Multichannel Setup V5 (1) .1
Multichannel Setup V5 (1) .1
Version 5.1
© 2009
Table of Contents
INTRODUCTION 1
B MULTICHANNEL FORMATS 11
1 Surround Sound Formats 11
1.1 Non-Discrete Formats 11
1.2 Discrete 5.1 Formats 12
1.3 The 6.1 and 7.1 Formats 13
1.4 More Channels for Surround Sound 15
2 New Storage Media 16
2.1 HD-DVD 16
2.2 Blu-ray Disc 16
2.3 Holographic Versatile Disc 17
3 Multichannel Audio Applications 17
3.1 Broadcasting Applications 17
3.2 Other Applications 18
F REFERENCES 51
1 Technical Papers 51
2 Website Sources 51
INTRODUCTION
A modern audio production facility must be able to serve productions in a large number of different formats.
The change from mono and stereo to multichannel reproduction has created many challenges, both in
converting existing production facilities to multichannel format and in new installations.
The multitude of audio formats that must be handled by a modern production facility include currently:
• Mono, stereo
• Matrix four channels
• Five channels (5.0 systems)
• Five channels with an LFE channel (5.1 systems)
• Six channels (with either discrete or matrix Rear Centre channel) with an LFE channel (6.1 systems)
• Eight channels including either a full range or a band-limited LFE channel (7.1 systems)
• Other multichannel formats and more channels: from 13.1 to 22.2 and more
This paper discusses multiple practical questions about monitoring loudspeakers, their set-up and possible
sources of problems and suggestions on how to avoid them. A brief overview of the current multichannel
formats and dedicated sections on low frequency effect (LFE) channel and Bass Management Systems are
also included.
Because the space has a strong effect on the radiation, it is important to understand the conditions.
The radiation space is typically characterised by a rough estimate of the solid angle (part of a sphere)
into which the loudspeaker is radiating (see Figure 1). As the loudspeaker/subwoofer is driven, it
creates a certain volume flow, which naturally spreads into all directions. As we limit the space seen
by the loudspeaker and at the same time we keep the total power identical the energy density
(intensity) in the limited radiation space increases. Hence, reducing the radiation space increases
Sound Pressure Level (SPL). Every halving of the radiation space doubles the SPL.
1/8th space
Half space Quarter space
Free space In an apex
Flush mounted In a corner
Free standing (corner limited by
Or next to a wall (limited by two walls)
three walls)
0 dB +6 dB +12 dB +18 dB
Figure 1 - Solid angle values, respective amplitude gain and loudspeaker alignments.
• Against one solid boundary, which is large compared to the wavelength, the radiation space
is 2π, and the theoretical amplitude gain is +6 dB at low frequencies. This applies typically
to flush mounting.
• A subwoofer is typically placed on the floor and against a wall, hence we get two
boundaries, and the radiation is now into solid angle π, and the amplitude gain is +12 dB.
It is essential to note that in this context the words like ”large”, ”close”, etc. refer always to the
wavelength. ”Large” compared for example to 10 m (34 Hz) is different from ”large” compared to 3.4
m (100 Hz) or 3.4 cm (10 kHz). This means that when a loudspeaker is placed ”far” from boundaries
at, say, 150 Hz, it probably is very ”close” at 30 Hz. The radiation space depends on frequency and
therefore it is important to be able to correct the response of the loudspeaker/subwoofer so that the
final acoustic performance stays as flat as possible.
In real life the depth and width of the cancellation dip varies depending on the level of the reflection,
but in most cases it is well audible. No equalisation will cure this situation because it originates from
interference; adding amplitude at the dip frequency will also boost the reflection and thus their sum
remains the same. This simple case deals with one mode only: reflection from one wall behind the
loudspeaker. Table 1 shows various distances between a wall and the front baffle of a loudspeaker
with their corresponding dip frequencies.
Distance from the wall (meters) First cancellation frequency (Hz) D = λ/4
0.10 858
0.20 429
0.40 214
0.60 143
0.80 107
1.00 86
1.20 71
1.40 61
1.60 54
1.80 48
2.00 43
2.20 39
2.40 36
2.60 33
2.80 31
3.00 29
Table 1 - First cancellation frequency at different distances from the wall
The first and best cure for the ‘wall behind the loudspeaker’ cancellation dips is to flush mount the
loudspeakers in a hard wall – also called ‘infinite baffle’ or ‘flush-mounting’ (which totally eliminates
this wall reflection and cancellation) – see Chapter 6.
Second best is placing the loudspeaker very close to the wall, which raises the cancellation frequency
higher. This works well when the loudspeaker is not too small. The risk is, with small loudspeakers
which inherently are less directional in mid frequencies that the dip just moves to the low mid-band
and causes even worse coloration. As seen above, the distances between 0 and 20 cm from the wall
let the loudspeaker response to be, in most cases, unaltered; i.e. the directivity of the loudspeaker is
high enough so that the rear radiation cannot cause a severe cancellation. Additionally, the low
frequency boost should be compensated for when the loudspeaker is mounted close to the wall (+6
dB).
Alternatively, the third cure is to move the loudspeaker considerably away from the wall: the
cancellation frequency goes down so far that it is below the low frequency cut-off of the loudspeaker.
Thus, the minimum distance ‘loudspeaker/wall behind’ depends on the loudspeaker low frequency
performance. However, at low frequencies and for large loudspeakers, the minimum distance
becomes very long and impractical. At the same time, the distances to other boundaries in the room
become similar to the desired distance to the wall behind the loudspeaker, and the reflections from
these other surfaces start to dominate the response.
The fourth cure is to make the wall so absorptive that the reflected amplitude is so small that it will not
cancel the direct sound. The thickness of a porous absorber has to be one quarter of the wavelength
of the frequency to be absorbed to become effective. This is the same distance that determines the
frequency of the cancellation dip, therefore, the absorber has to be very thick and, in most cases, this
is not done.
Although flush mounting loudspeakers (see Chapter 6) offers many benefits it is expensive and in
most cases, especially in smaller installations, the loudspeakers are left free standing. As stated
above reflections from various boundaries will characterise the performance of a free standing
loudspeaker.
The Genelec 8000 / 8200 Series bi-amplified loudspeakers should be placed so that a minimum gap
of 5 cm (2”) is left behind the loudspeaker for amplifier cooling and rear opening reflex port sound
radiation. In general, when positioning the loudspeaker’s front baffle further than 30 cm from the wall,
a reflection can cause a cancellation in the low frequency response and hence a loss in the bass
reproduction quality. For two-way loudspeakers, low frequency cancellations in the range 40…80 Hz
should definitely be avoided (see Figure 4). Low frequency cancellations in the 80…200 Hz range
should also be avoided where possible, but if this is not technically possible the overall sound quality
will still be perceived as good. Translating these frequency ranges into distance recommendations
shows that an acceptable response can be achieved at distances from the wall up to 1 m (see Figure
3). Beyond that, the 1 m…2.2 m range should definitely be avoided. Large loudspeakers placed at a
distance greater than 2.2 m may suffer from a cancellation in the very low frequency region around
their low frequency cut-off, thereby compromising the loudspeaker’s low frequency extension. So the
lower the low frequency cut-off, the further away the loudspeaker must be placed from this wall.
Figure 3 - Distances from a single wall to the front baffle of free standing loudspeakers.
Correct (green), acceptable (orange) and avoid (red).
AVOID ! Acceptable
2.2 m - 1 m up to 1 m
Figure 4 - Frequency domain notches and distances from the single wall
behind a free-standing loudspeaker and its front baffle.
• For larger loudspeakers placed away from the wall the necessary distance is far too long for
any practical rooms.
• In such cases, the distances to ceiling and walls are already smaller than the distance to the
wall behind the loudspeaker. Reflections from these surfaces are important and might
become more significant.
In practice free standing loudspeakers always suffer from many cancellations induced irregularities of
their frequency response.
When using subwoofers the additional crossover (at 85 Hz) between the two loudspeaker systems
changes the whole monitoring configuration. The subwoofer itself should be placed close to the
wall(s) in order to maximise its efficiency – maximum distance of 60 cm (see Figure 5). This also
eliminates most of the possible sources of cancellation dips in the subwoofer response, because the
subwoofer is acoustically close to the boundaries.
‘Satellite’ loudspeakers high-passed at 85 Hz do not have to reproduce very low frequencies so they
may be placed at a distance where low frequency notches do not occur in their pass-band. The
guidelines for placing ‘satellite’ loudspeakers are similar to the ones for free-standing loudspeakers
(see Figure 5).
The ‘acceptable’ distance extends out to 1.1 m due to the fixed low frequency cut-off of the ‘satellite’
loudspeakers. From 1.1…2 m, the loudspeakers may be placed without serious compromises from
the wall behind the loudspeaker’s reflection and corresponding cancellation effects. Although Genelec
subwoofers provide accurate phase control facilities at the crossover point, the ‘satellite’ loudspeakers
should not be placed too far (max. 2 m) from the subwoofer. If this is the case, the tonal balance
between the satellite loudspeakers and the subwoofer may differ considerably due to excitation of
different room modes by the sources.
GENELEC
Figure 5 - Distances from a single wall to the front baffle of the loudspeakers combined with subwoofer(s). Correct (green),
acceptable (orange) and avoid (red).
3 Other Reflections
The discussion above concentrated on one reflection only, i.e. the reflection from a single wall behind
the loudspeaker. Similar reflections also occur in other directions (side walls, floor or ceiling, console
top, equipment racks etc.) and their frequencies again depend on the distances to the relevant
boundaries. In general, loudspeakers should be placed as far as possible from reflecting surfaces.
This will place the reflection-induced irregularities to as low a frequency as possible, which is also
beneficial to the imaging. Although not always practical, one suggestion is that if there are many
reflecting surfaces (tables, computer screens, etc…), loudspeakers could be placed slightly higher
than the listening level, and then tilted down towards this listening position.
