
VS_FORMAT_DV 0x0C
VS_COLORFORMAT 0x0D
Reserved 0x0E
Reserved 0x0F
VS_FORMAT_FRAME_BASED 0x10
VS_FRAME_FRAME_BASED 0x11
VS_FORMAT_STREAM_BASED 0x12
VS_FORMAT_H264 0x13
VS_FRAME_H264 0x14
VS_FORMAT_H264_SIMULCAST 0x15

Revision 1.5 August 9, 2012 157

 USB Device Class Definition for Video Devices

VS_FORMAT_VP8 0x16
VS_FRAME_VP8 0x17
VS_FORMAT_VP8_SIMULCAST 0x18

A.7. Video Class-Specific Endpoint Descriptor Subtypes

Table A- 7 Video Class-Specific Endpoint Descriptor Subtypes
Descriptor Subtype Value
EP_UNDEFINED 0x00
EP_GENERAL 0x01
EP_ENDPOINT 0x02
EP_INTERRUPT 0x03

A.8. Video Class-Specific Request Codes

Table A- 8 Video Class-Specific Request Codes
Class-Specific Request Code Value
RC_UNDEFINED 0x00
SET_CUR 0x01
SET_CUR_ALL 0x11
GET_CUR 0x81
GET_MIN 0x82
GET_MAX 0x83
GET_RES 0x84
GET_LEN 0x85
GET_INFO 0x86
GET_DEF 0x87
GET_CUR_ALL 0x91
GET_MIN_ALL 0x92
GET_MAX_ALL 0x93
GET_RES_ALL 0x94
GET_DEF_ALL 0x97

A.9. Control Selector Codes

A.9.1. VideoControl Interface Control Selectors

Revision 1.5 August 9, 2012 158

 USB Device Class Definition for Video Devices

Table A- 9 VideoControl Interface Control Selectors

Control Selector Value
VC _CONTROL_ UNDEFINED 0x00
VC_VIDEO_POWER_MODE_CONTROL 0x01
VC_REQUEST_ERROR_CODE_CONTROL 0x02
Reserved 0x03

A.9.2. Terminal Control Selectors

Table A- 10 Terminal Control Selectors
Control Selector Value
TE_CONTROL_UNDEFINED 0x00

A.9.3. Selector Unit Control Selectors

Table A- 11 Selector Unit Control Selectors
Control Selector Value
SU_CONTROL_UNDEFINED 0x00
SU_INPUT_SELECT_CONTROL 0x01

A.9.4. Camera Terminal Control Selectors

Table A- 12 Camera Terminal Control Selectors
Control Selector Value
CT_CONTROL_UNDEFINED 0x00
CT_SCANNING_MODE_CONTROL 0x01
CT_AE_MODE_CONTROL 0x02
CT_AE_PRIORITY_CONTROL 0x03
CT_EXPOSURE_TIME_ABSOLUTE_CONTROL 0x04
CT_EXPOSURE_TIME_RELATIVE_CONTROL 0x05
CT_FOCUS_ABSOLUTE_CONTROL 0x06
CT_FOCUS_RELATIVE_CONTROL 0x07
CT_FOCUS_AUTO_CONTROL 0x08
CT_IRIS_ABSOLUTE_CONTROL 0x09
CT_IRIS_RELATIVE_CONTROL 0x0A
CT_ZOOM_ABSOLUTE_CONTROL 0x0B
CT_ZOOM_RELATIVE_CONTROL 0x0C
CT_PANTILT_ABSOLUTE_CONTROL 0x0D
CT_PANTILT_RELATIVE_CONTROL 0x0E
CT_ROLL_ABSOLUTE_CONTROL 0x0F

Revision 1.5 August 9, 2012 159

 USB Device Class Definition for Video Devices

CT_ROLL_RELATIVE_CONTROL 0x10
CT_PRIVACY_CONTROL 0x11
CT_FOCUS_SIMPLE_CONTROL 0x12
CT_WINDOW_CONTROL 0x13
CT_REGION_OF_INTEREST_CONTROL 0x14

A.9.5. Processing Unit Control Selectors

Table A- 13 Processing Unit Control Selectors
Control Selector Value
PU_CONTROL_UNDEFINED 0x00
PU_BACKLIGHT_COMPENSATION_CONTROL 0x01
PU_BRIGHTNESS_CONTROL 0x02
PU_CONTRAST_CONTROL 0x03
PU_GAIN_CONTROL 0x04
PU_POWER_LINE_FREQUENCY_CONTROL 0x05
PU_HUE_CONTROL 0x06
PU_SATURATION_CONTROL 0x07
PU_SHARPNESS_CONTROL 0x08
PU_GAMMA_CONTROL 0x09
PU_WHITE_BALANCE_TEMPERATURE_CONTROL 0x0A
PU_WHITE_BALANCE_TEMPERATURE_AUTO_CONTROL 0x0B
PU_WHITE_BALANCE_COMPONENT_CONTROL 0x0C
PU_WHITE_BALANCE_COMPONENT_AUTO_CONTROL 0x0D
PU_DIGITAL_MULTIPLIER_CONTROL 0x0E
PU_DIGITAL_MULTIPLIER_LIMIT_CONTROL 0x0F
PU_HUE_AUTO_CONTROL 0x10
PU_ANALOG_VIDEO_STANDARD_CONTROL 0x11
PU_ANALOG_LOCK_STATUS_CONTROL 0x12
PU_CONTRAST_AUTO_CONTROL 0x13

