Max For Live Ultimate Zen Guide

Max for Live Ultimate Zen Guide
Become a Max for Live Master and discover a new way of

using Ableton Live
Julien Bayle
This book is for sale at http://leanpub.com/Max-for-Live-Ultimate-Zen-Guide
This version was published on 2013-11-17
ISBN 978-2-9546909-4-0
This is a Leanpub book. Leanpub empowers authors and publishers with the Lean Publishing
process. Lean Publishing is the act of publishing an in-progress ebook using lightweight tools and
many iterations to get reader feedback, pivot until you have the right book and build traction once
you do.
©2013 Julien Bayle

Tweet This Book!
Please help Julien Bayle by spreading the word about this book on Twitter!
The suggested tweet for this book is:
I just bought Max for Live Ultimate Zen Guide written by @julienbayle & reviewed by
@markmsb / #ableton #book #m4l
The suggested hashtag for this book is #m4lbook.
Find out what other people are saying about the book by clicking on this link to search for this
hashtag on Twitter:
https://twitter.com/search?q=#m4lbook
Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i
About . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
About the review & edit of the book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
FAQ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Version . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
This guide is not . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
This guide is . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
If you find a mistake or a typo in this guide . . . . . . . . . . . . . . . . . . . . . . . . . . v
If you would like to find out about courses on Ableton Live, Max6 . . . . . . . . . . . . . v
1 Max for Live Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1 What ? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Ableton Live Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.3 Max for Live specific objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2 Live Object Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2.1 What is LOM ? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.2 LOM’s Objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.3 How to read the documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.4 Children, Properties & functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.5 Path, ID & navigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.6 One shot request with get and live.object . . . . . . . . . . . . . . . . . . . . . . . . 30
2.7 Observing properties in Live with live.observer . . . . . . . . . . . . . . . . . . . . . 34
2.8 Calling objects’ functions with live.object . . . . . . . . . . . . . . . . . . . . . . . . 35
2.9 Changing properties’ values by using set & live.object . . . . . . . . . . . . . . . . . 37
2.10 Controlling continuous parameters in Live with live.remote∼ . . . . . . . . . . . . . 39
2.11 List and use objects’ children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.12 live.thisdevice tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.13 Mode Preview & performances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3 JavaScript in Max for Live . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

3.1 How to use Javascript in Max for Live ? . . . . . . . . . . . . . . . . . . . . . . . . . 47
CONTENTS
3.2 Controlling Live API through JavaScript . . . . . . . . . . . . . . . . . . . . . . . . . 49
4 Creating your own Max for Live devices . . . . . . . . . . . . . . . . . . . . . . . . . . 56

4.1 preset & pattr . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.2 Freezing Max for Live’s devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.3 Presentation mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.4 Limitations of Max for Live compared to Max . . . . . . . . . . . . . . . . . . . . . 77
4.5 MIDI & audio specific objects for Live connection . . . . . . . . . . . . . . . . . . . 78
4.6 An example of MIDI instrument . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
4.7 Example of a sound generator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.8 MIDI FX example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.9 Audio FX example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Conclusions & perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

Introduction
I’m Julien Bayle¹ and I have used Ableton Live Digital audio workstation since its first release. I
was certified by Ableton² in 2010 and I teach not only Ableton Live Suite³ but Max6 and Max for
Live both to individuals and groups. I have released an eight and a half hour long training DVD
about Ableton Live 9 & Push⁴
Push & Live on the field
Max for Live was released in 2011, and I have had lots of requests to write a book since then.
Here we are! It is starting now and I hope that you will be totally happy and satisfied after reading
this guide. I want you to be able to design your own Audio & MIDI FX but also your own sound
generators and custom interfaces inside Ableton Live !
¹My website: http://julienbayle.net
²ableton.com: https://www.ableton.com/en/education/certified-training/france/julien-bayle-marseille
³My courses: http://julienbayle.net/ableton-live/
⁴My DVD about Ableton Live 9 & Push: http://www.elephorm.com/audio-mao/formation-ableton-live/ableton-live-9.html#a_aid=julienbayle
Introduction ii
In the first part of this guide, I’m going to explain Max for Live basics: what Max for Live exactly
is, what the differences with Max6, formerly named Max/MSP, are and also how it works with
Ableton Live itself.
In the second part, we will look at something already known by some of you: the famous (but
cryptic) Live Object Model, also named LOM within Max for Live’s Coding Mafia. It is the inner
structure of Live itself. We can reach, request and manipulate Live’s parameters thru it by using two
ways: Max for Live itself, but also the (also famous) MIDI Remote Scripts. We will explore Live’s
Application Programming Interface (also named Live API) and how to program it.
In the third part, I will talk about JavaScript as another way to reach and work with Live’s API.
You’ll like this part because it will push you into another field of programming within Max for Live.
In the fourth and final part, I will show you how we can easily design and build Max for Live
devices like MIDI instruments, sound generators, MIDI Fx and Audio Fx and also how we can play
and control Live using API from within the devices.
You can also follow me on twitter⁵ & facebook⁶
Are you ready?
⁵https://twitter.com/julienbayle
⁶https://www.facebook.com/julien.bayle
About
This guide is published & provided by the leanpub.com web platform but all rights are reserved to
Julien Bayle under © Julien Bayle 2013.
Copying or publication of any part of this guide is not allowed, at the moment
Being published on leanpub.com⁷, this guide can easily be revised and updated. You
will **automatically be notified about any updates in the future.
Stay Tuned.
⁷https://leanpub.com/Max-for-Live-Ultimate-Zen-Guide
About the review & edit of the book
This book has been entirely reviewed and edited by Mark Towers and I want especially thank him
for such involvement, serious, expertise and kindness !
Mark Towers teaching Ableton Live in Leicester
Mark Towers is an Ableton Certified Trainer from the UK, he runs a Foundation Degree in
Creative Sound Technology at Leicester College/De Montfort University, and teaches Ableton,
Sound Design and Programming with Max 6 and Max for Live. Mark also specialises in Electronic
Music Composition & Performance.
You can visit his website at marktowers.net⁸
⁸http://marktowers.net
FAQ
Version
The guide version is 1.0
This guide is not

• a whole course about Max6
• a whole course about Ableton Live
• a whole course (nothing can beat a course thru skype or face-to-face: ask my students!)
This guide is
• a guideline with the Max for Live world
• the guide you need to completely understand what Max for Live is and how it works with
Ableton Live itself.
• THE tutorial which will drive you to create your own Max for Live devices perfectly fitting
inside Ableton Live (presets’ storage, parameters’ automation, Live API etc)
• THE ONLY tutorial published & written by an Ableton Certified Trainer about Ableton Live
If you find a mistake or a typo in this guide

Please write to me using this page: http://julienbayle.net/contact⁹
If you would like to find out about courses on Ableton

Live, Max6
Please write to me using this page: http://julienbayle.net/contact¹⁰
⁹http://julienbayle.net/contact
¹⁰http://julienbayle.net/contact
1 Max for Live Basics
Max for Live isn’t a concept or a magical spell. There are lots of video tutorials out there that don’t
really explain it, so I will do that and I’m going to explain precisely what it is.
1.1 What ?
My own Max for Live definition is:
Max for Live is an interactive visual programming environment directly usable within
Ableton Live for designing devices.
We are going to review Ableton Live & we will revise it in the following pages, but at the moment,
let’s begin by explaining what interactive visual programming environment means.
What is an interactive visual programming environment ?

There are a lots of programming languages. Each one of them have their own characteristics and
paradigms which are basically the rules they are following and we have to follow in order to use
them.
Here is, for instance, an example of something I coded in C++ language for a generative installation¹:
¹Generative Digital installation named The Village / Where2Now:http://julienbayle.net/works/collaboration/the-village
Max for Live Basics 2
C++ language is based on text coding
This type of programming language requires a specific proramming workflow:
• we write the code

• we include some external programming libraries
• we compile
• we test
This series of steps is done cyclically: we modify, we compile, we modify again etc. There is
something I like to call a design disruption that comes from this necessary compilation step which
requires to stop coding and to compile & test before we begin to code again.
An interactive visual programming environment is interesting because it doesn’t have any design
disruption. As soon as we modify the program, we are seeing directly the consequences of our
modification
A visual environment also provides a way to learn programming more easily and less abstractly.
Indeed, we don’t have to create complex abstract structures for which we would have to remember
each part & detail but we are moving objects and linking them together.
Here is an example of what I mean with a picture:
A Javascript text code and its visual equivalent
On the left, there is a piece of JavaScript code. We are declaring a function named add_five which
requires an argument. Then, this function is called with the number 2 as an argument and the result
is stored inside the variable result which is displayed using the function display() that is not visible
defined and declared here. display() is only here as an example showing that we have to call or
use functions to display things in usual text-based programming language because the UI is often
separated from the code itself (UI means User Interface).
On the right, the same thing but designed with Max6. There are two UI objects providing both
features of displaying and modifying quantities. The object [+ 5] is the equivalent of the function
add_five and adds 5 to any values incoming from the top. Then, [+ 5] emits the calculation result
to the bottom.
BUT
I don’t mean that one type is better than the other.
Neither one has less features than the other.
Interactive visual programming is just simpler to learn and less abstract than the other,
especially for people with no programming background.
Max6 & Max for Live

Some of you will have already heard about Max6, formerly named Max/MSP.
Max6 is the latest version of what long-time users still call Max/MSP.
In the following pages, I will use the term Max to denote Max6 in order to mean Max6 environment.
But I will use Max for Live too, because it is very different from Max itself as we are going to find
out.
Max vs Max for Live
I’m going to explain why Max for Live is not equivalent to Max6.
Max framework & Max runtime Max is a development environment available here: http://cycling74.com/shop²
Max is an autonomous framework that doesn’t require you to own an Ableton Live license, and
even if you had one, it wouldn’t be necessary to launch Live to use Max.
When we program with Max, we create patches that we can edit & modify. A patch is a set of
objects more or less linked together.
If we want to share a patch with another user who doesn’t own a Max’license himself, we can create
an autonomous Max application. This application is basically our patch, a set of dependencies (all
objects required by the patch to be loaded) and the Max’ runtime.
runtime is available for free here: http://cycling74.com/downloads/runtime³ in case we want to
launch and use a collective. A collective is just the patch with all dependencies, but without the
runtime.
Check the next figure, it illustrates how this works:
²http://cycling74.com/shop
³http://cycling74.com/downloads/runtime
Max, Max Runtime, Patch, Collective and Application
Ableton Live & Max for Live If you understand what the previous figure means, you won’t have
any problem to understand the next pages.
Max for Live doesn’t provide any runtime.