The critical midrange response of the centre channel is often seriously compromised by strong
reflections. Placing this centre channel loudspeaker below the video screen means the loudspeaker
cabinet is closer to the floor and the floor reflections will color strongly the midrange response. So, the
centre channel loudspeaker should be placed above the video screen or TV, ensuring that it does not
suffer from first order ceiling reflections. Where the ceiling is low and hence might generate strong
reflections, some damping material can be applied to absorb these first reflections
It is well known, for example, that the first reflection created by the surface of a large mixing desk
induces 3…6 dB ripples starting at about 1 kHz – 2 kHz and extending up to higher frequencies. The
subjective effect of this comb filtering is that the loudspeaker does not sound so ‘open’. Also, the
vibration coupling between loudspeaker and console will especially affect small near-field
loudspeakers placed on the meter bridge of a console. This could lead to a compromised frequency
response particularly in the midrange area. To solve this common problem, Genelec developed a
special support called Iso-PodTM (Isolation Positioner/DecouplerTM) for its 8000 / 8200 series monitors.
The rubber based Iso-PodTM stand provides vibration decoupling between loudspeaker and console
meter-bridge. It is also designed so that the loudspeaker can be tilted (+/-7 degrees) enabling the
acoustical axis to be pointed precisely towards the listening position.
A loading-induced compromise present in the frequency response is a ‘bump’ in the range 100…250
Hz when a near-field monitor is placed close to a large surface or typically a mixing console (1...3 m2 ).
To compensate for this ‘bump’ the larger 8000 / 8200 Series monitors offer a ‘Desktop Low Frequency’
control that provides a reduction of 4 dB @ 160 Hz, so that a flat response may be retrieved in that
region as well. A statistic study1 has shown that this control is required in about 60% of cases where a
loudspeaker is mounted in such a way.
Early reflections, with high amplitude in relation to the direct sound, can smear the coherence of the
sound image and compromise the localisation of the sources in space. To avoid this, all reflecting
surfaces between the loudspeakers and the listening position should be minimised. Of course, some
equipment are inevitable, however symmetrical positioning of the equipment is essential. Even with
symmetry, reflections will remain and everything possible should be done to remove the reflecting
surfaces from the vicinity of the acoustic path – see Figure 6. It should also be remembered that the
smaller physically the loudspeaker, the less directional it is, and the more influenced it will be by its
surroundings.
Figure 6
(6a.) Example of a non-symmetrical layout (6b.) Example of a symmetrical layout
Reflections from computer screens and table surface are Reflection surfaces are minimized, away from the listening
totally different. This will create a front-back and left-right position and clearing the acoustical path where possible.
localisation smearing.
1
“Compensating the Acoustical Loading of Small Loudspeakers mounted near Desktops”, A. Goldberg, A. Mäkivirta, A.
th
Varla, Genelec Oy, Iisalmi, Finland, presented at the 117 AES Convention, San Francisco, U.S.A., October 2004.
As we know from current practice, the typical multichannel room is often not originally designed as a
multichannel room. In addition, as standard practices are still not well defined, the installations vary
from one room to the next. Clearly, loudspeakers used for multichannel work should have a well
controlled directivity, which leads to a high direct-to-reflected sound level ratio and a reduced
interaction with the nearby boundaries. The listeners will be located in the more direct field and will
therefore be able to listen to more of the program material and less to the room interaction. The
purpose of the Genelec Directivity Control Waveguide (DCWTM) is to control the radiation angle of the
tweeter and midrange elements such that the detrimental effect of diffractions from the loudspeaker
cabinet and the nearby walls and structures is minimized. By increasing the direct-to-reverberant
sound ratio, the localization and the uniform quality of the sound is improved, irrespective from the
loudspeaker location.
It should be remembered that the radiation load changes once the loudspeakers are flush mounted
and that adjustment of the loudspeaker’s frequency response is required to keep a flat response below
200 Hz.
When flush mounting small active loudspeakers, the amplifier ventilation has to be arranged from the
rear. This is best done, either totally from the rear of the wall where the loudspeakers are mounted, or
by leaving suitable ducts to the listening space. However, the space between the flushed loudspeaker
cabinet edge and the surrounding wall should never be left open without a facing panel. If there is no
facing panel, the cavity might form a resonator, which might have adverse effects on the frequency
response and perceived sound quality – see Figure 7 for construction detail examples.
In order to enhance the low frequency reproduction the wall in which the loudspeakers will be mounted
is made hard and dense. For two-channel monitoring conditions, this is ideal but for multichannel
monitoring, the front wall might form a large reflective surface off which the rear loudspeakers direct
sound might bounce. In some cases, the large loudspeakers’ front baffle themselves offer a reflective
surface for midrange and high frequencies. Depending on the size of the room these reflections may
cause image instability. It is therefore recommended that these large hard wall surfaces should be
covered with an absorbent, which is efficient enough at midrange and high frequencies but does not
absorb low frequencies. A 50 mm sheet of rock wool or similar is suitable for the job - see Figure 7
lower right drawing.
In an ideal case, all five loudspeakers should be flush mounted in a similar way to achieve the best
possible results. With all loudspeakers on the same radius, this will lead to new room shapes.
Various other issues in room design such as geometry of the control room, ceiling design, sound
diffusion, positioning and space for low frequency absorbers, midrange definition, clarity, etc… have to
be addressed separately by studio designers.
The very low frequency damping is an issue of its own and does not really vary from standard stereo
room design, except that there should be sufficient absorption areas on the sidewalls as well. The
typical rear wall absorbers are not sufficient to damp modes between sidewalls, which are excited by
the rear loudspeakers.
Multi-layered wall
(wood, concrete, bricks)
Concrete Wall
Sand Amplifier
bags removed from
enclosure
Concrete
structure
Amplifier
removed Rubber gasket
from enclosure (5-10mm / 1/4-1/2")
Vibration
Vibration Facing panel Isolators
Isolators fixed to the wall f ~ 2-8 Hz
(50-100mm / 2-4")
f ~ 2-8 Hz
Absorbing
material
Speaker
Rubber gasket
(5-10mm / ¼-½")
B MULTICHANNEL FORMATS
As the number of audio channel grows in the consumer and professional market, new multichannel
formats regularly appear. However, and for obvious reasons, most of the previously accepted
multichannel formats continue to be used. This chapter provides an overview of the encoding and
decoding technologies available today and introduces new developments and ideas.
The non-discrete systems are often called “Matrix 4:2:4” formats since the encoder receives two input
channels and delivers 3, 4 or more output channels after processing. During this process, it is
essential to use the proprietary encoder/decoder to monitor exactly what will be heard by the end
listener in each of the extracted channels.
• Dolby®* Surround
Dolby Surround is a matrix encoding process that in essence “folds” Left, Centre, Right, and
Surround channels onto a stereo soundtrack - Dolby Stereo (Lt/Rt). The passive Dolby
Surround decoder extracts three channels (Left, Right & Mono Surround) from two channels
Dolby Stereo (Lt/Rt). The left and right channels are full range and the mono surround
channel is band limited between 100 Hz and 7 kHz. In practice the surround channel is only
used for sound effects. Dolby Stereo encoded soundtracks can be played back in either,
mono, stereo or surround format.
works with any stereo, Dolby Surround, Dolby Pro Logic II, or Dolby Digital encoded
material.
• DTS® NEO:6TM*
DTS NEO:6 is an active matrix system that delivers up to six channels (L, C, R, Ls, Cs, Rs)
of audio from any two-channel source, analogue or digital. This format allows anyone with a
DTS NEO:6 decoder to simulate a surround sound experience, off of any software material.
• Logic 7 Matrix
Dr David Griesinger and Lexicon®* have developed this decoding system. It can be applied
to car audio, consumer decoder, etc. One feature of this matrix decoder is that it can extract
7.1 channels from conventional stereo program material. Two additional side loudspeakers
are required to replay this matrix format.
• Meridian©* TriField
Meridian TriField surround processing accepts any conventional two-channel sources, and
separate them into multiple channels. The extracted surround components of the original
recording are delivered to the rear surround speakers, and side surround speakers, if
present. The TriField processing technique allows the listener to adjust the width of the
image. Meridian recommends the use of this processing option for solo, chamber, or vocal
music.
In general, various audio quality problems might exist with matrix encoded surround systems. They
might suffer from a lack of channel separation and in some formats certain channels have a limited
frequency response. Some formats are also prone to centre channel collapse. In order to improve the
surround sound experience another group of surround systems were developed: the discrete formats.
These formats are commonly called discrete systems as all the main channels (Left, Centre, Right,
Surround Left & Surround Right) are full bandwidth (20 Hz – 20 kHz) but there is also an LFE channel
that is band limited. Note that the rear surround information is now in ‘stereo’ as there are two
different channels. Here are some of the current 5.1 formats:
• Dolby Digital
The Dolby Digital format is a discrete system that keeps the multiple channels fully
separated throughout the encoding and decoding processes. The six original discrete
channels are digitally encoded into a single bit-stream using the Dolby AC-3 (Audio Coding
3) encoding system. The codec can encode mono, stereo and 5.1 signals all in 20-bits/48
kHz quality.
PCM data stream, ’packs’ it (PPCM – Packed Pulse Code Modulation) at one end of the
chain and ’unpacks’ it at the other to provide a completely accurate replica of the original.
The difference between Dolby Digital, DTS and MLP is that MLP is loss-less, i.e. the original
signals are encoded without compression losses. MLP is licensed by Dolby Laboratories.
Motivated by the movie industry as well as the automotive, the video and computer gaming industries,
the companies developing multichannel formats have already gone beyond six channels of encoding
and replay. Most consumers are already familiar with 6.1 systems but high definition 7.1 formats are
now also available.
• Dolby E
This Dolby format encodes eight digital full bandwidth channels into two channels, with a
digital audio frame identical to the digital video frame for ease of editing. A 5.1 and stereo
mixes can be sent down a single AES-EBU cable with up to ten encode-decode cycles
possible before audible degradation. This professional format is not intended for consumer
use and will not reach viewers at home. Dolby E is optimized for the distribution of surround
and multichannel audio through two-channel to post-production and broadcasting
infrastructures, or for recording surround audio on two audio tracks of conventional digital
video tapes, video servers, communication links, switchers, and routers.