A.9.6. Encoding Unit Control Selectors

Table A- 14 Encoding Unit Control Selectors
Control Selector VALUE
EU_CONTROL_UNDEFINED 0x00
EU_SELECT_LAYER_CONTROL 0x01
EU_PROFILE_TOOLSET_CONTROL 0x02
EU_VIDEO_RESOLUTION_CONTROL 0x03
EU_ MIN_FRAME_INTERVAL_CONTROL 0x04
EU_ SLICE_MODE_CONTROL 0x05
EU_RATE_CONTROL_MODE_CONTROL 0x06

Revision 1.5 August 9, 2012 160

 USB Device Class Definition for Video Devices

EU_AVERAGE_BITRATE_CONTROL 0x07
EU_CPB_SIZE_CONTROL 0x08
EU_PEAK_BIT_RATE_CONTROL 0x09
EU_QUANTIZATION_PARAMS_CONTROL 0x0A
EU_SYNC_REF_FRAME_CONTROL 0x0B
EU_LTR_BUFFER_ CONTROL 0x0C
EU_LTR_PICTURE_CONTROL 0x0D
EU_LTR_VALIDATION_CONTROL 0x0E
EU_LEVEL_IDC_LIMIT_CONTROL 0x0F
EU_SEI_PAYLOADTYPE_CONTROL 0x10
EU_QP_RANGE_CONTROL 0x11
EU_PRIORITY_CONTROL 0x12
EU_START_OR_STOP_LAYER_CONTROL 0x13
EU_ERROR_RESILIENCY_CONTROL 0x14

A.9.7. Extension Unit Control Selectors

Table A- 15 Extension Unit Control Selectors
Control Selector Value
XU_CONTROL_UNDEFINED 0x00

A.9.8. VideoStreaming Interface Control Selectors

Table A- 16 VideoStreaming Interface Control Selectors
Control Selector Value
VS_CONTROL_UNDEFINED 0x00
VS_PROBE_CONTROL 0x01
VS_COMMIT_CONTROL 0x02
VS_STILL_PROBE_CONTROL 0x03
VS_STILL_COMMIT_CONTROL 0x04
VS_STILL_IMAGE_TRIGGER_CONTROL 0x05
VS_STREAM_ERROR_CODE_CONTROL 0x06
VS_GENERATE_KEY_FRAME_CONTROL 0x07
VS_UPDATE_FRAME_SEGMENT_CONTROL 0x08
VS_SYNCH_DELAY_CONTROL 0x09

Revision 1.5 August 9, 2012 161

 USB Device Class Definition for Video Devices

Appendix B. Terminal Types

The following is a list of possible Terminal types. This list is non-exhaustive and could be
expanded in the future.

B.1. USB Terminal Types

These Terminal types describe Terminals that handle signals carried over the USB, through
isochronous or bulk pipes. These Terminal types are valid for both Input and Output Terminals.
Table B- 1 USB Terminal Types
Terminal Type Code I/O Description
TT_VENDOR_SPECIFIC 0x0100 I/O A Terminal dealing with a
signal carried over a
vendor-specific interface.
The vendor-specific
interface descriptor must
contain a field that
references the Terminal.
TT_STREAMING 0x0101 I/O A Terminal dealing with a
signal carried over an
endpoint in a
VideoStreaming interface.
The VideoStreaming
interface descriptor points to
the associated Terminal
through the bTerminalLink
field.

B.2. Input Terminal Types

These Terminal Types describe Terminals that are designed to capture video. They either are
physically part of the video function or can be assumed to be connected to it in normal operation.
These Terminal Types are valid only for Input Terminals.
Table B- 2 Input Terminal Types
Terminal Type Code I/O Description
ITT_ VENDOR_SPECIFIC 0x0200 I Vendor-Specific Input
Terminal.
ITT_CAMERA 0x0201 I Camera sensor. To be used
only in Camera Terminal
descriptors.
ITT_MEDIA_TRANSPORT_INPUT 0x0202 I Sequential media. To be

Revision 1.5 August 9, 2012 162

 USB Device Class Definition for Video Devices

used only in Media

Transport Terminal
Descriptors.