More precisely, the only software able to open, edit & run Max for Live patches is Ableton
Live.
Actually, we refer to these patches as Max for Live Devices instead of just saying Max for Live
patches because these latter sit within Max for Live devices.
These devices are stored as files containing patches required by them to work, and also all
dependencies required too, especially in the case of those devices that have been frozen. We will
take a closer look at this later in this book.
For now, we are going to explore another figure in order to understand clearly what these different
entities are:
• liveset (file with the extension .als storing all liveset’s data like clips’ parameters, devices’
parameters, MIDI mapping and much more)
• liveset project (folder containing one or more livesets, samples related to these livesets and
much more)
• device (file with the extension .amxd containing all elements required by the patch to be
loaded and used as a device in Live)
• device in edit mode inside a … Max-style editing window
Liveset, liveset project, device and device in edit mode
You could think of Max for Live’s runtime as being Ableton Live itself and Max for Live’s editor as
being Max run by Live.
One of the first direct consequence is :
as soon as we load a Max for Live device, we can also edit it !
You can also click on the edit button of a Max for Live device and you’ll see the underlying Max
patcher open and provide you with the whole patch related to this device.
Max for Live’s device EDIT button
Now we are a bit more aware of how Max for Live, and Max & Ableton Live are related, let’s
continue with licensing.
Use & licensing
In order to use Max for Live, we have to own an Ableton Live license, obviously. But not only, of
course.
I only own Max for Live Then, I can use and edit Max for Live’s devices.
But I cannot use Max outside of Ableton Live, I mean: as the pure Max framework.
I only own Max I CANNOT use Max for Live’s devices, nor edit them, of course.
Obviously, I can use Max independently outside of Live or even connected to Live using MIDI or
audio by using audio inter-application routing application like Soundflower⁴ on OSX or Jack⁵ on
OSX or Windows.
I own Max & Max for Live I can use Max for Live’s devices.
I can also use Max patches independently of Live and make them communicate with Max for Live’s
devices using OSC, for instance.
⁴Soundflower: http://cycling74.com/products/soundflower
⁵JACK: http://jackaudio.org
1.2 Ableton Live Basics

Ableton Live is an audio & MIDI workstation. We call this kind of software a DAW (standing for
Digital Audio Workstation).
Let’s explore what Ableton Live is a litte more.
Projects & livesets

A project is a folder inside of which we can store livesets.
It can contain multiple livesets and it is a convenient way to organize our music.
We can find imported audio sample files and also samples created within the liveset itself. When we
use the feature Collect All & Save, all presets and elements required by the liveset are copied inside
this project folder too in order to make it totally independent from any external files.
I’d suggest reading the Live manual if you want more details about these fundamentals.
Tracks
Live’s livesets are structured around the tracks’ concept, whatever the mode we use (session or
arrangement modes)
A track can be an audio or a MIDI track.
Each track can contain a set of clips which are the most basic unit inside of which we can have MIDI
sequences or audio samples, depending on the type of track. MIDI tracks can contain MIDI clips and
audio tracks can contain audio clips.
In a track, only one clip can be playing at a time.
Each track also contains a mixer and a send FX section, and of course a devices’ chain that can be
seen at the bottom of the screen in both modes.
An audio track can contain:
• audio clips inside the clips area (clipslots in session mode or timeline in arrangement mode)
• audio effects inside the devices’ chain
A MIDI track can contain:
• MIDI clips in the clips area (clipslots in session mode or timeline in arrangement mode)
• MIDI effects inside the devices’ chain, before the MIDI instrument
• MIDI instrument inside the devices’ chain
• audio effects inside the devices’ chain, after the MIDI instrument
Devices
We have refered to devices since the beginning of this book but we didn’t define them…
Devices are those entities that we can drag from Live’s browser and drop on to the tracks’ device
chain.
We can see in the browser (on the left) a big list of categories. Version 9 of Live can be a bit confusing
at first sight.
Live’s Browser
We find the different devices within each of the following categories:
• Instruments
• Audio Effects
• MIDI Effects
• Max for Live
• Plug-ins
Category names are quite self-explanatory.

In the Instruments category , we can find Live’s native instrument devices like Operator, Analog
or Sampler. They receive MIDI messages from Live itself through MIDI clips or from outside (from
another application within the same computer or from another computer or MIDI controller) and
can produce audio depending on the MIDI messages received.
In the Audio Effects category: there are Live’s native audio effects like delays, compressors or beat-
repeat. They can process audio signals.
In the MIDI Effects category: there are Live’s native MIDI effects like Pitch, Chord and some others.
They can transform MIDI messages.
The Plug-ins category contains all non-native devices like VST for OSX and Windows or Audio-
Units (AU) for OSX plugins. They are add-on devices usable in Live but that are not delivered with
Live or even designed by Ableton.
Then, we have the Max for Live category.
There are 3 types of devices in this category:
• Max MIDI Effect

• Max Instrument
• Max Audio Effect
We have the same classification in 3 categories.
Max for Live’s 3 different types
Why?
Because any device that we want to use has to be placed in a device’s chain in Live. And, as we have
just seen, a device’s chain belongs to a track and a track can be either a MIDI or an audio track with
all consequences raised.
When I’m teaching Ableton Live, I use the following image to illustrate the concept of MIDI effects,
MIDI Instruments, and Audio Effects, look at this now:
MIDI & Audio in Live’s devices chain
Bar graphs represent MIDI signals and VU-meters represent audio signals.
If we start from the left that is the entry point of a signal inside the devices’ chain (the signal can
come from inside Live itself or outside, as I said), we can see a Bar graph.
Because this device is included in-between two Bar graphs, that means we have a MIDI Effect here.
It takes an incoming MIDI signal and outputs a MIDI signal: this is the basic definition of a MIDI
Effect.
With the same reasoning, we can see that the next device is in-between a Bar graph on its left and
a VU-Meter on its right. It receives a MIDI signal and produces an audio signal: this is a MIDI
Instrument.
Then, the last device is in-between two VU-Meters. It is an audio Effect.
In the same way we are going to discover that Max for Live’s devices are nothing but Live’s devices,
with specific features like edit mode of course, and this is why they have to be from one type or
another such as MIDI effect or audio effect etc.
We are also going to see that in the case of a Max for Live device that doesn’t have to
process MIDI or audio but to be an interface between Live and a Wiimote controller for
instance, we can use any of the 3 types. Indeed, this won’t be important except for the place
where we will want to place it in Live’s device chain.
Let’s continue with Live’s basics.

MIDI Mapping
MIDI Mapping provides a way to associate MIDI messages like Note or Control Change with a
livesets’ parameters through Live’s GUI (GUI stands for Graphical User Interface)
MIDI Mapping Edit Mode enabled
We can enable the MIDI learn mode by clicking on the MIDI text button at the top right in Live’s
GUI (or by pressing CMD + M on OSX & CTRL + M on Windows)
Once enabled, we can select either a parameter, elements like clipslots, but also scene selectors, we
can then move a knob on our hardware controller. Then the mapping/association is made: the liveset
parameter is linked to the hardware controller’s knob.
PAY ATTENTION HERE

The liveset’s MIDI mapping is stored inside the liveset itself and nowhere else !
Question: what can we do if we want a particular set of parameters to be mapped each time to the
same set of knobs, whatever the liveset used ?

Indeed, because we may be playing our live performances quite often, we want to be able to retrieve
our setup when we are working on a liveset, on stage etc. without having to remap everything
everytime.
Of course, we can keep the same liveset and copy-paste it as a template. As we just saw, MIDI
Mapping parameters will be copied with the liveset’s file itself. It works fine. A lot of people do that.
But we can do better by using MIDI Remote Scripts.
By the way, be a little bit more patient here: Max for Live is coming. But we have to keep all of that
in mind before we dive in completely !
MIDI Remote Scripts

Some of you probably already know about MIDI Remote Scripts.
We are talking about scripts for programming languages that don’t require a proper compilation
step.
When we are typing JavaScript code, for instance, we don’t need to compile it before to test it: an
interpreter can run it directly. JavaScript is often used within web pages with HTML and in that
case, the interpreter is just our web browser itself.
What are those scripts ?
MIDI Remote Scripts are Python-based scripts. Python is a programming language. They are
directly interpreted (and run) by Ableton Live itself and provide ways for:
• controlling Live (act on Live)

• grab feedback from Live (observe Live)
By checking the MIDI Sync tab in Live’s preferences, we can see in the top level part some
parameters related to Control Surfaces (sometimes referenced as Remote Surfaces)
Enabling & Disabling MIDI Remote Scripts in Live
Each member of the list corresponds to a MIDI Remote Script available from Live’s installation
folders.
Each one is named after a controller. We can see APC20, APC40, Push and also LividCode, for
instance. The latter involves the Livid Instruments’ Code controller on which I am currently working
on (a secret project, by the way)
Generally, these scripts are designed by the hardware’s designers themselves for hardware handling
in Live.
For instance, APC40’s script provides a way to directly control the clip matrix using its buttons and
also to grab all clips’ states.
Basically, these scripts are an interface between these controllers and Live itself.
Here is a minimalistic schematic summarizing these concepts.
Livid Code controlling Live through MIDI Remote scripts OR classic MIDI mapping
Black arrows show the relationships between MIDI Mapping and MIDI Remote Scripts concepts
AND Live’s Preferences window.
In blue, we can see the control/feedback themselves.
In the MIDI Remote Scripts method, we can see that the configuration is depending on Live itself,
but in the MIDI Mapping method, the configuration is based on the liveset itself. Of course, in this
latter case, the MIDI interface configuration is depending on Live itself too, but please keep in mind
that MIDI Mapping IS stored inside the liveset’s file (the .als file)
Where can we find the MIDI Remote Scripts on our computer?
These scripts are within Live’s folder hierarchy.