• Dolby TrueHD™
Dolby TrueHD is a lossless encoding technology developed for high-definition disc-based
media. It provides up to 18 Mbps data rate, supports up to eight full-range channels of 24-
bit/96 kHz audio and extensive metadata including dialogue normalization and dynamic
range control. Dolby TrueHD was selected as the mandatory format for HD-DVD and as an
optional format for Blu-ray Disc. Note that HD-DVD and Blu-ray disc standards currently limit
their maximum number of audio channels to eight. This new format is also supported by
High-Definition Media Interface (HDMI™), the new single-cable digital connection for audio
and video signals.
• DTS-ESTM
First, the DTS-ESTM Discrete format is a 6.1 format that uses a 5.1 DTS data stream with
additional extension bits that allows the decoder to retrieve an additional discrete Rear
Centre channel. The DTS-ESTM Matrix version ‘folds’ and ‘unfolds’ the rear centre channel
from the matrix encoded surround left and right channels.
• DTS-HDTM
DTS-HD Master Audio™ format feature variable bit rate (up to 24 Mbps), bit-for-bit (lossless)
encoded streams with up to 7.1 channels in 24-bits/96 kHz. DTS-HD High Resolution
Audio™ format feature constant bit rate (lossy) encoded streams up to 6 Mbps for Blu-ray
Disc and up to 3 Mbps for HD-DVD. DTS-HD was selected as optional audio format for both
Blu-ray and HD-DVD discs, but DTS decoders will be built into all HD-DVD or Blu-ray disc
players. DTS-HD Master Audio™ (lossless), DTS-HD High Resolution (lossy), and DTS
Digital Surround (lossy) encoded streams each contain a constant bit rate sub-stream (Core)
that is backward compatible with all DTS Digital Surround decoding technologies in the
market.
• Logic 7
The Logic 7 decoding matrix can have a stereo input signal or a discrete 5.1 input signal.
This matrix will conserve the front L-C-R channels but also extracts four different surround
channels; two side channels and two surround channels. Two LFE channels are also
extracted for ‘better low frequency envelopment’ and replayed over two subwoofers.
Many new multichannel ideas are in development and the range of applications is broadening all the
time. Here are some of the formats using more than eight channels in either encoded or non-encoded
forms:
• Dolby AAC
Dolby® AAC is an enhanced version of the core MPEG-AAC (MPEG-2 and MPEG-4 AAC)
that adds proprietary Dolby intellectual property and supports data rates from 14 kbps
(mono) to 128 kbps (stereo) and up, enabling wider frequency range at lower bit rates.
• Ambisonic System
The Ambisonic surround sound system encodes sound directions and amplitudes and then
reproduces them over practical loudspeaker systems in such a way as to fool the ears of
listeners into thinking that they are hearing the original sounds correctly located. This can
take place over a 360 degree horizontal only soundstage (Pantophonic systems) or over the
full sphere (Periphonic systems).
Systems using the so-called 'B' format signals to carry the recorded information require three
and four channels respectively for full encoding of sounds to the kind of accuracy achievable
with first order microphones (cardioids, figure eight, etc). Practical reproduction requires a
minimum of four (Pantophonic) or eight (Periphonic) loudspeakers. The important thing to
note is that there is no need to consider the actual details of the reproduction system when
doing the original recording or synthesis, since if the ‘B’ format specifications are followed
and suitable loudspeaker/decoder setups are used, the Ambisonic system will work. In all
other respects the two parts of the system, encoding and decoding, are completely separate.
• VBAP
Vector Base Amplitude Panning2 (VBAP) is a method for positioning virtual sources via
multiple loudspeakers. The number of loudspeakers can be varying and they can be placed
in an arbitrary 2-D or 3-D positioning. Since it uses at one time the minimum number of
loudspeakers needed (one, two or three), VBAP produces virtual sources that are as sharp
as is possible with current loudspeaker configuration and amplitude panning methods.
• RF64
The RF64 file format is designed to meet the requirements for multichannel sound in
broadcasting and audio archiving. It is based on the Microsoft RIFF/WAVE format and Wave
Format Extensible for multichannel parameters. Additions are made to the basic
specification to allow more than 4 GB file sizes when needed. A maximum of 18 surround
channels, a stereo down-mix and also bit-stream signals with non-PCM coded data can be
stored in the file.
This list is not exhaustive as there are other multichannel surround formats, like IMAX or OMNIMAX
for example. Many more multichannel formats will be developed in the future since the interest for
multichannel audio and its wide-scale application is growing all the time.
This new ‘High Density Digital Versatile Disc’ can hold high definition video, audio and other data.
HD DVD is similar to the competing Blu-ray Disc, which also uses the same CD-size (120 mm
diameter) optical data storage media and 405 nm wavelength blue laser. HD DVD is promoted by
Toshiba, NEC, Sanyo, Microsoft, and Intel, among others.
HD DVD has a single-layer capacity of 15 GB and a dual-layer capacity of 30 GB. A triple-layer disc
which offers 45GB of storage was announced. HD DVD can offer both the current DVD and HD DVD
formats on one disc, which means that special HD DVD discs will play in any DVD player, old or new.
HD DVD can be mastered with up to 7.1 channel surround sound using uncompressed PCM, Dolby
Digital and DTS formats also used on DVDs. In addition, it also supports Dolby Digital Plus and the
lossless formats Dolby TrueHD and DTS-HD. HD DVD discs support encoding in up to 24-bit/192kHz
for two channels, or up to eight channels at 24-bit/96kHz.
Blu-ray Disc (BD) is also a high-density storage optical disc meant for high-definition video and data.
The Blu-ray standard was jointly developed by a group of consumer electronics and PC companies
called the Blu-ray Disc Association (BDA), spearheaded by Sony. It is currently competing with the HD
DVD format.
A single-layer Blu-ray disc has enough storage capacity (25 GB) for approximately two hours of high-
definition audio/video and a dual-layer (50 GB) disc for about four hours of HD video. TDK recently
announced that they have created a working Blu-ray disc capable of holding 200 GB of data (six 33
GB data layers).
2
The VBAP method has been developed by Dr. Ville Pulkki at Helsinki University of Technology
For audio, BD-ROM supports up to 7.1 channels using uncompressed PCM, Dolby Digital and DTS
formats also used on DVDs. In addition, it also supports Dolby Digital Plus and the lossless formats
Dolby TrueHD and DTS-HD. Blu-ray discs support encoding in up to 24-bit/192kHz for up to six
channels, or up to eight channels at 24-bit/96kHz.
The Blu-ray movie region codes are different from the DVD region codes; A – North America, South
America, East Asia except China. B – Europe and Africa. C – China, Russia and other countries.
While it is not compulsory for manufacturers, the Blu-ray Disc Association recommends that Blu-ray
drives should be capable of reading DVDs for backward compatibility.
Six companies have formed the Holographic Versatile Disc (HVD) Alliance (CMC Magnetics, Fuji
Photo, Nippon Paint, Optware, Pulstec Industrial, Toagosei) which now includes more than 19
companies. The HVD technology will be able to record more than one terabyte (1 TB) of data, 200
times the capacity of a single layer DVD, on a disc the same size as today’s DVD at a rate of over
1Gbps, 40 times the speed of the DVD. A maximum capacity of 3.9 TB on a 12 cm disc was
demonstrated. It employs a technique known as collinear holography, whereby two lasers, one red
and one blue-green, are collimated in a single beam. The blue-green laser reads data encoded as
laser interference fringes from a holographic layer near the top of the disc while the red laser is used
as the reference beam and to read servo information from a regular CD-style aluminum layer near the
bottom.
Optware has announced the release of a 200 GB disc in early June 2006, and Hitachi Maxell in
September 2006 with a capacity of 300 GB. InPhase Technologies has developed a holographic
format they call Tapestry MediaTM*, capable of storing up to 1.6 TB. With such high end storage
capacities, these technologies seem better than either HD DVD or Blu-Ray Disc, however, the reader
currently costs approximately US$ 15,000, and a single disc currently costs approximately US$ 120.
• Digital TV broadcasting in either high or single definition TV format. The DVB transmission
will occur via either terrestrial or satellite network
• TV and radio commercials and jingles
• Internet: radio & TV specific productions for web diffusion (news, sports, interactive
programs)
• Radio: digital audio broadcasting, (DAB) via either terrestrial or satellite network (AOL
incorporated Dolby AAC in its proprietary Ultravox streaming media platform)
• The Digital Radio Mondiale system (the digital replacement for radio broadcasting under 30
MHz) has been developed on the MPEG-4 AAC audio coding technology
Conventional monitoring rooms intended for stereo reproduction are not very suitable for surround
sound use. Likewise, a perfect five channels production is more likely to sound dull and uninteresting
once replayed in stereo only.
• ITU–R BS.775–1
”Multichannel stereophonic sound system with and without accompanying picture” (Geneva,
1992-94)
• ITU–R BS 1116–1
”Methods for the subjective assessment of small impairments in audio systems including
multichannel sound systems” (Geneva, 1994-97)
The ITU is also working together with the IEC, SMPTE and ISO/MPEG organisations. Once again,
more standards and recommendations are issued on a regular basis and this list is not complete. The
following chapters highlight the most important points of these ITU recommendations and discuss
some other essential points of multichannel reproduction.
Much research work was been done over the past 50 years to determine the best angle for two
loudspeakers reproducing stereo material. As we know, 60 degrees has long been considered to give
the optimum stereo width vs. mono phantom image. With multichannel audio, we have a standard
recommendation (ITU-R BS 775-1) for the positioning of the five loudspeakers in space (see Figure 8
below).
It has to be noted that this recommendation is subject to discussion by some producers, engineers,
etc. and that they use alternative angles for positioning the surround channels.
