Unfortunately, audiounit.js does not make the actual signal processing code any easier. You still have to write an Audio Unit in C++, which is not an easy task for a beginner programmer.
Because Xcode templates are difficult to write and seem to change with every version, audiounit.js is distributed as a node.js script. To install, first install node from the official website. Then, install audiounit.js with
sudo npm install -g audiounitjs
if you're still confused, there's a quick screencast on the install process.
First, create a configuration json file as described below. Then, run
audiounitjs myfile.json, which will create a project scaffold for you. Then it's a simple matter of programming - edit the
audio.cpp file to write the audio source, and edit the
ui folder to create your beautiful UI.
The Xcode project generated by the script has 3 targets:
Creating a project will provide a minimal example of a volume-changing playthrough effect. This simply connects the input to the output, adjusted by a user-settable gain.
Two other examples are included in the
examples directory, each exemplifying an important feature.
You can look at the "config_example.json" file to get you started.project This is the name of the output Xcode project file. It must be unique as it is used as a class prefix for Cocoa (if it isn't unique, your plug-in may have namespace collisions when other plug-ins are loaded). type The audio unit plug-in type. Common ones are 'aumu' for synthesizers, 'aumf' for effects that receive MIDI, and 'aufx' for effects that don't receive MIDI. A full list of types is in the apple document titled "Audio Unit Component Services Reference". plugin_id The ID for your plug-in. This must be four letters long and unique for each plug-in. manufacturer_id Your manufacturer id. name The name of your plug-in or app. May contain spaces. description A human-readable description that will be shown in certain plug-in hosts. company Human readable company name that will be shown in plug-in hosts width, height The height and width of your UI for the plug-in and mac app. The iOS app of course will be the size of the device's screen preferred_iPhone_output Set this to either "Speaker" or "Receiver" depending on where you want the output to happen.
index.html in your UI folder.
To communicate with the C++ audio processing code, use the
Volume in your C++ code, There will be an object called
Get()- returns the current value of the property or parameter
and the following fields:
OnChange- set this to be a function that will be called whenever the parameter or property changes.
Parameters have the following additional methods:
Set(value)- sets the current value.
BeginGesture()- notifies the plug-in host that the user has started to change the parameter
EndGesture()- notifies the plug-in host that the user has ended the current gesture
and the following additional fields:
OnBeginGesture- set this to a function that will be called whenever the plug-in host starts a gesture.
OnEndGesture- set this to a function that will be called whenever the plug-in host ends a gesture.
so, if you have a parameter called
var v = AudioUnit.Volume.Get(); would return it's current value, while
AudioUnit.Volume.Set(.2); would set the value to
Additionally, the AudioUnit object has three methods of its own for sending MIDI to your Audio Unit:
AudioUnit.NoteOn(note, velocity)- call this to trigger a note-on event
AudioUnit.NoteOff(note, velocity)- call this to trigger a note-off event
AudioUnit.SendMIDI(b1, b2, b3)- call this to send a three-byte MIDI message.
To write your C++ audio processing code, simply create an Audio Unit as described by the apple document entitled "Audio Unit Programming Guide". You can look at the provided examples for some more concrete hints.
fivescope example for more information.
In the non-plugin targets for iOS and Mac, all MIDI received by the system will be sent to your Audio Unit. MIDI is handled as in the Audio Unit standard - see the
monosine example for more information.