On windows:
• C:\ProgramData\Ableton\Live x\Resources\MIDI Remote Scripts\ on Windows 7 & Vista

• C:\Documents and Settings\All Users\Application Data\Ableton\Live x\Resources\MIDI Re-
mote Scripts\ on Windows XP.
On OSX, we have to find the file Live.app in Macintosh HD/Applications/Live x.x.x./ and right-
click on it, then choose Show Package Contents. At last, if we look inside this folder: Contents/App-
Resources/MIDI Remote Scripts, we can see this:
MIDI Remote Scripts on OSX
Wouldn’t it look like the previous Remote Surfaces list we saw in Live’s preferences ? Of course, it
is directly related.
Hack and Script’ edit
In the previous snapshot, we can see the opened folder LividCode and some files inside it with the
extensions: .py and .pyc.
.py & .pyc file format The file extension .py is related to Python scripts’ source codes. We can
easily read and modify them. These are basic text files.
Actually, if we have a better look at the content of the folder, we can also see .pyc files. Those aren’t
editable. They are binary files.
These are bytecodes⁶. They can be considered as executables, binary files are directly executable by
the Python interpreter, which is Live itself, as we already discovered.
If we modify a Python script source code, the next time we launch Live, and choose the correspond-
ing Remote Surface in Live’s Preferences, Live will take the .py source code and will pre-interprete
it (this is a kind of pre-compilation) in order to optimize the execution and will create this .pyc
bytecode file (one for each .py used)
Retrieve source codes from bytecodes Obviously, as we are more or less hackers, we want to
read what is in these source code .py files. Indeed, we first need to understand how the Live API
works and how we can design our own MIDI Remote Scripts.
As an early 2013 beta-tester for Ableton’s PUSH controller, I especially wanted to read the source
codes for this controller and I decompiled bytecodes early February 2013.
I published the whole sources and I’m doing my best to maintain and update them as soon as a new
version of Live is released.
Everything is available here on Github:
https://github.com/gluon/AbletonLive9_RemoteScripts⁷
I know this helped and still helps a lot of you and especially application developers like those who
made Conductr⁸ for iPad.
Because nothing is officially documented, I tried to make some documentations:
• Push & hacks⁹

• MIDI Remote Scripts¹⁰
• Doc for python Live object¹¹
⁶http://en.wikipedia.org/wiki/Bytecode
⁷https://github.com/gluon/AbletonLive9_RemoteScripts
⁸Conductr for iOS: http://www.conductr.net
⁹http://julienbayle.net/ableton-push
¹⁰http://julienbayle.net/ableton-live-9-midi-remote-scripts
¹¹http://julienbayle.net/ableton-live-9-midi-remote-scripts/#liveAPI
• Docs for _Framework & _Tools objects¹²
I won’t teach you how to program using Python language here, but how to do the same things by
using Max for Live itself.
Let’s dive now inside Max for Live.
1.3 Max for Live specific objects

Max for Live contains a specific set of objects available while patching. Some of the objects are only
usable within Max for Live, but others can be used in Max too, outside of Live, I mean.
Here is a global picture with specific objects of Max for Live.
Max for Live specific objects
We can see two series of objects:

¹²http://julienbayle.net/Ableton-Live-Scripts-Documentation
• at the top, objects able to provide control/feedback from Live itself,

• at the bottom, Graphical User Interface’s objects (GUI)
I will describe objects at the top in the next chapter.

live.arrows, live.button, live.text & live.tab are button type objects. We push, we select.
live.dial is basically a potentiometer, a knob.
live.number displays values but also provides a way to modify them.
live.menu presents a list to the user who can select an element in it.
live.drop makes the connection between a Max for Live device and the file system of the computer.
We can drag’n’drop files on to this area and this object pops out the absolute path of those files. It is
really interesting if we want to process files and to load them to memory [buffer∼] objects in Max
for Live for instance. There is no need to type a path, it finds it for you.
live.gain∼ displays the audio signal level and provides a way to change the gain too.
live.meter∼ only displays the audio signal level.
live.slider is a basic fader.
At last, live.grid & live.step are helpers and GUI to configure other objects.
live.grid works well with chucker∼ by providing a display and control to cut an incoming signal
into pieces and to replay those pieces in another order.
live.step is a step sequencer.
We will talk about these objects in the next pages of the book because it is very important to know
how to use each of them.
Because Live brings some features I want you to master, for instance, if you want to create a Max
for Live device and be able to automate some of its parameters in Live, or to save them with the
liveset, you have to apply specific actions.
Now, let’s keep focused and explore the famous Live Object Model (LOM)
2 Live Object Model
This part of the book is very important and I’m going to try to explain and describe all of the
following concepts as simply as possible in order to help you understand all that you’ll need to build
your own interfaces using Max for Live.
By the term interfaces, I mean devices that can interact with Live itself by requesting data from
it (like grabbing the tempo in real-time and sending it as an OSC-transported value to another
program for instance) but also by controlling Live itself (change the value of a parameter, change
the transport state, etc)
At first, please keep in mind and in sight the Max window. We can display it in Max (and Max for
Live) by typing:
• CMD + M on OSX
• CTRL + M on Windows
We can easily display data in the Max window like debugging information, but also log some data
flows. It is a useful way to visualise data flows and learn how things work.
We can also display the Max window directly from Max for Live’s devices title bar. By right-clicking
on them and choosing the last option at the bottom Open Max Window, we can show it.
Opening the Max window from the device
2.1 What is LOM ?

The Live Object Model represents Live’s internal structure. I mean, in the program, in the software
itself. Actually, it would be easier to say: it is a map that guides us to everything accessible within
Live Object Model 21
the Live application through programming.

At first sight, we can see some elements which we already know about like tracks, clips, devices
Check this big schematic provided in the Max documentation in the specific Max for Live part:
Live Object Model as available in the documentation by Ableton and Cycling’74
It looks cryptic and complex. It is a bit complex, but not that much!
We can also find this big picture on-line here¹
If you want to work with this schematic under your eyes, you can easily reach it via the Max’ help:
¹http://www.cycling74.com/docs/max6/dynamic/c74_docs.html#live_object_model
Reach the Live Object Model in the documentation in 3 clicks only
As well as being a big schematic overview, the Live Object Model also includes text on the same
documentation page related to the schematic.
When we click over the different rounded corner rectangles in the schematic, we go directly to
the text section corresponding to the rectangle.
Those rectangles are what we call objects (the O of LOM)
2.2 LOM’s Objects

Objects are an elementary entity required by anyone who wants to use Max for Live and the LOM.
There are 19 objects available:
• Application & Application.view

• Song & Song.view
• Track & Track.view
• Clipslots
• Clip & Clip.view
• Device & Device.view

• DrumPad
• Chain
• ChainMixerDevice
• DeviceParameter
• MixerDevice
• Scene
• CuePoint
• ControlSurface
These objects provide access to specific elements of Live, visible or not, as we are going to find out.
Each object has its own characteristics & properties. For instance, tempo is a property of the Song
object. The looped or not looped state of a clip is defined as a property of the Clip object.
Application & Application.view

These objects are related to the Live application itself. For instance, if we want to display a user
interface in a Max for Live device depending on the version of Live, we can request this property to
the object Application.
The notation & suffix .view only means that the object is a little bit more related to elements of the
graphical user interface of Live itself (and not exactly the liveset itself). “.view” suffixes also appear
for other objects than Application.
Song & Song.view

Song is related to the liveset itself.
Track & Track.view

Track is related to the tracks themselves. Actually, it is related to one track only, each track being
addressed by multiple instances of the object Track.
Clipslot
Clipslot is related to slots inside which we can put clips into, in session view.
Clip & Clip.view

These objects are related to clips themselves and how they can be displayed.
Device & Device.view

These objects are related to devices themselves and how we can display them.
DrumPad
DrumPad is related to the drumpad of drumracks. It is quite specific, but very useful.
Chain & ChainMixerDevice

These ones are about Racks and especially Racks’ chains.
DeviceParameter
DeviceParameter Devices’s parameters are all the parameters (dials, faders, buttons) available in
Max for Live’s UI devices.
MixerDevice
This one defines a track’s mixer.
Scene
This one defines scenes. A scene is a row in the clips’ grid in session mode.
CuePoint
This one defines arrangement mode’s markers.
ControlSurface
ControlSurface is an interesting object providing a way to access MIDI controllers through MIDI
Remote Scripts but from Max for Live. I especially worked on that for Livid Instruments and their
Code Controller. You can check this page² for more information.
²http://julienbayle.net/livid-code
2.3 How to read the documentation

Each object is described and broken down in the same way.
Let’s inspect the object Scene as an example:
Scene object’s documentation
What does this all mean? Lets take another look:

Scene object’s documentation explained
We can always see the shaded elements, even if there is no info/object listed.
2.4 Children, Properties & functions
Parent & Children objects’ relationships