L Sub R
300
1100
LS RS
The ITU recommendation specifies quite clearly a number of points. In the front array, the Left & Right
loudspeakers are spread 60 to 70 degrees apart, with the Centre loudspeaker in the middle. Ideally,
all loudspeakers should be of the same type to achieve coherence in the sound field. In any case, the
three frontal L-C-R loudspeakers really have to be of the same type so that no coloration changes
occur when panning sounds across the front stage. This part of the practice seems to be clear and
most parties agree with it.
Concerning the surround loudspeakers, the ITU recommends a positional window between ±100
degrees and ±120 degrees from the centre line. This is applied where there are two loudspeakers to
reproduce the two surround channels. If more than two loudspeakers are used (such as four), an
equal number of loudspeakers have to be placed symmetrically on either side of the centre line on a
circle between ±60 degrees and ±150 degrees. Despite the fact that most recording engineers choose
the ±110 degrees position, there are different views and opinions about these angles. Depending on
the source material and the type of surround effects desired, the choice for the positioning of the
surround loudspeakers can differ.
It is also very important that there is symmetry in the multichannel installation in relation to the room
boundaries. All reflections created by the various nearby boundaries should be identical from left and
right in the time domain, so that the spatial information and the panning of the sources stay as stable
as possible. Furthermore, we also recommend that the listening position is located in the front half of
the room, so that the engineer listens to the maximum of direct-to-reverberant energy. The further
away from the direct energy the engineer is the more reflected energy from the rest of the room
he/she will receive.
Once the array is in place, the loudspeakers must be pointed towards the engineer’s listening position
to obtain the optimal on-axis reproduction. Since all channels are fully discrete, equal care must be
taken for each channel when doing the final multichannel mix down and hence the orientation of the
loudspeakers should be very precise. Some installers use lasers to ensure that they have positioned
the loudspeakers correctly!
Finally, if a projection screen is required in such a multichannel set-up, the screen type should be
acoustically transparent.
It is firmly recommended that the multichannel set-up places loudspeakers at equal distance from the
listening position and is symmetrical relative to room boundaries. In other words, all sources of sound
have to have the same ’arrival time’ at listening position. For arrival time to be identical, the sources
have to be positioned at an equal distance from that listening position, on a circle radius. This places
constraints on the room design but it is essential to respect this criterion. As an example, if an audio
element is panned from Centre to Left, and the Centre loudspeaker is offset forwards or backwards by
25 mm (1”), one can expect amplitude ripples in the 500 Hz region due to signal arrival delays.
The vertical positioning of the loudspeakers is also important, but less critical that the horizontal. The
brain has very high capability to localise information in the horizontal plane (azimuth angle), but it is
not so good in the vertical plane (zenith angle). The physical position of our ears explains that fact
very clearly.
The ideal positioning of the three front loudspeakers is where the three acoustical axes (as defined in
the data sheet for each product) are positioned at exactly the same height. The ITU-R BS 775-1
recommends that all loudspeakers should be ideally placed at the same height.
However, as mentioned above the resolution of the ear/brain in the vertical plane is about 3 degrees
above ear level horizon and 3 to 10 degrees below ear level horizon. Precisely because of this
behaviour of the ear/brain, our vertical localisation tolerance is about 7 degrees. Within these 7
degrees, there is a localisation blur (inaccuracy in positioning a sound source) in the vertical plane.
This allows two sources to be positioned at slightly different heights without the brain noticing the
height difference. This useful human hearing limitation can be exploited when planning the centre
channel position, for example. Some tolerance in loudspeaker height can be allowed, and this without
disturbing the engineer who is doing multichannel work and panning sound sources across the L-C-R
loudspeaker array – see Figure 9.
Less than
7 Degrees
Listening
Position
Figure 9 - Ear/Brain Vertical Localisation Tolerance Applied to Loudspeaker Positioning
The ITU is quite strict on the identical height positioning for the three front loudspeakers, however it
gives some allowances for the surround loudspeakers. They can be placed higher than the front
loudspeakers, and tilted down towards the listening position up to an angle of 15 degrees.
Other recommendations, such as the EBU Tech 3276 and the Japanese HDTV Multichannel Sound
Forum, specify that all channels have to be placed at the same height, including the surround
channels. They can then be all tilted down towards the listening position if required. The Japanese
HDTV Multichannel Sound Forum specifies that, “…All loudspeaker axes should be placed at the
same height between 1.2…2.0 m, more than 1.2 m is recommended…”
Concerning the overall height of the surround system, the ITU states that, (…) the frontal
loudspeakers should ideally be placed at a height approximately equal to that of the listener’s ears
(…). This height is specified as 1.2 m from the floor.
This is an important guideline but acoustically when using large loudspeakers the interaction with the
floor (below 400 Hz) and the first reflection cancellation (typically between 100 Hz…200 Hz) can be
very serious if the loudspeakers are placed too low and close to the floor surface. As seen in stereo
control rooms, the overall height of the loudspeakers is dependent on the room geometry and the
listening distance. The larger the loudspeakers the further away from the floor they should be placed
to reduce cancellation phenomena. Furthermore, in standard stereo control room construction, studio
designers always observe the accepted rule that the main loudspeakers should not be placed higher
than 15 to 20 degrees from the ear level horizon. For our brain, if this loudspeaker vertical angle is
higher than the above recommended values, the aural and visual coherence becomes inaccurate in
our perception.
As in stereo rooms, the relative loudspeaker height in control rooms is dependent upon:
However, some recent research (Dr. David Griesinger4) suggests that for best results in terms of
psycho-acoustical perception (spaciousness and envelopment), there should be two subwoofers
placed on either side of the room and driven with the same signal but with a 90 degree phase
difference between them.
One misleading ‘recommended’ place for the subwoofer is in the front, on the floor and in the middle of
the room, equidistant from the sidewalls. This position can, in fact, be a serious acoustical
compromise, since the subwoofer sits in the first pressure minimum of the lateral standing wave.
Hence the frequency response in that location will most likely display serious irregularities.
3
“Perceptibility of Direction and Time Delay Errors in Subwoofer Reproduction”, Juhani Borenius, AES Preprint
2290; and “Loudspeaker Reproduction: Study on the Subwoofer Concept”, Christoph Kügler and Günther Theile,
AES Preprint 3335.
4
“Speaker Placement, Externalisation, and Envelopment in Home Listening Rooms”, D. Griesinger, AES
September 1998.
Recommended positions for subwoofers are (subwoofer always on the floor in these cases):
• Close enough to the front wall and slightly offset from the middle of the room to avoid the first
pressure minima position.
• In a corner, close to both front and sidewalls. This position will maximise the efficiency due
to the corner loading.
• Adjustment of the gain (input sensitivity) and frequency response (Bass Roll-off) of the
subwoofer is necessary since the acoustical loading has changed relative to calibrated
anechoic conditions.
• The subwoofer can also be flush mounted in the front wall but the discussion of the position
of the source relative to the room remains valid.
• The phase adjustment on the subwoofer at the crossover frequency is important to achieve
and keep a flat frequency response in the crossover region.
Sometimes the rear and/or side loudspeakers also need low frequency extension. There are two
possibilities:
• The bass management system can also include these rear/side channels and their low
frequency content will be reproduced through the subwoofer(s) as well. In this case, it is
highly important that all ‘satellite’ loudspeakers are placed on the circle radius so that the
arrival time difference between each loudspeaker and the subwoofer is the same. It must be
also noted that the acoustical behaviour of the room at low frequency will greatly affect the
way the low frequencies are summed and replayed by one single subwoofer. The low
frequency acoustical characteristic of the room should be as identical as possible in the front
and rear part of the room.
• A separate subwoofer is used for the rear loudspeakers. The location of this subwoofer
should be in the rear part of the room and positioned in a similar fashion to the subwoofer for
the front channels. If the room has very different low frequency acoustical characteristics
between the front and the back part, the option of dual subwoofers might provide much
better monitoring results than the single subwoofer choice.
In general, regardless of the obvious benefits of low frequency extension, the use of the subwoofer(s)
and the LFE channel is currently causing many problems. These should be overcome with proper
bass management system and better defined standard practices of the use and content of the LFE
channel. As mentioned several times already, we are in an evolving phase of multichannel
reproduction and more research and experiments are needed to find the best ways to utilise these
available channels.
7.5 cm (3")
Concrete
wall
Absorbing Absorbing
material material
Concrete Concrete
wall wall
Figure 11 – Subwoofer flush mounting recommendations showing various mounting solutions and materials.
Note: concrete wall structures should be preferred over heavy wooden walls.
The positioning of the listening point at the centre of an imaginary circle is as important as the
positioning angles. It is strongly recommended, once again, that all loudspeakers be of the same type
and brand to guarantee similar and coherent responses of the multichannel monitoring system. In a
small room, small multichannel systems are suggested and it is a good idea to locate the
loudspeakers as close as possible to the walls or boundaries whilst respecting the identical radius
from the listening position.
Once the loudspeakers are against a boundary the large baffle created by the walls provides a
hemispherical radiation load for the omni-directional low frequencies (below 200 Hz) and hence a level
gain of +6 dB will be seen below 200 Hz. Consequently, the bass response of the loudspeakers has
to be adjusted so that the frequency response remains flat. The alternative is to reduce the radius of
the loudspeakers but this can lead to problems with low frequency cancellation dips from the wall
behind the loudspeaker. Optimisation of this was discussed in Chapter A2.
Also, it is important to position small loudspeakers on separate stands as far as possible from any
reflective surfaces. Although it is common practice, loudspeakers should not be placed on the meter
bridge of a console. Conventional two-way loudspeakers should be placed vertically, not horizontally,
unless the design is coaxial. If placed horizontally, lobbing of the polar pattern will destroy imaging
when the engineer moves slightly off-axis. In addition, a phase cancellation may be seen at the
crossover frequency. For the same reason, two-way centre channel loudspeakers should also be
placed vertically.
• In a small room, the loudspeakers should be placed near walls to avoid cancellations.
• In a large room, the loudspeakers should be placed far away from walls as this gives a better
direct-to-reverberant energy ratio, but be careful of cancellations.