B.3. Output Terminal Types

These Terminal types describe Terminals that are designed to render video. They are either
physically part of the video function or can be assumed to be connected to it in normal operation.
These Terminal types are only valid for Output Terminals.
Table B- 3 Output Terminal Types
Terminal Type Code I/O Description
OTT_ VENDOR_SPECIFIC 0x0300 O Vendor-Specific Output
Terminal.
OTT_DISPLAY 0x0301 O Generic display (LCD,
CRT, etc.).
OTT_MEDIA_TRANSPORT_OUTPUT 0x0302 O Sequential media . To be
used only in Media
Transport Terminal
Descriptors.

B.4. External Terminal Types

These Terminal types describe external resources and connections that do not fit under the
categories of Input or Output Terminals because they do not necessarily translate video signals to
or from the user of the computer. Most of them may be either Input or Output Terminals.
Table B- 4 External Terminal Types
Terminal type Code I/O Description
EXTERNAL_ VENDOR_SPECIFIC 0x0400 I/O Vendor-Specific External
Terminal.
COMPOSITE_CONNECTOR 0x0401 I/O Composite video connector.
SVIDEO_CONNECTOR 0x0402 I/O S-video connector.
COMPONENT_CONNECTOR 0x0403 I/O Component video connector.

Revision 1.5 August 9, 2012 163

 USB Device Class Definition for Video Devices

Appendix C. Video and Still Image Formats

C.1. Supported video and still image formats

This specification is designed to be format-agnostic, and will support any present or future video
or still image format. The video and still image formats supported by the device are reported to
the host software via Format descriptors (see section 3.9.2.3, "Payload Format Descriptors").

C.2. Proprietary video formats

New or proprietary video and still-image formats must be defined outside of this specification
via Payload Format Specifications. . The host software will require a matching video encoder or
decoder module.

Revision 1.5 August 9, 2012 164

 USB Device Class Definition for Video Devices

Appendix D. Optical and Digital Zoom

Optical and digital zoom are functionally independent, so each will be discussed separately in the
following sections. Although functionally independent, users will expect a single zoom control
that integrates both.

D.1. Optical Zoom

Although lens groups can be quite sophisticated, this specification describes a simple two-lens
system, which is sufficient to model optical zoom. Given objective and ocular lens focal lengths
(Lobjective and Locular), magnification (M) can be calculated as follows:
Lobjective
M =
Locular

The objective lens is the one nearest the subject, while the ocular lens is the one nearest the
viewer, or in our case, the camera sensor. A zoom lens varies the objective focal length.

Since magnification is a ratio of the objective and ocular focal lengths, the Units used to specify
these focal lengths can be of any resolution supported by the device. In other words, these Units
do not need to be specified in real physical units (millimeters or fractions of inches). The only
requirement is that the two focal lengths are specified in the same units.

Note that when Lobjective < Locular, the lenses are at a wide-angle setting. The subject will appear
smaller than life, and the field of view will be wider.

Locular will be a device-specific constant value for each camera implementation, so it will be
specified within the static Camera Terminal descriptor. If a camera implements an optical zoom
function, Lobjective can vary within a specified range. In order to properly represent the range of
magnification, Lobjective will be specified as a range Lmin to Lmax, which will also be specified
within the static Camera Descriptor.

Finally, the variable position within the range of possible Lobjective values will be specified via a
dynamic Camera Zoom Control, as integral values Zmin, Zmax, Zstep, and Zcur. See sections
4.2.2.1.12, "Zoom (Absolute) Control" and 4.2.2.1.13, "Zoom (Relative) Control". This allows
the Units of the objective lens focal length to be de-coupled from the Units used to control zoom.
For simplicity, Zstep will be constrained to equal one (1). Values of Lmin and Lmax are constrained
to be non-zero integral numbers; however, for the purpose of the following calculations, Lcur will
be a real number.

Note: A typical choice for Locular would be half the length of a diagonal line of the imager
(CCD, etc.), however there is no requirement for this value to be a direct physical measurement.

Given a known Zcur, the current objective focal length (Lcur) can be calculated as follows:

Revision 1.5 August 9, 2012 165

 USB Device Class Definition for Video Devices

(Zcur - Zmin) * (Lmax - Lmin) + Lmin

Lcur =
(Zmax - Zmin)

From this, the relative magnification can be calculated as follows:

Lcur
M =
Locular

Working from the opposite direction, given a known magnification (M), Lcur can be calculated as
follows:

Lcur = M * Locular

From this, the current Zoom control value (Zcur) can be calculated as follows:

(Lcur - Lmin) * ( Zmax - Zmin) + Zmin

Zcur =
(Lmax - Lmin)

To further simplify the calculations, Zmin can be constrained to be zero (0). The camera designer
will choose the values and ranges of the remaining variables according to the capabilities of the
device.