Children object concept is fairly straight forward: some objects have some children objects.
For instance, a liveset has tracks. Track instances are the children of the object Song (i.e. liveset).
There is a hierarchical relationship between tracks & a song within the Live model.
Another example: tracks own clipslots and these latter can contain clips or not.
We can ask an object for a list of its children.
It is really important to grasp this concept in order to navigate inside the LOM.
Properties
Objects’ properties are very varied and are intimately related to the object itself.
For instance, the object Scene owns a property defining its color and another one defining its name.
In the property part of the documentation, we can see something related to the accessibility of this
property
There are 3 types of access to properties:
• get
• set
• observe
get is the way to request the current value of a property when we request it. This is a one-shot
access.
set means we can change the value of the property.
observe means we can observe the property. It allows us to constantly monitor a value. By using
this feature, as soon as a property’s value changes, we are informed of its new value.
We are shortly going to learn how to use these 3 ways of accessing properties.
But before, we have to learn how to navigate within the LOM.
Functions
A function is an action launched through a considered object.
For instance, the function fire launched for a clip will trigger it with all consequences resulting. For
instance, if the clip’s launch mode is toggle and the clip is already playing, calling the fire function
will stop it exactly as if we were to click on the small triangular button on the clip itself in Live’s
GUI.
2.5 Path, ID & navigation

In order for us to access and apply actions to LOM’s objects such as property value modifications or
function calls, we need to find the instance of the object that we want to work with.
Object / Instance & ID

An object is an abstract definition of a more concrete element of Live in the LOM.
An object instance is an effective materialization of a LOM object.
For instance, a car could be an object. It defines the global concept of what a car is. And that specific
green car near my home is an instance of a car. It is a concrete and special materialization of the car
concept, of the car object.
When we are playing with the LOM, we need to reach precise object instances.
Within Max for Live, the way to reach object instances can be done by using unique IDs which
describe each instance of each object, said more simply: In Max for Live, we can reach objects’
instances by using their internal IDs
These IDs are created and set as soon as we start to reach the corresponding instances. I mean,
nothing is defined by default before this action.
For instance, if I’m trying to reach the Master’s volume by requesting it through the LOM, the
system is creating an ID, a way to reach it. For the whole Live session, this ID is used for that precise
LOM object. It also can persist through sessions (i.e. between two different launches of the Live
application)
How can I find the ID of the object I want to reach ?

We have just learned that we cannot predict IDs values before the first request to each LOM objects.
So we need a tool for this and it is the Max for Live object named [live.path].
[live.path]
live.path translates an objects canonical, absolute or relative path into a corresponding ID
for the object
We can follow another analogy here: on the Internet, we can reach Google’s website by typing
google.com directly into the address bar. But Internet communication protocols use IP address to
make messages travel and indeed require IP addresses. Domain Name Servers (DNS) are here for
that purposes: they convert names into IP address.
[live.path] is like the DNS for IDs in Max for Live.
Follow object or path ?

[live.path] provides three outputs.
live.path’s three outputs
Help documentation tells us that:
• the first one provides IDs in follow object mode

• the second one provides IDs in follow path mode
• the last one is the dumpout
If the last one is the generic dumpout providing a way for objects to dump information while they
receive particular messages, the role of the first two is a little bit more difficult to understand.
Let’s imagine we are accessing a path like live_set tracks 2 and we take the ID from the follow
object output.
The ID is the one for the track which has the index 2 (it means the third track, as all indexes start
from 0)
If I keep that ID and I send it to a live_set tracks 2 object, this object will begin to handle the third
track.
Ok.
But now, imagine that we move this third track to another place in the liveset (at the end of my
tracks for instance)
The same ID we already have is STILL related to the freshly moved track (we chose follow object
mode)
But imagine now another scenario. If the ID comes from the output follow path and I move the third
track to the beginning of my tracks in the Liveset. We are okay to say the path itself has changed,
right? Indeed, the track originally had the index 2 (as it was the third track) but because I moved it
to the beginning of my tracks in Live, it now has the index 0. So if I play with the same ID I stored
for instance, I won’t play with the track I just moved, but with the new track that is in the position
of track 3 in the Liveset, which now has the index 2 (it previously had the index 1, but has now
moved up one as a result of my track rearranging)
So, be careful with which output you take the ID from.
2.6 One shot request with get and live.object

I like to experiment fast in Live using a basic setup with Max for Live. This one allows me to rapidly
prototype things and to have it all in the same folder. Let’s do that first.
Creating a Max for Live rapid testing setup
• Open Live, create an empty liveset and save it with a name related to your test.
• drag’n’ drop a Max for Live device from the MIDI Effect type in a track.
• Click on the edit button of this device
• Save this device in the same folder as your Live project
This is a basic guideline that you can follow to quickly create a rapid prototyping environment.
Now, edit your device and create a patch as in the figure below:
LOM’s testing
This structure provides a simple way to test and understand IDs, paths, and to make one-shot
requests to Live by using live.object.
There are two ways of using [live.path]:
• setting the path as an argument. Arguments are assigned to objects by putting a space after
their name, then the arguments itself.
• sending a message containing the keyword path followed by the path itself.
Both are syntactically equivalent, but the second is more interesting when we require dynamic
systems. For instance, if we want the patch to be able to make requests about multiple tracks, we
will need to build paths dynamically depending on the track we need information from. We cannot
do that when the index of the track is set as an argument.
Here, from our small Max for Live device, we want Live to tell us what the tempo is and if the
transport is playing or not.
If we check the LOM documentation, we can see that tempo and playing status are available as
properties of the object Song. Indeed, these values depend on the liveset itself.
I made a small cut to show specifically the data we need to see within one table:
We need the object Song here
All that we need are these two properties :
• tempo
• is_playing
tempo data type is float, which basically means it is a number with a decimal point i.e 120.50. We
see that we can also perform a get, a set and even an observe request on this property.
is_playing data type is a bool that stands for a boolean. Its value can be true or false. We can also
perform a get, a set and an observe request on this property.
Now, we know which properties we can request and to which object we need to perform those
requests.
Let’s check the canonical path of the object Song: live_set
Let’s pass it to [live.path] and this one will produce a result formatted like this: ID <number> where
<number> is the numerical ID of the requested property.
BE CAREFUL OF ID 0
If ID 0 is the result of [live.path]
Then it means that no object fits with the path passed to the [live.path]
ID 0 is often a clue saying you made an error in the path.

Object [trigger], abbreviated as [t], provides a way to control the outgoing flow of messages. Here,
[t b l] means that if the trigger object receives something, it will process it as a list that will be
transmitted from the rightmost outlet then after this operation, it will send out a bang from its
leftmost outlet.
The message element linked to its rightmost outlet is only here for control and display purposes, it
shows us what ID is sent out of [live.path]. This value is transmitted to the [live.object] which will
be linked to the object’s instance related to the transmitted ID, here: the Song object.
From this moment, we can transmit multiple requests to [live.object] and it will process them
accordingly to the LOM, relatively to the object Song, that means the liveset itself.
Then we send two messages to it:
• get tempo
• get is_playing
And [live.object] answers with the two responses :
• tempo 120.00
• is_playing 0
[route] object provides a way to route incoming messages to many outlets depending on a prefix
(tag) in the received messages. It also un-tags these incoming messages (here it removes the tempo
and is_playing prefixes)
Play a bit with this patch.
Launch some requests, observe the results, then change the tempo and activate the transport in Live.
It works fine but we need to relaunch the requests to get the updated values.
Requesting values occasionally

1/ find the property you want to know the value of
2/ find the object providing this property and remember the object’s canonical path (the
canonical path can be found under the object name in the LOM documentation)
3/ give this canonical path to live.object
4/ request the live.object with messages like : get <property name> and this one will
answer with messages like <property name> <current value of the property>
This is good! But we have to make a request every time we change the value and if we change the
value we know the new one anyway, so why would we have to request it… ?
It doesn’t make sense because what we need is the updated tempo value at anytime.
2.7 Observing properties in Live with live.observer

As mentioned previously, [live.observer] provides a way to be informed about a property’s value
change whilst the property is observed by the [live.observer] itself.
This means, we can give an ID to [live.observer], we tell it what property we need it to follow, and
it does the job.
Here is the corresponding example patch:
Observe tempo & transport’s state
I have consciously left the previous patch to the left with the get / [live.object] technique.
We can see that the ID itself is the same in both cases. This is totally normal (and safe) as we can
provide the same path which navigates to the Song object to both [live.path] objects.
Here, we have the [t b l] object again, and it transmits the ID to the [live.observer].
Please, be aware of this:
[live.observer] can ONLY observe one property at a time !
If we need to observe two properties, we need two [live.observer] objects !

Then, to define which property the [live.observer] should follow, we send it this message:
property <property name>

From this moment, [live.observer] transmits from its outlet the value of the property. If this latter
changes, a new value is sent out.
Play with it as you did with the previous patch.
Change the tempo, start & stop the transport while looking at the patch. You will see the
[live.observer] objects update with the same correct values.
Observing a property
1/ find the property you want to know the value of
2/ find the object providing this property and remember the object’s canonical path
3/ pass the canonical path to [live.path] which finds the ID for [live.observer]
4/ send the message property <property name> to the [live.observer] which will send
out the current value of the property.
This is the way to always be informed with up to date properties’ values.
2.8 Calling objects’ functions with live.object

We are going to learn now how to call functions.
When we read LOM’s documentation, we can see that each object owns a set of functions.
We are going to create a small patch that provides a way to trigger clips.
Seeking the object required

Let’s scroll through the LOM documentation. We can see an interesting object, the ClipSlot.
Indeed, without worrying about whether there is a clip or not inside the clip slot, we can use the
function fire and this function will, as read in the LOM, trigger the clip if there is one, or trigger the
stop button of the clipslot if there is not.
This is exactly what we want to.
The canonical path to reach each clip slot in our clips’ grid is:
live_set tracks N clip_slots M
with N being a track’s index & M as scene’s index (N and M are the coordinates of the clip slot,
starting from zero in each dimension)
Building the patch

Here is a possible example patch.
Triggering clip slots
Here the canonical path has been built dynamically.