Large loudspeakers should really be flush mounted to achieve the best possible performance.
Detailed recommendations for flush mounting are given in Chapter A6. If this is not possible and the
large loudspeakers have to be free standing, they should then be placed against a solid and
structurally heavy boundary to avoid low frequency cancellations and the frequency response adjusted
accordingly – see Chapter A1. For very large loudspeakers this is not possible, as the depth of the
loudspeaker is large enough to put the loudspeaker at a distance from the wall that will cause a
cancellation in the bass region. Again, note that the Japanese HDTV Multichannel Sound Forum also
specifies that: “flush mounting is desirable to avoid reflections from rear walls”.
The loudspeaker models that are most appropriate for a particular installation depend on a few factors:
• Room size
• Listening distance
• Listening level
• Program spectrum content
By following the three simple steps below and using the selection table (see Table 2), you will be able
to select the appropriate loudspeaker models and subwoofer type for every application:
• Calculate the room volume and find the highest row in the table column ‘Room Volume up to’
that is not smaller than your room volume.
• Measure the listening distance to the centre of the listening area and find the highest row in
the table column ‘Listening Distance up to’ that is not shorter than your listening distance.
• If there are two different rows selected in the previous two steps, select the models from the
row that is lowest in the table, i.e. the larger system of the two if there are two different lines
recommended.
Notes:
These recommendations are for the minimum sized system that can be expected to give a suitable
SPL for most professional audio applications. Larger systems offer higher SPL and an increased
directivity control, so do not be afraid to select larger models in the range than those indicated. The
main point to be concerned about when up rating the system is to keep the whole system in
balance, so do not select very large loudspeakers (1036A’s) for the front channels and small ones
for the rear channels (8030A's) together with a small subwoofer (7070A) for the LFE channel and
the bass management! Additional separate subwoofers can be used on the rear channels too,
although it is not detailed here. If space or finances are limited, the rear channel models can be
compromised slightly by selecting the next sized model down in the Genelec range. For example,
use 8040A's instead of 8050A's.
The column labelled "Typical Listening Distance" is, in our experience, the average distance
between the loudspeaker and the listening position in studio control rooms.
Examples:
1. If the room is 4 m (13') wide, 7 m (23') long and 3 m (10') high, then the room volume is 84 m3
(2990 ft3). This limits the loudspeaker selection to 8040A / 1030A or larger. If the listening distance
is then measured to be say 1.9 m (6' 3") then the selected front loudspeakers are confirmed as
being 8040A / 1030A or larger.
2. If the room is 6 m (19 1/2') wide, 13 m (43') long and 2.5 m (8') high, then the room volume is 195
m3 (6700 ft3). This limits the loudspeaker selection to 1034B or larger. If the listening distance is
then measured to be say 5 m (16'5") then the selected front loudspeakers should then be 1035B or
larger as the 1034B should only be used up to 4.5 m (14'9").
75 (2,600) 2.0 (6'6") 1.3 (4’3") 8030A (1029A) 8030A (1029A) 7050B 7060B
7070A or
85 (3,000) 2.2 (7'2") 1.4 (4’7") 8040A (1030A) 8040A (1030A) 7060B
2 x 7060B
7070A or 7071A or
95 (3,400) 2.3 (7'6") 1.5 (4’11") 8050A (1031A) 8050A (1031A)
2 x 7060B 2 x 7070A 4)
7070A or 7071A or
110 (3,900) 2.4 (7'10") 1.6 (5’2") 1032A 1032A
2 x 7060B 2 x 7070A 4)
7071A or 7073A or
125 (4,400) 3.5 (11'5") 2.3 (7’6") 1037C 1037C
2 x 7070A 2 x 7071A 4)
7071A or 7073A or
170 (6,000) 4.0 (13'1") 2.5 (8’2") 1038B, 1038BC 1038B, 1038BC
2 x 7070A 2 x 7071A 4)
7073A or
200 (7,000) 4.5 (14'9") 2.8 (9’2") 1034B, 1034BC 1038B, 1038BC 2 x 7073A 4)
2 x 7071A
7073A or
240 (8,500) 4.7 (15'5") 3.0 (9’10") 1039A 1038B or larger 2 x 7073A 4)
2 x 7071A
400 (14,000) 5.5 (18'0") 3.5 (11’5") 1035B 1038B or larger 2 x 7073A 3 x 7073A 4)
400 (14,000) 5.5 (18'0") 3.5 (11’5") 1036A 1038B or larger 2 x 7073A5) 3 x 7073A 5)
Table Notes
1)
If the system is planned to be eventually upgraded to surround, it is recommended to select the
subwoofer model from the ‘5-channel surround’ column for future SPL compatibility. In addition,
selecting the larger of the two subwoofers will give additional headroom and lower distortion in a
stereo system.
2)
This table assumes that the bass management system built into the subwoofer is used. This is a
situation normally encountered in the music industry. Fewer or smaller subwoofers maybe sufficient
if the subwoofer is required to reproduce the LFE channel only. This is a situation normally
encountered in the movie industry.
3)
When using the digital input, the 8130A cannot be used with the LSETM Series subwoofer
analogue crossover filters. The subwoofer can be used for reproducing the LFE channel only.
Additional subwoofers of the same type may be required in a larger room with bass heavy program
4)
material.
5)
Subwoofers are not necessarily required for a stereo 1036A installation as these loudspeakers are
already full range. For surround systems, use the monitor section in the console to route the LFE
signal to the front loudspeakers. Alternatively, subwoofers can be used to reproduce the LFE
channel only.
When two or more subwoofers (except for the 7050B as it does not have a sum output) are positioned
close to one another mutual coupling is the fortunate by-product. This is due to the long wavelengths,
associated with low frequencies, causing constructive superimposition. For mutual coupling, the
subwoofers must be placed within ½ a wavelength of one another (85 Hz upper crossover frequency
½ wavelength is approximately 2 m). For example, two subwoofers give a 6 dB increase in acoustical
output at the listening position – see Table 3.
Dolby provides loose guidelines on loudspeaker placement for the different format playback
configurations shown in Figure 12. These various playback configurations do not follow the ITU-R BS
775.1 recommendation mentioned earlier. Loudspeakers are not positioned on a circle radius and no
angles are specified for the surround left and surround right channels. Also, the various loudspeaker
placements shown are closer to a ‘conventional’ consumer home theater set-up than a professional
control room environment.
600 - 900
Back surround Left back Right back
surround surround
In the 5.1 playback configuration, the front left-right angles can spread from 45 degrees (typical movie
theatre set-up) to 60 degrees (ITU set-up). The 6.1 playback configuration displays a single ‘back
surround’ loudspeaker, but a second alternative set-up is possible and illustrated in the following
chapter. For the 7.1 playback configuration (also used for the various matrix formats that provide up
to 7.1 channels of audio), Dolby suggest angles between ± 135…150 degrees from the front centre-
line for the ‘left and right back surround’.
The conventional five channels ITU set-up is illustrated in grey in Figure 13. The additional sixth
channel is the rear centre channel. There does not seem to be an agreement between parties
whether a single loudspeaker or two loudspeakers should be used to reproduce this rear centre
channel. Some guidelines suggest that using two loudspeakers between ±135 - 160 degrees from the
centre line would provide better envelopment and stable sound-field, especially if the circle radius is
large. When the radius is smaller, the use of a single loudspeaker at 180 degrees seems to be
sufficient and appropriate, but might start to interfere with the centre front channel. Further experiment
and research seems to be needed to agree on an appropriate and standardized set-up.
Center
Sub
Left Right
30 0
110 0
Multiple suggested configurations exist for 7.1 audio reproductions. There are seven configurations
that are supported on new A/V receivers, which feature loudspeaker re-mapping and user definable
down mixing to 5.1. Dolby TrueHD and DTS-HD configurations agree on the loudspeaker setup
illustrated in Figure 14. Dolby suggests angles between ± 135…150 degrees for the surround
channels.
Center
Sub
Left Right
30 0
150 0
Figure 14 - 7.1 Loudspeaker Set-up with two side loudspeakers at 90 degrees and two surround loudspeakers at 150 degrees
Dr. David Griesinger has conducted extensive investigations regarding the positioning of rear
channels. The Logic 7 surround sound format is one of the results of his work – see Figure 15.
This format uses 2-channel stereo or 5.1 program material as the input signal and extracts, via a
matrix system, four different surround channels. Two of these channels are called side channels and
are placed at ±90 degrees from the centre line. The two other channels are placed between ±110
degrees (ITU compatible) and about ±140 degrees.
Dr. Griesinger is also advocating for the use of two subwoofers placed on either side of the room (±90
degrees) to provide a better “envelopment effect.” The subwoofers have a 90 degree phase shift
between them.
Center
Left Right
30 0 Sub
110 0
~140 0
Surround Left Surround Right
Figure 15 - Logic 7 Loudspeaker Set-up with four surround channels and two de-correlated subwoofers.
Tomlinson Holman is behind the development of this TMH Sound System format – see
Figure 16. Designed mainly for movie theatres, it includes the standard 5.1 loudspeaker set-up and
placement (ITU set-up in grey), additional Extreme Left and Extreme Right channels at ±60 degree (for
a better frontal spread), a Centre Rear channel (for better rear imaging) and two channels at ±45
degrees both in the horizontal and vertical plane (to give some sense of height). This set-up also uses
two subwoofers that have a 90 degree phase shift between them.
Center
Left Right
Sub
Horiz 45 0 Horiz 45 0
Vert 45 0 30 0 Vert 45 0
110 0
Rear Center
Figure 16 - TMH Sound System Loudspeaker Set-up, with eight channels in the horizontal plane, two channels at 45 degree
and two de-correlated subwoofers.
In the scope of the next generation broadcasting system, the Japanese Broadcasting Corporation
(NHK) has developed this ‘ultimate’ 22.2 multichannel audio system for ultrahigh-definition video with
4000 scanning lines5. The 22.2 reproduction system has three vertical layers of loudspeakers with 2
LFE channels, namely 3 loudspeakers in the bottom layer, 10 loudspeakers in the middle layer and 9
loudspeakers in the upper layer – see Figure 17. The source materials used to develop and evaluate
this system were various multi-track recordings and different mix-downs of these tracks.