As an example, substituting some plausible values for each of these variables:

Lmin = 800
Lmax = 10000
Zmin = 0
Zmax = 255

The current Objective focal length (Lcur) can be calculated as follows:

Zcur 9200
* + 800
Lcur =
255

The current Zoom control value (Zcur) can be calculated as follows:

Revision 1.5 August 9, 2012 166

 USB Device Class Definition for Video Devices

(Lcur – 800) * 255

Zcur =
9200

When choosing a camera sensor to match a lens system, the camera designer may need to
consider a multiplier effect caused by a sensor that is smaller than the exit pupil of the ocular
lens. This multiplier will not be represented explicitly in the USB Video Class specification,
since its effect can be represented via adjustments to the Lobjective values.

Note The Zcur value can be mapped to the physical lens position sensor control/status register.

D.2. Digital Zoom

Digital zoom is applied after the image has been captured from the sensor. Thus, digital zoom is
independent of optical zoom, and is a function of either the Processing Unit or host post-
processing. Although digital zoom is independent of optical zoom, users have come to expect
that camera implementations will not apply digital zoom until full optical zoom has been
realized. This will be enforced by the host software. There is no requirement for the device to
enforce this, but it is recommended.

Digital zoom is represented as a multiplier of the current optical magnification of the captured
image. In order to change the amount of digital zoom, the multiplier is changed through a range
from 1 to some maximum value mmax, and mmax will be specified in the Processing Unit
Descriptor. The position within the range of possible values of multiplier m will be expressed via
a Processing Unit Digital Multiplier Control, as Zmin, Zmax, Zstep, and Zcur. See section
4.2.2.3.16, "Digital Multiplier Control". This allows the multiplier resolution to be described by
the device implementation. Zstep will be constrained to equal one (1).

Given a known Zcur, the current multiplier mcur can be calculated as follows:

(Zcur - Zmin) * (mmax - 1) + 1

mcur =
(Zmax - Zmin)

From this, and referring to the optical zoom values of Lmax and Locular described in the previous
section, the total magnification M can be calculated as follows:

Revision 1.5 August 9, 2012 167

 USB Device Class Definition for Video Devices

Lmax
mcur
M = *
Locular

Working from the opposite direction, given a known magnification M, the multiplier mcur can be
calculated as follows:
Locular
mcur = M *
Lmax

From this, the current Digital Multiplier Control value (Zcur) can be calculated as follows:

(mcur - 1) * ( Zmax - Zmin) + Zmin

Zcur =
(mmax - 1)

For simplicity, Zmin can be constrained to be zero (0). The camera designer will choose the
values and ranges of the remaining variables according to the capabilities of the device.

As an example, substituting some plausible values for each of these variables:

mmax = 40
Zmin = 0
Zmax = 255

The current multiplier (mcur) can be calculated as follows:

Zcur 39
* + 1
mcur =
255

The current Digital Zoom control value (Zcur) can be calculated as follows:

(mcur – 1) * 255
Zcur =
39

In addition to the Digital Multiplier Control, devices may optionally support a Digital Multiplier
Limit control, allowing either the camera or the host to establish a temporary upper limit for the
Zcur value. This control may be read-only if the limit can only be established via physical camera

Revision 1.5 August 9, 2012 168

 USB Device Class Definition for Video Devices

configuration. If this control is used to decrease the limit below the current Zcur value, the Zcur
value will be adjusted to match the new limit.

D.3. Relationship between Optical and Digital Zoom

As mentioned in the preceding sections, users expect to use a single control on the device (or
from within an application on the host) to traverse the entire range of optical and digital zoom.
Further, users expect that digital zoom will not be active except at full optical zoom.

The following diagram illustrates the relationship between optical and digital zoom, and the
constraints on the zoom control variables:

Zmin Zmax
Zcur == Zmax
Optical Range Zmin <= Zcur <= Zmax

Zcur == Zmin Digital Range

Zmin <= Zcur <= Zmax
Zmin Zmax
Increasing Magnification

Figure 4-8 Relationship between Optical and Digital Zoom

D.4. Absolute vs. Relative Zoom

The equations and examples given in the previous sections describe independent, absolute
optical and digital zoom controls. However, based on users’ expectations that devices provide a
single relative zoom control allowing them to move across the entire zoom range (from wide to
telephoto and back again), many cameras will implement a relative zoom control that supports
increasing and decreasing the zoom parameters without actually specifying the parameter values.
Devices that allow only relative zoom control should still report the optical focal lengths and
maximum digital multiplier in their respective descriptors, as well as maintain read-only absolute
optical and digital zoom controls. This way, the host software will always be able to determine
the current state of the zoom values.

Revision 1.5 August 9, 2012 169