This means that when we push each button, the static message placed under the buttons will send
out the following message:
path live_set tracks $1 clip_slots $2
$1 & $2 are what we use to call placeholders. These are variables (known as replaceable arguments)
that get their values from incoming messages, in the same order of the incoming list.
I chose to put the index of the track first, then, the index of the scene (our clip slots coordinates)
For instance, if I send the message (2 1) to my message path, this will transmit path live_set tracks
2 clip_slots 1 to [live.path]
And [t b l], after having transmitted the ID received from the [live.path], will trigger a bang sent
to the message (call fire) and then, as soon as this message goes into the [live.object], the function
fire will be called and the corresponding action triggered.
Here, the clip slot is triggered exactly as if we were to push its stop button or if there was a clip the
play/stop button of the clip.
Put some clips in the liveset between scene 0 and 1 and track 0 to 3 and play with the patch.
Calling objects’ functions

1/ find the object and its function you want to call
2/ pass the canonical path of the object to [live.path] which finds the ID for [live.object]
3/ send the message call <function name> to [live.object]
Functions act on the liveset and thus modify, as a consequence, other properties.
For instance, triggering a clip with the transport stopped will also start the transport. The value of
the is_playing property of the object Song will change to 1.
But we can also change some properties’ values without using functions.
2.9 Changing properties’ values by using set &

live.object
We are going to create a button to start and stop the transport by changing the value of the property
is_playing.
Don’t forget to check what you can and cannot do with each property you want to use!
Indeed, is_playing allows us to get, set and observe and this is also why I’m often using it during
my training sessions in order to illustrate how we can use all of a property’s features, globally.
Look at that patch:
Apply a set on a Live object
It looks strangely like the first example we built with get.

Here, in place of get is_playing, we have:
set is_playing $1
This is a basic way to change a property’s value.
[toggle] provides an easy way to change a value from 0 to 1 and from 1 to 0 with a nice and cheap
user interface: a check button.
Push it several times, it will trigger Live’s transport.
Changing a property’s value

1/ find the property you want to modify
2/ find the related object and its canonical path
3/ send this canonical path to [live.path] which will find the ID for live.object
4/ send the message set <property name> <new value> to [live.object] and the property’s
value will be modified.
We can do this for continuous parameters too, I mean, a parameter that takes a range of values and
not only on and off, or 0 and 1.
2.10 Controlling continuous parameters in Live with

live.remote∼
Imagine that you need to control the value of a track’s mixer panning of track 3 and you want to
make it follow a continuous sine wave, for instance.
We can totally do that with [live.object].
But I must warn you about this, and would not encourage you to do it!
Why ?
Because when we do this with the set / [live.object] technique, all modifications are recorded in
Live’s UNDO history
By using set / [live.object] technique with continuous parameters, we would pollute very quickly
the UNDO history.
BE AND WATCH CAREFULLY

You will find a lot of Max for Live devices programmed in this way.
Be careful, not all Max for Live device designers are very aware about this…
More, and this is apparently a consequence of this UNDO history fact, system performances aren’t
good in this case.
This is also why we have another tool available:
[live.remote∼]
As we can see, there is the famous tilde ∼ at the end of the name.
It means that this object is able to process signals generated by the MSP part of Max for Live, or at
least entering into MSP.
Look at the next patch
Home-made LFO controlling a track’s panning
Easy, isn’t it?

But let’s look at it closely, especially because I’m using a specific path.
Look:
live_set tracks 3 mixer_device panning
And at this very moment, I have usually 4 or 5 people asking me the question:
“Hey! panning isn’t an object! Because we want to modify it, this is a property and a property doesn’t
own an ID so live.path won’t take it…”
“Very good question”, I used to say: but look at the LOM a bit better.
DeviceParameter & MixerDevice, two singular objects
Panning is a child of MixerDevice. And when we are checking its type, we can see that it is a
DeviceParameter. Thus this IS an object.
If we are checking objects which are DeviceParameter, they don’t have a lot of properties. We can
see the property value, that is the current value of the considered parameter and min/max which
are the minimum and maximum values allowed for this parameter in Live.
Access column in the table Children

BE CAREFUL HERE (too)!
On the previous snapshot, we can only see a ‘get’ in the MixerDevice’s children table in
front of panning.
It means that we can only get a list of the MixerDevice’s panning children.
It doesn’t mean that we can only use get for the underlying DeviceParameter object.
Said differently: In children tables, get, set, observe are only related to the relationship
parent/children of the concerned object. In order to know what we can do with child objects,
we need to look at the type of child object itself.
[translate] just translated a note value (this is the name of the relative time format in Max,
sometimes) to Hz value in order to set the sine wave frequency to modulate the panning of the
track. This allows, for instance, to have one sine wave cycle per bar, or for 1/2 bar etc.
In Max for Live, the default transport is Live’s transport.
Important notice:
live.remote∼ holds the parameter and locks it restricting any manual modification while
it controls it.
Indeed, in Live, the parameter linked to the live.remote∼ is colored in grey.

It means we cannot automate it or modulate it.
In order to break the link between [live.observer], [live.object] & [live.remote∼] and
an object, we have to send them a message (ID 0).
If we send another message (a real other ID) to [live.remote∼] for instance, and if this ID
was already associated with another [live.remote∼], we make another association and the first
associated [live.remote∼] is freed.
2.11 List and use objects’ children

Each object has some children, or almost all.
We are going to try to list all tracks’ IDs for instance from the Song object. Tracks are children of
the Song object.
Look at that patch
Playing with children objects
On the left, we are focusing on the liveset itself and we are observing the property ‘tracks’. Indeed,
children in that case are also properties.
We can see that in the LOM:
Observing child objects as properties
We can observe all Song’s children of the type tracks. We could do the same with scenes
or return_tracks.
This is a tip to keep in mind and is a not a very well known one.
It provides a way to:
• keep an up to date list of tracks’ IDs

• know the exact current number of tracks in the liveset
[zl.len] measures the length of the list coming from live.observer. We know that this list contains
ID’s which use two list entries each **
Imagine for instance that you need to keep a trace of the states of all clip’s in all tracks. If we were
to create a new track or remove one, we would have to update our clips’ grid. This is a nice way to
trigger things when we add or remove a track!
Let’s check the patch on the right.
We can see that we can send other types of messages to [live.path]:
• getcount <child> provides the number of children of the type
All this information is sent out the right outlet of the [live.path] named dumpout .
2.12 live.thisdevice tool

[live.thisdevice] is another Max for Live specific object. It is really simple.
live.thisdevice
This object sends a bang from its leftmost outlet when the Max for Live device in which it is placed
is loaded/initialised.
We use it in place of the classic [loadbang] object in Max for Live devices.
It provides also a way to observe the state of the device itself: it sends a 0 or a 1 from its middle
outlet depending on whether the device is disabled or enabled.
This can be useful to eventually switch off or on a CPU expensive process when we don’t need it,
for instance.
At last, the rightmost outlet provides the edition mode that can be:
• Preview mode
• Non-Preview mode
What is it all about ?
2.13 Mode Preview & performances

I was waiting for probably the end of the chapter about the LOM to talk about Preview mode and
especially the computer performances with Max for Live.
Mode Preview
In Preview mode, we can edit and alter the Max for Live device exactly as if it was still connected
to Live. If we act on the patch itself, this can have an effect on Live itself.
In that mode, the device is grayed in the devices’ chain in Live and indeed, we cannot control the
device using Live’s UI.
In non-Preview mode (they call it Preview Off, anyway), this is totally reversed.
From a performance point of view and from the different experiences I have had and also some
discussions with Max and Live’s developers:
The best way to test our devices is to close the Max edit window and to test it in real
conditions
Normally, Preview mode provides a way to avoid this step, and indeed for 90% of the time it is
enough.
BUT, I often see some cases where the Live API was a bit confused and I had to close the Max edit
window in order to test it in real conditions in Live.
Especially, if you need to test for performance, you need real conditions. That sounds obvious, but
keep it in mind
Conclusions about the LOM
I strongly suggest that you dig and continue the road inside the LOM with these links and your
documentation:
• LOM³
• Live API Overview⁴
• live.path help⁵
• live.object help⁶
• live.observer help⁷
• live.remote\∼ help⁸
³http://www.cycling74.com/docs/max6/dynamic/c74_docs.html#live_object_model
⁴http://www.cycling74.com/docs/max6/dynamic/c74_docs.html#live_api_overview
⁵http://www.cycling74.com/docs/max6/dynamic/c74_docs.html#live.path
⁶http://www.cycling74.com/docs/max6/dynamic/c74_docs.html#live.object
⁷http://www.cycling74.com/docs/max6/dynamic/c74_docs.html#live.observer
⁸http://www.cycling74.com/docs/max6/dynamic/c74_docs.html#live.remote~
3 JavaScript in Max for Live
JavaScript¹ is an object-orientated script programming language.
JavaScript, also known as JS, is implemented inside Max which interprets it with a JS 1.6 engine.
Through the object [js], we can directly write our own scripts and execute them in Max.
The interesting thing is that we can make calculations, store values and other pure programming
wizardry, but especially we can manipulate a lot of Max’ features directly with JS like MSP and Jitter
stuff.
Indeed, lots of Max parts are exposed to JS and Max for Live parts too.
It means we can use JS to reach the LOM.
3.1 How to use Javascript in Max for Live ?