LFE 1 LFE 2
30 0
120 0
180 0
Above Viewer Level Speakers
Viewer Level Speakers
Floor Level Speakers
Figure 17 – 22.2 Japanese Broadcasting (NHK) Loudspeaker Setup for ultra-high definition pictures
5
“Advanced Multichannel Audio Systems with Superior Impression of Presence and Reality”, Kimio Hamasaki, Koichiro
Hiyama, Toshiyuki Nishiguchi, and Kazuho Ono, NHK Science & Technical Research Laboratories, Tokyo, Japan.
th
Paper 6053, presented at the 116 AES Convention, May 2004, Berlin, Germany.
In a 5.1 audio production, the five main channels (Left, Centre, Right, Surround Left and Surround
Right) are all full bandwidth, i.e. 20 Hz to 20 kHz. Before encoding, the LFE channel is NOT band
limited, i.e. it is just another full bandwidth channel. Figure 18 illustrates the audio bandwidth of each
channel in a 6.1 production, where the 6th rear surround channel is also full bandwidth at the recording
stage.
1.
Full
Full
band
band 1.
Full
Full
band
band
2. Full band 2. Full band
3. Full band 3. Full band
4. Full band Encoding Data Stream 01001100... 4. Full band
5. Full band 5. Full band
6. Full band 6. Full band
LFE Full or Pre filtered LFE low passed
After the LFE channel is being encoded it usually has a limited bandwidth (hence the label ‘.1’). This
encoded bandwidth will range from 20 Hz up to various upper cut-off frequencies depending on the
encoding formats – see Table 4. Note that some recent multichannel formats deviate from this
principle as the LFE channel stays at full bandwidth even after the encoding stage. In these cases, the
naming ‘.1’ is consequently not technically correct.
(1)
Table Note : in all DTS ‘Coherent Acoustic’ codec the LFE channel is band limited at the encoding
stage at 120 Hz. In the DTS-HD codec only [DTS-HD Master Audio™ (lossless) and DTS-HD High
Resolution Audio™ (lossy)] the DTS-HD decoder applies a low-pass filter at 100 Hz (-3 dB) with a 60
dB/octave roll-off onto the LFE channel.
The .1 channel was originally designed for use in theatres where the main channels could not
reproduce the lower frequencies and additional headroom was required at these low frequencies to
reproduce high SPL. It is often given different names depending on its use:
The use of the LFE channel is not consistent throughout the audio industry, as the needs of Movie
Theatres (Cinemas), Home Theatres, Digital TVs and Music only production are all very different.
Also, sound engineers new to multichannel surround sound mixing are experimenting with new ideas
and techniques all the time.
Various multichannel encoding/decoding processes exist, but one consistent feature for all formats is
that all main channels remain full bandwidth and the LFE channel has variable upper cut-offs (or even
a full bandwidth) as already detailed.
If the LFE channel is recorded (to tape, hard disk, etc) at the same nominal level as all the main
channels then a loud explosion effect could easily overload the recording machine. If this were the
case, this would defeat the object of having a special channel for loud sound effects as the main
channels would have to be recorded at a lower level to accommodate the ‘big bangs’ on the LFE
channel. To overcome this problem some encoding systems - Dolby Digital & DTS Digital Surround -
require the LFE channel to be monitored at +10 dB. In doing so, the audio level recorded ‘to tape’ will
have an additional 10 dB of available headroom as the engineer will naturally reduces the LFE
channel level by 10 dB on the mixing desk output to maintain the sound level balance between the
channels. Note that this increase in headroom is at the expense of 10 dB of signal-to-noise ratio, but
as the LFE channel is eventually band limited in the encoder, it is a price worth paying.
In Dolby Digital and DTS Digital Surround formats, between the LFE channel monitor section fader
and the acoustic output of the monitoring system, there must be a +10 dB level change - see Figure
19. This can happen in various places in the audio chain (see Chapter D 3), but it must be
implemented during the production stage. Note that there are NO level changes between the console
outputs and the inputs to the tape or hard disk recorder. The net effect is that there is a 10 dB
additional headroom on the LFE channel level recorded to tape compared to the main channel levels.
All the tracks are then encoded into a single bit-stream using various encoding schemes (most
commonly with Dolby Digital or DTS Digital Surround) and replayed accordingly - see Figure 20.
LFE Channel
+10dB
Figure 19 - Complete Multichannel Audio Production Chain (Dolby Digital & DTS Digital Surround)
+10 dB
0 dB 0 dB
Full band
Full band
LFE LFE
Multitrack
LFE
+10dB
L C R LS CS RS LFE L C R LS CS RS LFE
Monitoring Bus Recording Bus
Output Output
Mixing Console Monitoring Section
Figure 20 – Mixing console 6.1 monitoring section with LFE channel initial gain structure for both the monitoring and the
recording output busses. The LFE channel output level is then reduced to achieve additional 10 dB of recorded headroom for
the LFE channel.
The 10 dB gain on the LFE channel has to be implemented during the production stage to achieve
correct level before encoding the data. Only one +10 dB gain is needed but two choices exist for the
placement of the gain stage in the monitoring path:
1. If a large format console is used, the output matrix of the monitoring section will usually provide an
internal +10 dB gain assignable on the LFE channel buss. In this case, no additional +10 dB gain,
as provided by the Genelec bass management system, should be used.
2. If the monitoring section of the console cannot provide a +10 dB gain on the LFE channel buss,
then the ‘LFE +10 dB’ DIP-switch provided on the Genelec bass management (7060B, 7070A,
7071A and 7073A) should be used.
As many different pieces of equipment are connected and used in a control room, the production
engineer has to be aware of the detailed audio path of the monitored signal. The +10 dB gain on the
LFE channel may, or may not, be necessary depending on the signal source. Here are some cases
when the 10 dB LFE boost of the Genelec bass management system should not be used:
• If the +10 dB gain on the LFE channel is already implemented by another device (mixing
console monitoring output or an external device).
• When producing an audio format that does not require the use of the +10 dB gain on the
LFE channel at the decoding stage, such as DVD-Audio (MLP), SACD (DSD/DXD), etc.
Please note when Dolby Digital or DTS Digital Surround recordings are “re-mastered” on
DVD-Audio tracks the +10 dB gain will have to be utilized.
• When Dolby Digital Encoder/Decoder devices are connected in the monitoring path. In this
situation, the 10 dB gain on the LFE channel will already be provided by the decoder.
• If a DVD player with built-in decoder is connected and monitored through the console output
buses, then the +10 dB gain on the LFE channel is already provided by the decoding stage.
As seen above the LFE channel will be band limited after encoding, therefore many engineers suggest
that it is highly important to filter the LFE channel at the production stage.
In the Dolby Laboratories publication, ‘5.1 Channel Production Guidelines' it clearly states about LFE
channel:
“A low-pass filter must be inserted into the LFE signal path during the mix process to
ensure proper monitoring. Furthermore, the filter must be applied to the signal being
recorded so that the results will be consistent, whether delivered by Dolby Digital or Linear
PCM.”
“Normally, in low end decoders, one cannot change the low-pass frequency say to 100
Hz or 120 Hz.”
That is mainly why this pre-filtering should occur in the production stage for all consumer multichannel
productions.
Even if various recommendations exist concerning what should be put in the LFE channel, technically
anything is possible. This has even more truth with formats like MLP (DVD-Audio) and DSD/DXD
(SACD) that have an LFE channel that has full bandwidth after encoding (!). The production engineer
has mainly two choices:
• Coherent Signals
The main channels and the LFE channel receive the same low frequency content. The LFE is
used just as an additional track to put some more low frequency energy. That is in some
ways dangerous because there might be non-coherent radiation of the same programme
content in the room by both the subwoofer and the main loudspeakers. This results in
unpredictable acoustical summing of audio, which leads to inconsistent reproduction highly
dependent on the listening environment.
• Non-Coherent Signals
The LFE channel has different audio content than the low frequencies mixed in the main
channels. This is the general guideline and practice followed by the movie industry that is far
safer in terms of the acoustic summing found in different listening rooms.
Various bodies have given recommendations on how to use the LFE channel and what to put in that
channel. This list is not exhaustive, but gives some ideas and suggestions and various opinions that
exist amongst audio professionals:
subwoofer channel is an option at the receiver and thus should only carry the additional
enhancement information. (…) The LFE channel should be capable of handling signals in the
range 20 Hz – 120 Hz”.
The use of the LFE channel for music only productions is strongly discouraged by many
experienced recording engineers because:
- All main channels contain full bandwidth that can allocate any low frequency information.
- No LFE channel leads to less potential replay error at the consumer end.
In ‘story telling’ TV productions, it is highly recommended not to use the LFE channel and its
content to avoid low frequency masking on the essential dialogue parts.
The ITU recommendation is thus very important for the broadcast industry that has to produce
multichannel material that is 5.1 and 5.0 compatible.
As there are six (or more) channels to be replayed, it may seem logical to connect each output of the
multichannel source directly to the appropriately positioned loudspeakers and to the subwoofer(s).
However, without taking care that the entire bandwidth of each channel is properly monitored (with
bass management or redirecting the LFE channel correctly), losses of important parts of the audio
bandwidth or even serious cancellation effects can be induced, both electrically and acoustically.
Firstly, unless they are full bandwidth, each loudspeaker replaying the five or six main channels will
have a -3 dB low frequency cut-off that is higher than 20 Hz. Therefore, the lowest frequencies of the
main channels will not be reproduced and monitored. This is a serious compromise as all five or six
main channels are full bandwidth even after encoding and decoding.
Secondly, Genelec subwoofers have their crossover point fixed at 85 Hz. After many investigations
and subjective listening tests a fixed 85 Hz frequency has been chosen for best acoustical results. If
one connects the encoded LFE channel to the subwoofer and the main loudspeakers are not
connected to the outputs of the subwoofer, only audio material up to 85 Hz will be monitored via the
subwoofer. No information above that frequency will be heard – even though the LFE upper cut off
can be much higher!
That is why the Genelec 6.1 Bass Management system that is built-in to the LSETM series subwoofers
should be used, as this enables the rest of the bandwidth to be replayed (see E Bass Management
System chapter).
In small studios with basic acoustic treatment and geometry, the frequency response of a loudspeaker
system below 100 Hz is dominated by the modal response of the room. Strong low frequency
standing wave patterns can be observed in small rooms that have parallel walls. Therefore, it is very
difficult to achieve consistent low frequency response from multiple full-range loudspeakers in such
limited space.
One solution to this basic acoustical problem is to employ a system called ‘Bass Management’. Using
active electronic filters and crossovers, one can extract the low frequency information from the main
channels and route that information to a single subwoofer feed. The low frequencies are now
originating from one single source that can be placed in an optimum position in the room.
Furthermore, the LFE channel can also be monitored via this subwoofer and added to the low
frequencies of the other main channels. Therefore, the Bass Management’s basic and main goal is to
ensure that the entire audio bandwidth of all channels can be accurately monitored.
Dolby Laboratories
“Bass Management allows the user to redirect low-frequency information from any of the
main loudspeakers to the subwoofer.”
Now, for the subwoofer bandwidth, a fixed crossover point has been carefully chosen at 85 Hz. This
means that for the Bass Management function of the six main channels, the crossover point between
the subwoofer and the main channels is set at 85 Hz.
If we look at the bandwidth of the LFE channel after encoding, (see Chapter D1) we observe that this
bandwidth can vary quite significantly. For that reason, the Bass Management for the LFE channel
bandwidth provides three different setting possibilities for reproduction.
Note that the setting of the LFE subwoofer bandwidth will not at all affect the encoding of the LFE
channel in the various formats, but does provide different replay bandwidths only.
• Six main channels of inputs & outputs are provided for low frequency redirection. Each
channel is identical and has a fixed crossover frequency set at 85 Hz.
• Bass management Bypass Function is provided for the six main channels. The LFE channel
is not affected by the bypass function, regardless of the LFE bandwidth and redirection
settings. The bypass function can be activated via an optional remote control box (RJ11) or
via the ¼” jack bypass connection.
• Selectable +10 dB LFE gain for the LFE channel monitoring. This function can be either
selected in the bass management located on the subwoofer’s side panel, or remotely via an
optional remote control box (RJ11 connection).
• ‘Slave’ subwoofer function, selected with the ‘Sum In’ DIP-switch. This selection must be
done to the ‘slave’ subwoofer to guarantee proper operation and all other switches of that
particular switch group should be set to the ‘OFF’ position. The ‘slave’ subwoofer receives
only one input (LFE IN/SUM IN) and follows automatically the parameters selected in the
‘master’ subwoofer.
• The discrete ‘Sum Out’ output sends the total subwoofer signal from the ‘master’ unit to
another subwoofer(s). This gives the possibility to daisy chains multiple subwoofers to
achieve higher SPL at the listening position relative to the room size.
• The three-colour power LED mounted on the side panel of the subwoofer indicates signal
‘On’ (green), ‘Clipping’ (orange) and ‘Driver Protection’ (red). This status LED can also be
mounted remotely in any control surfaces, using one of the optional (RJ11) remote control
boxes.
• Variable input sensitivity from +12 dBu to -6 dBu (except for 7073A: variable from +6 dBu to
-12 dBu)
The typical connection of the Genelec 6.1 Bass Management is shown in Figure 21. All channels
including the LFE channel are connected through the Bass Management. The main six channels are
then connected from the output of the bass management to each of the main loudspeakers. A single
output from the ‘master’ subwoofer (Sum Out) is fed into the ‘slave’ subwoofer unit (LFE In/Sum In).
L C R
LFE
SL SC SR
The first possibility for the LFE input bandwidth is the reproduction of the LFE channel by the
subwoofer up to 85 Hz and then the redirection of the energy above 85 Hz to the front centre
loudspeaker. This means that all information present on the LFE channel will be reproduced through
the monitoring system, whatever the upper cut-off frequency of the LFE channel – see Figure 22.
Any noise, distortion artefacts and other unwanted sounds on the LFE channel can be monitored in
addition to the LFE signal itself. In addition, any type of LFE content (coherent or non-coherent audio
material) will be properly monitored without unpredictable summing of the low frequencies between
the different channels. This should be the default setting for the bass management as it gives full
range monitoring of the LFE channel.
SL
SC
SR
Figure 22 - Multichannel Monitoring with Genelec 6.1 Bass Management System. Redirection of the LFE channel signal above
85 Hz to the front centre loudspeaker.
The second possibility for the LFE input bandwidth is the reproduction of the LFE channel by the
subwoofer up to 85 Hz, without any redirection above 85 Hz to the front centre loudspeaker. This
setting is comparable to pre-filtering the LFE channel in the mixing console using an 80 Hz LP filter.
No information above 85 Hz present on the LFE channel will be heard through the monitoring system
– see Figure 23.
This setting is not recommend for everyday use as the LFE channel has a bandwidth up to 120 Hz in
Dolby Digital and DTS formats and wider bandwidths in some of the other formats. A good use for this
setting is to simulate the effect of some (usually cheaper) consumer decoders that do not replay the
LFE channel information that is above 80 Hz when the bass management is used. It is important for
the mixing engineer to be aware of such a limitation so that the multichannel mix translates well in the
home environment.
This configuration provides a replica of what will happen in many home surround system situations
and yields to consistent reproduction of the low frequencies below 85 Hz. Also, any type of LFE
content (coherent or non-coherent audio material) will be properly monitored.
As mentioned earlier, Dolby Laboratories and other companies recommend that, in DVD mixes for
consumer home applications, the LFE channel is monitored and recorded with an 80 Hz low pass filter
inserted in the console output buss. This second LFE channel bass management setting is a quick
and easy way to simulate the monitoring part of that recommendation.
L
C
SL
SC
SR
Figure 23 - Multichannel Monitoring with Genelec 6.1 Bass Management System. LFE Input band-limited to 85 Hz.
The third possibility for the LFE input bandwidth is the reproduction of the LFE channel by the
subwoofer up to 120 Hz without any redirection above that frequency. This means that no information
above 120 Hz on the LFE channel will be heard through the monitoring system – see Figure 24.
This setting has been provided so that production facilities can emulate the replay systems that exist
in movie theatres and cinemas. For these applications, there are strict rules on how the replay system
should perform. Dedicated subwoofers with a bandwidth from 20 Hz – 120 Hz reproduce the low
frequency content for the 35 mm movie soundtracks. However, it must be noted that when mixing
music and sound effects for film release, engineers always use non-coherent low frequency
information between LFE channel and main channels. In other words, the low frequency LFE channel
content is different from any other main channel low frequency audio content so that, potentially,
unpredictable acoustic summing in the room is avoided.
L
C
SL
SC
SR
Figure 24 - Multichannel Monitoring with Genelec 6.1 Bass Management System. LFE Input band-limited to 120 Hz.
The LFE channel is unaffected by the bypass function as it affects the main channels only. This
means that the low frequency extension of all main channels, provided by the subwoofer, is bypassed
and the full bandwidth is monitored via the main loudspeakers only. The low frequency reproduction
of the main channels will extend as low as the main loudspeaker low frequency extension (i.e. for the
8040A: -3 dB @ 45 Hz). At the same time, the LFE input remains untouched and the settings
concerning ‘routing to the centre channel’ and ‘bandwidth’ (85 Hz or 120 Hz) will be unchanged – see
Figure 25.
The bypass function can be activated in two different ways in the 7060B, 7070A, 7071A and 7073A
subwoofers:
1. Via the mono ¼” jack connector on the subwoofer side panel using a standard mono
shorting switch.
2. Via the side panel RJ11 remote connector. The optional remote control box provides
switching and status LED for the bypass function.
Note that the ¼” jack bypass switch should be set to ‘OFF’ for the RJ11 bypass remote control to work
and display the status of the bypass function.
ed
R
ss
pa
SL
By
SC
SR
LFE Redirection to Centre speaker
85 Hz upper cut-off
120 Hz upper cut-off
Figure 25 - Multichannel Monitoring with Genelec 6.1 Bass Management System. All main channels have the subwoofer
bypassed, and the LFE channel remains unchanged.
When using the bass management RJ11 connection, two remote control boxes can be connected to
the subwoofer RJ11 socket. Both remote boxes have dual RJ11 sockets for daisy chaining the control
signal from one box to the next.
The subwoofer input sensitivity is variable from +12 dBu to -6 dBu (except for the 7073A,
variable from +6 dBu to -12 dBu). Typical when a subwoofer is placed close to a wall, it
experiences acoustical loading which can be compensated by reducing the input sensitivity by 6
dB. So, the purpose of this sensitivity control is to allow the user to adjust the subwoofer
acoustical level to be aligned with the main loudspeaker’s level.
Bass Roll-off adjustments (-2 dB, -4 dB and -6 dB) can be useful to retrieve a flat subwoofer frequency
response. When subwoofers are placed on the floor and against one or two walls, the effect is not
only an in-band gain in the frequency response, but often a change in the frequency response shape.
The adjustment of the Bass Roll-off will enable the user to achieve a flat subwoofer response.
When the connection to the Genelec Bass Management is properly done, it is possible and necessary
to adjust the Phase at the crossover point once the multichannel system is set-up in the room.
Genelec Bass Management allows for 0, 90, 180 or 270 degrees @ 85 Hz. This guarantees that the
main loudspeakers and the subwoofer are in phase at the 85 Hz crossover point. An 85 Hz tone
generator is provided in the Genelec bass management so that it is easy do the necessary on-site
phase adjustments.
This Phase adjustment has to be done for each subwoofer placed in the room, as they possibly will
have different phase adjustments depending on their different physical location in the control room
relative to the ‘satellite’ loudspeakers. This is especially important in the case where one uses a
master/slave subwoofer configuration.