At first, we need to create a [js] object with the JS script filename as an argument.
Usually, I do it like this:
1/ create the [js] object with the argument <myScriptName.js>

2/ double-click on the [js] object while pressing CMD on OSX or CTRL on Windows
3/ editing the script and save it with CMD + S on OSX or CTRL + S on Window
¹https://developer.mozilla.org/en-US/docs/Web/JavaScript?redirectlocale=en-US&redirectslug=JavaScript
JavaScript in Max for Live 48
JS in Max for Live
We can see a script named calcul.js (which is basically a file on my hard disk) and this script can
be edited in Max, and even in Live itself as soon as we can see the [js] object in the device’s UI.
If we click on the message (addition 1 2 3 9), JS takes the first element of the list and tests if there
is a named function.
Here, it is the case and it passes the other elements of the list as arguments to the function, I mean:
1 2 3 & 9.
In the JS code, we can play with each argument passed to a JS script in Max by using the array
arguments. In each array’s slot, we have an argument available.
arguments.length is the length of the table, the number of elements contained by the Array.
In the code, I made a loop that checks all arguments and stores them and makes a sum 2 elements
per 2 elements.
outlet(<outlet’s index>,<value>) pops out the value <value> from the outlet with the index
<outlet’s index>.
This is a basic calculator.
I won’t and unfortunately cannot make a whole JS course right here. Or fortunately for us, maybe.
I’d strongly suggest you to check these links :
• JavaScript’s official documentation²

• The four JS’s tutorials in Max’ help
• JS in Max³
I’d really push you to study JS in Max because it can be a massive helper while patching.
Indeed, as soon as we have to handle repetitive tasks like loops in patch, we are often
building a lot of spaghetti-styled patches. With JS, a basic ()for loop can be the only one
statement needed, which can be much more simple.
3.2 Controlling Live API through JavaScript

The Live API is the programming interface that provides a way to access the LOM.
The different Max for Live objects we have already played with also provide a way to access the
LOM.
Now, let’s check how to do that with JS.
Live API & JS

Two important documents are :
http://www.cycling74.com/docs/max6/dynamic/c74_docs.html#jsliveapi⁴
http://www.cycling74.com/docs/max6/dynamic/c74_docs.html#live_api_overview⁵
The principle is simple to understand: we can use all the LOM’s objects with JS.
²https://developer.mozilla.org/en-US/docs/Web/JavaScript?redirectlocale=en-US&redirectslug=JavaScript
³http://www.cycling74.com/docs/max6/dynamic/c74_docs.html#javascriptinmax
⁴http://www.cycling74.com/docs/max6/dynamic/c74_docs.html#jsliveapi
⁵http://www.cycling74.com/docs/max6/dynamic/c74_docs.html#live_api_overview
Act on the liveset
Let’s look at a small example in which I will show you how to trigger clips using the clipslots object
exactly how we did before but by using JS.
JS triggering clips
Here is the code:
inlets = 1;
outlets = 1;
clip_slot_grid = new Array();
function scan()
{
root = new LiveAPI(this.patcher, "live_set");
num_scenes = root.getcount("scenes");
num_tracks = root.getcount("tracks");
for (x = 0 ; x < num_tracks ; x++)

{
clip_slot_grid[x] = new Array();
}
for (x = 0 ; x < num_tracks ; x++)

{
for (y=0 ; y < num_scenes ; y++)
{
clip_slot_grid[x][y] = new LiveAPI(this.patcher,"live_set","tracks",x,"cli
ots",y);
}
}
function fireClipSlot(x,y)
{
clip_slot_grid[x][y].call("fire");
}
In the JS code, I declared an Array clip_slot_grid. This one stores all the LOM objects I want to keep
available like, here, the whole clips slots of my liveset’s clips grid.
The function scan() scans the whole clips’ grid and stores at the place clip_slot_grid[x][y] the
clipslots object in the track x at the scene y (be careful, we start at zero each time)
Please keep in mind that it is the objects’ instances that are stored!
If we do it like this, then we have access to each function, and each property of all objects directly
by using the right element of our Array. Easy, isn’t it?
Let’s now describe a little bit more of our script:
root = new LiveAPI(this.patcher, "live_set");
Here, we are initializing the API and I decided to store the object Song directly in the variable
named root. This is the highest object in my hierarchy. This basically represents the liveset as you
already now know.
Then, as we know we can request the number of children of each type from the Song object. With
Max for Live objects, we should use the message getcount <child> with [live.path]
Here, we call the method getcount of the Song object which is stored in root and here is how we do
that:
num_scenes = root.getcount("scenes");
num_tracks = root.getcount("tracks");
We store those values in two variables: numtracks and numscenes

We have a double nested for () loop in which we store each clipslots object instance in clip_slot_-
grid[x][y] place of the Array, as we described before.
Here is the code:
clip_slot_grid[x][y] = new LiveAPI(this.patcher, "live_set", "tracks", x, "clip_s\

lots", y);
By doing this, we have an easy way to use a function named fireClipSlot(x,y) defined further with
the argument x & y which are respectively track index and scene index, the coordinates of the clip
slot we need to trigger. Thus, we have the opportunity to call fireClipSlot function (x, y) defined
below.
This function basically calls the function fire of the ClipSlot object instance stored in clip_slot_-
grid[x][y].
If we look at the patch itself, we have several elements.
We can firstly select the tracks’ and scenes’ indexes. These two integers are stored in [i] objects.
These latter store the value they receive in their rightmost inlet without transmitting it from their
outlets. Indeed, if we want them to send out stored values, we have to send a bang to the their
leftmost inlet. This happens when we click on the [button] object.
[pack] object prepares a list with the track’s index followed by the scene’s index previously chosen
and transmits the whole message to [prepend]. This latter adds a prefix according to the value it
owns as an argument. Said differently, it adds the keyword fireClipSlot in front of the track’s index
and the scene’s index , and this is the way to call the corresponding function in the JS object.
Aiming at LOM’s object using JS

We have to use this kind of instantiation:
new LiveAPI([callback], <path or ID>);
[callback] is the name for callback functions and is optional. We’ll see later what a callback function
is.
This call with new LiveAPI() creates and returns an object. This object is stored inside a variable in
order to keep it and to provide an easy way to retrieve it. With this, we can call functions related to
these stored objects, change property’s values: finally, this is what we did with [live.object].
The <path or ID> part shows that we can also use an ID, if we already have it. We can also use the
canonical path instead of the ID if we don’t know the latter. The Live API will call internally and
automatically the equivalent of [live.path] will convert the path into an ID.
Properties’ observing and callback functions
We just learned how to act on the liveset through a small example which showed how to store
objects’ instances in Arrays.
Let’s check now how we can observe as we did with [live.observer] but within the JS.
I’m going to show you how we can observe a live parameter’s value and transmit it to the patch
(outside of the JS) by using Javascript.
Look at the next patch:
Callbacks in JS
Here is the code:

inlets = 1;
outlets = 1;
var temp;
function setup()
{
root = new LiveAPI(this.patcher, "live_set view");
}
function learn_this()
{
param = new LiveAPI(this.patcher, callback_value, root.get("selected_parameter")\
);
param.property = "value";
}
function callback_value(args)
{
outlet(0,param.get("value"));
}
We have two functions calls that we perform outside of the JS object:
• setup()
• learn_this()
setup() stores the instance of the Song.view object inside the variable root.
learn_this() provides a way to keep the parameter just selected with the mouse (or with a MIDI
controller through the MIDI Mapping) and to observe its value.
In the JS, we have to change a bit of the syntax of the object’s instantiation using new LiveAPI().
Indeed, we add the name of the callback function inside the call AND we define the property to
observe like this: <objcet>.property = <property name>
How does it work?
Look at the function learn_this()
I instantiate that basically with the name of the callback (previously described as an optional
argument):
param = new LiveAPI(this.patcher, callback_value, root.get("selected_parameter"));
Please, I didn’t define a path or an ID… I could add the word: WOW!
However, please focus on the arguments passed to the object’s constructor LiveAPI():
root.get("selected_parameter")
root contains the object instance Song.view which provides a property named selected_parameter.
This one always contains the ID of the current selected parameter in Live. Interesting. Let’s resume:
Then, I’m defining the property I want to follow.
param.property = "value";
A DeviceParameter provides always the property value. This latter stores, as its name describes,
the value of the property itself.
If I summarise, we can observe the value of the parameter selected in Live and fixed by the call of
the function learn_this in Max for Live. If this value changes, the callback is called.
The callback function is called each time the property’s value changes.
It avoids making CPU expensive polling in which we would make a huge amount of request per
second even if the property’s value doesn’t change. It would consume a lot of our CPU’s cycles for
nothing!
The listeners/callbacks strategy is widely used in the UI design field.
Unlink an observer with JS When we want to unlink an observer, we need to send “ID 0” to it.
However, even developers are a bit confused about this. And I don’t say that without having
discussed it with them.
JS uses a garbage collector. Everything that isn’t used anymore, for instance outside the scope of a
function, is progressively destroyed and the underlying memory is freed.
However, I (we) noticed that even by manually destroying objects using the function delete, even if
we take care about emptying its value before by putting the value null inside of it, sometimes, the
listener wasn’t unlinked… I guess this will be improved in later versions.
Practically, this can be annoying in some specific cases.
You can maybe read more about that on my own blog and eventually comment too :
http://julienbayle.net/blog/2012/02/ability-to-cleanly-destroy-liveapis-callbacks⁶
We are now ready to create Max for Live devices.
⁶http://julienbayle.net/blog/2012/02/ability-to-cleanly-destroy-liveapis-callbacks/
4 Creating your own Max for Live
devices
One of the most interesting opportunities Max for Live offers us, is the ability to create our own
devices for Live. We are going to see that we can create very different types of devices.
Before this, we are going to explore some required concepts.
4.1 preset & pattr

When we are talking about Max for Live devices, we need to think of them as Ableton Live devices.
This means that our Max for Live devices have to be totally and smoothly integrated in Live’s presets
system.
Importantly, each device’s parameters and settings must be:
• storable within liveset files