The 7050B features selectable 85/120 Hz low pass filter for the LFE channel reproduction bandwidth,
selectable 0/+10 dB LFE input sensitivity, phase and ‘Bass Roll-off’ adjustment as well as input
sensitivity adjustment between +12 dB to -6 dB.
In a stereo configuration, the main 8020A’s or 8030A’s can be connected to the 7050B in the same
manner as for a 5.1 connection scheme. Alternatively a pair of 8020A’s or 8030A’s can be connected
by routing the signal cables from the source to the input connectors of the main loudspeakers and
another pair of cables from the main loudspeakers' output connectors to the "IN" connectors on the
7050B. In this configuration the volume controls on the main loudspeakers affect the playback level of
the 7050B too. The "Bass Roll-off" switch (switch 2) on the main 8020A’s/8030A’s loudspeakers must
also be switched to "ON" (switch 2) for the same reasons.
C SL SR
L R
SERIAL
NUMBER
SENSITIVITY ROLL-OFF
FOR 100 dB -6 dB
SPL @ 1m -4 dB PHASE
-2 dB
IN
-270°
-90°
+3 0 -180°
-3 85 Hz
+6
-4 ON
+12 -6
120
LFE IN dBu LFE POWER FRONT FRONT FRONT REAR REAR MAINS INPUT
BANDWIDTH 50 / 60 Hz
LFE + 10 dB
LEFT CENTER RIGHT LEFT RIGHT 120 W 230 V~
LFE
Other engineers in the audio post-production and movie industry sectors want to use subwoofers to
reproduce the LFE channel only. In doing so, they emulate the replay conditions in movie theatres. In
that situation, the audio content of the LFE channel has to be carefully considered as cancellations
may occur between the audio replayed by the main loudspeakers and by the subwoofer.
Whatever the opinions are, it is also possible to monitor accurately all main channels without a
subwoofer when using multiple large full range loudspeakers. The important point is to be able to
monitor each channel in its full bandwidth. Therefore, in this case, the five or six full range
loudspeakers will reproduce the five or six main channels and the LFE channel should be monitored
via one, or more, of the full range loudspeakers. Re-routing is required and can be done inside the
mixing desk’s multichannel monitoring section or in a separate multichannel monitor controller – see
Figure 27. The most widely used routing for redirecting the LFE channel is to send it to the centre
channel. There are also other set-ups where one redirects the LFE channel to front left and front right
loudspeakers (note that decoders do this if there is no subwoofer) or even to all three front
loudspeakers (this is acoustically the best solution for SPL handling reasons).
Note that when routing the same LFE channel to more than one loudspeaker, the total SPL of that
channel will increase. For example, routing the LFE to the front left and right loudspeakers will add 6
dB SPL to the overall LFE channel level, therefore, the output level of the LFE channel should be
reduced accordingly in the console monitoring matrix output.
20Hz 20kHz
L
C
SL
SC
SR
The main goal in the alignment of a multichannel system is to achieve that the subwoofer sound
output level (within the band limited subwoofer frequencies) is the same as the sound output level
compared to the main loudspeaker system (at other frequencies). The LFE output of the mixing desk
or decoder should be connected to the LFE input on the subwoofer.
To ensure and achieve repeatable results in the finished production, the SMPTE (Society of Motion
Pictures and Television) organisation has developed standard monitoring levels for cinema
postproduction work. For film mixing, the SMPTE reference is 85 dB, with full bandwidth pink noise,
read with a sound pressure level meter (SPL) set on C weighted/Slow scale.
For release of film material on television, various standards state that the operating mixing level
should be somewhat lower so that low-level dialogues, which are easily heard in a quiet and
acoustically well treated control room, are mixed slightly higher. This is to ensure that in the home
environment with higher background noise levels the dialogue is clearly heard. However, for music
mixes, there are no standardised levels – as for stereo – because the level that the engineer chooses
is arbitrary and based on personal taste, as is the level chosen by the end user.
Thus, one absolute reference level does not really apply for all multichannel surround sound
applications, so the following alignment procedures do not mention specific values but rather the
relative levels between the various loudspeakers at the listening position.
6.3 Calibration of the Level and Frequency Response with an MLS Signal
First, there is no point in setting the channel levels until the loudspeakers have been calibrated for
their individual frequency responses. Furthermore, there is no point in calibrating the frequency
responses using the acoustic tone controls if there are fundamental acoustical problems in the room.
Level setting is the last setting to be made once all the other issues have been resolved.
The acoustical performance of the main channels and the subwoofer should have a flat frequency
response for accurate monitoring. First, make sure that the switch labelled ‘LFE +10 dB’ is set to ‘off’
position. Then proceed as follows:
• Calibrate the five or six main channel frequency responses using an MLS or similar
measuring system with the subwoofer bypassed or disconnected.
• Then connect the Genelec subwoofer as described in the supplied operating manual and
adjust the subwoofer level, bass roll-off and phase controls (relative to the centre channel)
so that the measured frequency response of the subwoofer and near field monitor is
extended smoothly down to 19 Hz (7060B, 7070A, 7071A and 7073A).
Note that there should be no 10 dB level changes at low frequency (in the subwoofer bandwidth)
compared to the mid and high frequencies as all the headroom level changes of the LFE channel are
done electrically in the mixing desk, the bass management system or the decoder.
If MLS type equipment is not available for aligning the system then follow the guidelines that can be
found in the operating manual for adjusting the frequency response. Remember though that there is
still the need to align the system for level. Below are listed two alternative methods but one should
note here that the accuracy of these methods depends greatly on the quality and the low frequency
response of the SPL meter. First, make sure that the switch labelled ‘LFE +10 dB’ is set to ‘off’
position in the subwoofer bass management.
• Level Calibration using a 1/3 Octave Real Time Analyser, Broadband Pink Noise and
an SPL Meter
Connect the Genelec multichannel system and play broadband pink noise signal (20 Hz – 20
kHz) through the subwoofer and one of the main channels, for example, the centre channel.
Set the level of each band on the RTA Analyser to read the same value, both in the
subwoofer bandwidth and in the main loudspeaker bandwidth. The specific absolute
reference level depends on your application area as mentioned earlier and can be checked
using the sound pressure level meter. Place pink noise through the whole system and adjust
the level of each channel on the mixing desk to give the same acoustical level in the room.
The level depends on the level of the signal but if it is set to -20 dBFSrms (-18 dBFSrms in
Europe) the SPL meter should read 85 dB for cinema/theatre work, 80-85 dB for television
and 85-95 dB for music.
¾ Filtered pink noise from 500 Hz to 2 kHz for the mid band frequencies of the main
loudspeakers. This range is well away from the subwoofer’s bandwidth and suffers less
from room cancellation effects.
Note: If the recorded level of both the filtered pink noises is –20 dBFSrms (-18 dBFSrms for
Europe) then an absolute level calibration can be made according to your application area, for
example, 83 dB SPL for theatrical work, 78-83 dB SPL for television and, typically, 83-93 dB
for music. This level is 2 dB lower than for broadband pink noise as there will be less energy
in the room due to the limited bandwidth signal.
¾ Connect the Genelec multichannel system then play the 500 Hz to 2 kHz filtered pink
noise to adjust each of the main channels individually. Set the SPL Meter to C-weighting
& Slow scale, and note the reading (say it is 83 dB SPL). All five channels have to be
adjusted to this same level.
¾ Next, play the 20 to 80 Hz filtered pink noise through the subwoofer. The correct
adjustment should give a reading 3 dB lower than the one for the main channel
loudspeakers (in our example, 80 dB SPL). The reason for the difference in level reading
is that most SPL meters have a built in HP filter. If there is no HP filter in the SPL meter
then the reading should be the same as for the main loudspeakers as the bandwidth, and
hence energy in the room, of the two signals is the same.
F REFERENCES
1 Technical Papers
The following technical papers and recommendations were used in compiling this paper:
• EBU, Technical paper 3276 – 2nd Edition; “Listening conditions for the assessment of sound
programme material: monophonic and two channel stereophonic. Multichannel Sound:
Supplement 1”. February 1999, Geneva.
• Surround Sound Forum (VDT, IRT and SRT), Recommended Practice SSF-02/1-E-2,
“Multichannel Recording in 3/2 Format”.
• DTS, Technical White Paper, “Delivering High Quality Multichannel Sound to the Consumer”,
SMF. Smyth, WP. Smith, MHC. Smyth, M. Yan, T. Jung, AES Copenhagen, May 1996.
• “A Proposal for the High Quality Audio Application of High Density CD Carriers”, Acoustic
Renaissance for Audio, R.J. Stuart 1995.
• Digital Audio Denmark, Technical papers; “The advantages of DXD for SACD”, Resolution
Magazine, July/August 2004, pages 60-61.
• EBU, Technical paper 3306: “RF64: An extended File Format for Audio”, Geneva, January
2006.
2 Website Sources
The following Websites sources were also consulted for compiling this paper:
• Philips www-eu.sv.philips.com/mpeg/html/multichannel.html
• MPEG Audio FAQ http://drogo.cselt.stet.it/mpeg/faq/faq-audio.html
• Via Licensing (by Dolby) www.vialicensing.com/products/mpeg4aac
• Ambisonic System www.ambisonic.net
• VBAP www.acoustics.hut.fi/research/abstracts/vbap.html
• Digital Audio Denmark (DXD) www.digitalaudio.dk/technical_papers
• Merging Technologies (DSD/DXD) www.merging.com
• European Broadcasting Union www.ebu.ch
• HD-DVD http://en.wikipedia.org/wiki/HD-DVD
• Blu-ray Discs http://en.wikipedia.org/wiki/Blu-ray_Disc
• Holographic Versatile Disc (HVD) http://en.wikipedia.org/wiki/Holographic_Versatile_Disc
• Holographic Storage www.inphase-technologies.com/technology
* All products indicated by a trademark (TM) or registration (®) symbol are trademarked and/or
registered by their manufacturers.