• storable as device presets
For this purpose, it is required to use the pure Max’ preset system combined with the Max for Live
one, as we are going to see.
Max for Live specific objects VS the Native Max objects

Let’s make a small experiment.
Look at this device:
Presets and devices

Creating your own Max for Live devices 57
There are Max for Live specific UI objects ([live.dial] & [live.text] at the left) and also two native
Max objects ([dial] & [toggle]) at the right.
With Max for Live, ONLY specific Max for Live UI objects (those with a name beginning
with “live.”) are stored within livesets, as presets and when we are copying/pasting a device
from one track to another.
I can prove this: I set the two dials’ values to their maximum and I clicked on the [live.text] and
the [toggle].
All values have been modified
Then, I copy/paste the same device to another track named 6 MIDI

Copy/paste only keeps the values and settings of Max for Live specific UI objects
The copy/paste operation only kept the values and settings of Max for Live specific UI objects. The
native Max UI objects resorted to their initial/default values.
Let’s check now how it works when we try to save the current settings of the device as a preset. I
did that and named the preset mod.
mod preset saving
Then, I paste that preset in a new track named 7 MIDI

Only Max for Live specific UI objects’ states are stored in presets
Only Max for Live specific UI objects’ states are stored in presets.
The same goes for the liveset: only specific UI objects’ settings are stored with the liveset itself.
Automation and MIDI mapping works like that too.
Natively, we can ONLY apply automation in Live with Max for Live UI objects. As proof, check
the next snapshot: we can see my own UI elements in the automation list but not the default ones
(named default dial & toggle)
Only Max for Live UI objects are available for automation and modulation purposes
But there is a way to use all Max elements as we would like: pattr with the attributes parameter
mode enable enabled.
MIDI mapping can only be used with Max for Live UI objects. It CANNOT be used with other pure
Max UI objects excepted by using MIDI parsing, of course.
Only Max for Live UI Objects can be used for keyboard and MIDI Mapping
pattr & parameter mode enable

I’d suggest you to read the pattr’s documentation at this point. You can find it here¹ and some very
useful tutorials too:
• tutorial 1²
• tutorial 2³
Look closely what I added to my patch:

¹http://www.cycling74.com/docs/max6/dynamic/c74_docs.html#pattr
²http://www.cycling74.com/docs/max6/dynamic/c74_docs.html#pattrchapter01
³http://www.cycling74.com/docs/max6/dynamic/c74_docs.html#pattrchapter02
pattr helps native Max UI Objects
I added a [pattr] object with a name as an argument for each native Max UI object I want to use as
if they were Max for Live objects.
If I look at the inspector for each [pattr] object, here is what I can see:
Parameter Mode Enable enables some Max for Live features for Max objects
Importantly, we can see the Parameter Mode Enable attribute checked and enabled.
As soon as we check it, some new parameters appear below in the list within the inspector. We can
recognize these parameters: these are the same ones visible in any Max for Live UI Objects.
As soon as we check it, we link the [pattr] object to Live’s parameter’s system.
I won’t describe all attributes but only those which are very important to our study.
Parameter Visibility
Parameter Visibility defines how Live will handle this parameter
It sets up the parameter handles by the [pattr] object, i.e. the one linked by a patch cord to the
middle outlet of the [pattr] (also called bindto). This parameter can be:
• automatable and storable

• only storable
• hidden
In the first case, the Max UI Object linked to the [pattr] will be potentially automatable and be
stored with the liveset and the device’s presets too.
In the second case, this parameter will only be stored, and not available for automation.
In the last case, it basically won’t be handle by Live.
You may be thinking: why would we use hidden when we could simply not connect the Max UI
object to a [pattr] ? The reason is easy to understand. We could build a preset bank within this
device and handle its presets directly inside of the device and not in Live.
autopattr & pattrstorage
You probably noticed some objects like [autopattr] & [pattrstorage].

[autopattr] links all objects that are not attached to [pattr] to the [pattrstorage] system, this latter
is also a way to have an internal preset bank for all linked objects. For instance, this is also the way
to attach all Max objects to [pattrstorage].
The patch, the clientwindow & the storagewindow
If we send the message clientwindow to [pattrstorage], we can see that a window appears: the
clientwindow displays all parameters handled by the related [pattrstorage].
If we send the message storagewindow to [pattrstorage], we can see that another window appears:
the storagewindow shows a set of slots in which we can store whole settings for all parameters.
One of the main interests of [pattrstorage] is the fact we can retrieve all parameters directly by
recalling a slot.
We can also morph between parameters stored in 2 separate slots. This is why we have a column
named interpolation and we have a specific setting for each parameter. We can even choose the
interpolation’s curve
[pattrstorage]’s help can really help you to understand the full range of features: http://www.cycling74.com/docs/m
docs.html#pattrstorage⁴
A table summarizing parameters in Max for Live & Live

Here is what you need to keep in mind about the handling of Max for Live & Live parameters.
⁴http://www.cycling74.com/docs/max6/dynamic/c74_docs.html#pattrstorage
Max for Live and Live parameters
4.2 Freezing Max for Live’s devices

Max’ patches can use native objects (those delivered when you buy the software) and subpatchers.
In that case, our patch is autonomous and can be shared with other programmers directly, without
taking special precautions, by supposing, of course, that we have the same version of Max.
But consider the following situations:
• we have external objects we coded ourselves with the Max6 SDK

• we use abstractions referring to other patches on our hard drive
Then, we’ll have dependencies and if I send you the patch, your computer won’t be able to solve
them if I only send you the root patch
With Max for Live, we have two ways:
• we can handle dependencies ourselves by copying them and sending them alongside with the
main patch
• we can just let the Freeze feature work for us
Obviously, the second solution is the one I would strongly suggest.
Handling dependencies manually

In the first case, we need to list the whole dependencies of our patch by clicking on List Externals
& Subpatcher Files like this:
List all dependencies
Here is the result, in the example case :
The dependencies list
We can see here that we don’t have specific dependencies: the first column shows us only Max
Objects.
Freezing devices
With Max, we can create collectives & standalone applications as explained at the very beginning
of this guide. But we cannot do that with Max for Live because there is no runtime , except for Live
itself.
But we have the device freezing feature:
Freeze button
As soon as we freeze it, it becomes non-editable. We have to unfreeze it in order to edit it again.
A frozen device can be saved, of course. And at this very moment, the .amxd file of the device will
contain ALL dependencies of the main root patch. Usually, its size is increased a bit compared to the
unfrozen file, especially if you are using a lot of dependencies.
Then, I can safely send you the new frozen device (the .amxd file) because all dependencies required
to use it are included!
This is the main way to distribute devices.
4.3 Presentation mode

There are two display modes in Max, and also in Max for Live:
• Patching Mode (Editing Mode)

• Presentation Mode
Patching Mode
Patching mode can be locked or unlocked. If the patch is locked, we cannot edit it. This is just a way
to play and test it without modifying it by mistake.
We can switch between locked and unlocked mode by pushing the small lock icon at the bottom left
as we can see here:
Locking patches
When we have finished editing our patch, we may want to only allow some UI elements on the
screen, but not the whole set of elements.
This is exactly the purpose of Presentation mode
Presentation mode
Every object in Max can be set up to be visible in Presentation mode or not. It is possible for all
objects in Max and not only UI objects.
Build your own Presentation UI

There are three steps:
• choose and set some objects in Presentation mode

• place your UI elements on the display as you want to use them in live performance, for instance
• set up the patch itself in Presentation mode
We select objects that we want to see in Presentation mode. We can use multiple selection by
selecting all objects while keeping the SHIFT key pressed.
Then we display the Inspector for our selection and we check the attribute Include in Presentation:
Include some objects in Presentation mode

We can also use some shortcuts:
• CMD+SHIFT+P on OSX or CTRL+SHIFT+P on Windows to include objects in

Presentation
• CMD+ALT+P on OSX or CTRL+ALT+P on Windows to remove them from Presen-
tation
We can easily check if an object is set up to be in Presentation mode or not: if they are, there is a
rose border outlining them:
We can easily see which objects will be in Presentation mode or not
Then, we can switch to the Presentation mode by clicking on the icon showing a small screen as
shown here:
Enable/Disable the Presentation mode
We can also use the shortcut:
• CMD+ALT+E on OSX or CTRL+ALT+E on Windows for Presentation mode switch-

ing
In Presentation mode, we only see the objects we included in it.

Presentation mode
We can move them and place them exactly as we want to design our interface.
We design our final device’s UI
Then, we set up the patch in order to make it open directly in Presentation mode by default: this is
the only way to have a Max for Live device UI in Live that fits with the Presentation mode.
Let’s right click on the patch itself and choose Patcher Inspector:
Set up the patch iteself
Then, we have to check Open in Presentation:

Check the Open in Présentation attribute
Save the device (CMD+S or CTRL+S) and close the edit Max window.
Check the render in Live:
Finished device in Live
We now have our own device UI.

Now, with all the knowledge we have, I’m about to give you some real device examples.
This guide isn’t a proper Max course: probably, some concepts will be a bit tricky to
understand but I’m going to describe things very simply.
4.4 Limitations of Max for Live compared to Max

Max for Live is Max operated by Live, which means we have some limitations coming from Live
itself.
We have also identified in the first chapter: if you also own a Max license, you can run it outside of
Live and Max won’t be limited.
Max Limitations
It is possible to make data communications using [send] & [receive] between two devices in Live.
BUT Cycling’74 developers warn us about latency/performances.
In real-life, it really depends on what kind of data flow we want to use between objects.
If we only send data sporadically, not in continuous flows, we won’t feel any latency. I have tested
and used to do that sometimes without any problems. Be careful if pure continuous data flows,
anyway.
As we have also discovered, with Max for Live, we cannot create collectives or standalone
applications.
Audio Limitations
Max for Live doesn’t directly use Max audio drivers as if it were Max standalone.
Indeed, Max for Live’s device’s inputs & outputs are its proper inputs and outputs in the devices’s
chain in Live.
Max for Live devices are limited to 2 audio channels (left/right) and we also know that Max (more
exactly MSP) can handle up to 512 channels.
Then, [send∼] & [receive∼] cannot send signals between devices: we have to use Live’s own
routing features.
MIDI limitations
Max for Live doesn’t use its own MIDI drivers too.
When we are editing a Max for Live device that can send and receive MIDI data (MIDI FX or
instrument) then it can be confusing as at this particular moment (when Max for Live is in edit
mode, i.e. the Max edit patcher is opened but still operated by Live), Max uses its own MIDI drivers!
However, in normal use, Max for Live devices can receive MIDI information from the previous
devices in the chain, or of course from the clip inside the same track. Of course, all of these depend
on the routing and the monitor setting’s of the considered track in Live
4.5 MIDI & audio specific objects for Live connection

4 objects are required:
• [midiin]
• [midiout]
• [plugin∼]
• [plugout∼]
midiin is the MIDI input

In a device that accepts MIDI information incoming (i.e. a MIDI FX or an instrument), we can grab
all MIDI data sent to it inside the patch by using [midiin].
midiout is the MIDI output

Following the same logic, if a MIDI device can send MIDI data (i.e. a MIDI FX), we have to send all
data we want to go out of the device to the object [midiout].
plugin∼ is the audio input

If a device can receive an audio signal (i.e. an audio FX), we can grab this signal from within the
patch by using this object [plugin∼].
plugout∼ is the audio output

If a device can send out an audio signal (i.e. and audio FX or an instrument), we have to send that
signal inside the patch to the object [plugout∼].
4.6 An example of MIDI instrument

A MIDI instrument is a Max for Live device which owns a MIDI input object [midiin] & and audio
output object [plugout∼].
By drag’n’dropping the default Max for Live Instrument into a track in Live, it creates a device
already including [midiin] and [plugout∼] objects:
A MIDI instrument ready to edit
Here is a very basic synthesizer I built.

Small Synthesizer
We can choose the waveform as sine (Max object [cycle∼]) or saw wave (Max object [saw∼])
There are also two rotary knobs for volume and decay.
We grab all MIDI messages incoming to the device by using [midiin] and we filter them by using
[midiselect] in order to keep only MIDI notes.
Then, with a bit of Max’ wizardry, we keep only MIDI notes with a velocity greater than 0. Meaning,
we drop all MIDI notes off messages.
Then, we send the pitch number of all incoming MIDI notes to the message note $1 which transmits
the result to the proper synthesizer object.
Its structure is very simple and based on the object [poly∼].
Cheap synthesizer
[poly∼] is an object providing a way to build polyphonic synthesizers, i.e. synthesizers with multiple
voices, i.e. able to play more than one voice at the same time, with a voice number limit of course.
Beyond that limit, i.e. if we send another note to the synthesizer, we can set it up for instance to cut
the previous notes off.
We have to build another patch. This one is saved as a proper .maxpat file and we can instantiate it
inside [poly∼] object.
The proper synth itself
I want to focus on three specific objects:
• [mtof]
• [line∼]
• [selector∼]
The first one converts a MIDI note pitch number into a frequency to tune correctly both oscillators
[cycle∼] & [saw∼] in order to play the correct note.
The second one provides a simple way to apply an envelope to anything: here we use it to create an
amplitude envelope to our signal.
The third is basically a signal selector. We can switch it by sending a command to the leftmost input
and it will commute inputs to the output, or not. 0 means all off: no input signal is connected. 1 means
the signal in the second input is connected to the output, 2 means the third input is connected etc.
Then, check how the implementation works: we just have to set the voice patch filename as the
[poly∼] object’s argument.
We can also see that the output signal goes through another object that will multiply its value. It
provides a basic way to control the signal amplitude, i.e. its volume.
At last, the signal is sent to both inputs of [plugout∼] which is basically the plug to Live inside the
devices’ chain.
4.7 Example of a sound generator

The name synthesizer is used to describe a device that takes notes and then can produce sounds
according to these received notes.
I’m using here the term sound generator to define a device that doesn’t take any notes but will
generate & produce continuous sounds. A bit like a drone machine would do.
We could use a MIDI instrument type of Max for Live device as we could use an audio FX too. The
main thing is: it has to own an audio output to produce sounds in Live.
Let’s figure out how we can do that.
This basic sound generator shows some interesting concepts with few objects.
The sound generator integrated in a MIDI track
Let’s check the patch itself.

Basic and cheap sound generator
At the top, we can see an object [cycle∼] which is the basic sinusoidal oscillator in Max. We can
choose its frequency by using its first inlet.
The signal produced enters into [*∼ 0.] that multiplies its amplitude by the value of another signal.
This other signal is another [cycle∼] for which I designed a basic system to choose its frequency
by using transport relative time values, thanks to the object [translate]⁵ which converts relative
time quantities (here in note values) into absolute time quantities (here Hertz). If Live’s tempo is
modified, [translate] generates the new frequency value. It works a bit like an observer, internally.
We have here quite a fast amplitude modulation also named ring modulation. Lower values would
have made me name it tremolo.
The resulting signal is controlled by another volume control.
⁵The object translate
On the same logic, a noise generator is modulated by another sine wave controllable exactly as the
one described before.
Then, both signals are aggregated into another [*∼ ] which provides a way to control the global
volume.
I’m aggregating here both signals also in order to produce some small digital distortions. I
like this, just my cup of tea :)
The whole signal is separated into two equal signals entering in [omx.peaklim∼]⁶ that is a nice and
native Max sound limiter.
4.8 MIDI FX example

Let’s check another example: a MIDI Fx.
MIDI FX
We have to choose semi-tones interval (named delta space) and then we have to define a note
duration in ms.
⁶OctiMax limiter object
Let’s check the patch itself:
The MIDI Fx Patch
We grab only the notes in this devices and we unpack pitch and velocity into two separate ways.
We add a random number to the original pitch. This one can be positive or negative.
Then, we pass this new pitch calculated with the original velocity to [makenote] that will generate
note off messages after a certain amount of time.
4.9 Audio FX example

Let’s now create a basic and kind of ring-modulator, home-made!
A ring-modulator is an FX well-known and available in Live’s device Frequency Shifter.
Here is the device integrated in Live.
Audio FX
Here is the patch:
Audio FX patch
Here, we have the ring modulation, already described before and you also have a nice example of
dry/wet circuits controlled only with one rotary knob.
Conclusions & perspectives
I hope that you learned a lot with this guide.
There are a lot of new concepts coming from Max for Live for both type of users: Max users & Live
users.
We can notice, by seeing the big part of this guide focusing on the API & LOM, that Max for Live,
besides and beyond MSP & Jitter, IS THE WAY to control Live and especially in live performances
conditions.
We can for instance easily create a MIDI/OSC converter, hack scripts for our own custom controller,
handle a clips’ matrix with algorithms and all of this by using built-in and powerful Live features
(UI, audio interface, clips management, presets etc)
This guide is not the end, but just the beginning !
Indeed, I will update it, add things, fix typos etc. If you bought it, you will be informed about each
major update directly by email.
Take care and be creative,
Julien

Max For Live Ultimate Zen Guide

Uploaded by

Copyright:

Available Formats

Max For Live Ultimate Zen Guide

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Max For Live Ultimate Zen Guide

Uploaded by

Copyright:

Available Formats

Max for Live Ultimate Zen Guide

Become a Max for Live Master and discover a new way of

This version was published on 2013-11-17

©2013 Julien Bayle

About the review & edit of the book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv

1 Max for Live Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2 Live Object Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3 JavaScript in Max for Live . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

3.2 Controlling Live API through JavaScript . . . . . . . . . . . . . . . . . . . . . . . . . 49

4 Creating your own Max for Live devices . . . . . . . . . . . . . . . . . . . . . . . . . . 56

Conclusions & perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

Push & Live on the field

Mark Towers teaching Ableton Live in Leicester

This guide is not

If you find a mistake or a typo in this guide

If you would like to find out about courses on Ableton

What is an interactive visual programming environment ?

C++ language is based on text coding

This type of programming language requires a specific proramming workflow:

• we write the code

Here is an example of what I mean with a picture:

A Javascript text code and its visual equivalent

Max6 & Max for Live

Max vs Max for Live

Max, Max Runtime, Patch, Collective and Application

Max for Live doesn’t provide any runtime.

Liveset, liveset project, device and device in edit mode

as soon as we load a Max for Live device, we can also edit it !

Max for Live’s device EDIT button

Use & licensing

1.2 Ableton Live Basics

Projects & livesets

A MIDI track can contain:

We find the different devices within each of the following categories:

Category names are quite self-explanatory.

• Max MIDI Effect

We have the same classification in 3 categories.

Max for Live’s 3 different types

MIDI & Audio in Live’s devices chain

Let’s continue with Live’s basics.

MIDI Mapping Edit Mode enabled

PAY ATTENTION HERE

same set of knobs, whatever the liveset used ?

MIDI Remote Scripts

What are those scripts ?

• controlling Live (act on Live)

Enabling & Disabling MIDI Remote Scripts in Live

Where can we find the MIDI Remote Scripts on our computer?

These scripts are within Live’s folder hierarchy.

• C:\ProgramData\Ableton\Live x\Resources\MIDI Remote Scripts\ on Windows 7 & Vista

MIDI Remote Scripts on OSX

Hack and Script’ edit

• Push & hacks⁹

• Docs for _Framework & _Tools objects¹²

1.3 Max for Live specific objects

Max for Live specific objects

We can see two series of objects:

• at the top, objects able to provide control/feedback from Live itself,

I will describe objects at the top in the next chapter.