repository·v3·Indexed 19 days ago
https://github.com/gibber-cc/gibberishA high-performance audio API for the browser designed for fast synthesis using per-sample techniques. It provides low-level primitives and high-level synthesizers, effects, and sequencers, featuring capabilities such as single-sample feedback loops and audio-rate modulation of scheduling. Includes various instruments like FM, Complex, Karplus, and TR-808 emulations (Kick, Snare, Hat, Cowbell, Conga), as well as polyphonic implementations and unit generator routing.
The PolySynth object is a polyphonic implementation of the Synth instrument. It acts as a bus (Gibberish.Bus2) using the Gibberish.mixins.polyinstrument mixin.
Behavior:
maxVoices is set during instantiation.PolySynth instance affect all child voices.a = PolySynth({ maxVoices:3 }).connect()
a.chord([ 330,440,550 ])Sequencer2 objects are used to schedule sequence calls to methods, property changes, and the execution of anonymous functions. Unlike standard Sequencer objects, Sequencer2 supports audio-rate modulation of timing via its rate property.
By default, rate is 1. A value of 2 causes events to occur twice as fast, while 0.5 causes them to occur half as fast. You can map any mono audio signal (an ugen) to the rate property to modulate the speed of events dynamically.
To use a Sequencer2, you must provide a target and a key. The key determines which property or method on the target will be controlled. If the key refers to a method, that method is called; if it refers to a property, that property is assigned a new value from the values array.
kick = Gibberish.instruments.Kick().connect()
seq = Gibberish.Sequencer2({
target: kick,
key: 'note',
values: [110],
timings: [11025],
rate: Gibberish.binops.Add(
1,
Gibberish.oscillators.Sine({ frequency:.25, gain:.75 })
)
}).start()Gibberish provides two types of bus unit generators to sum multiple inputs:
Bus: Sums mono inputs.Bus2: Sums stereo and mono inputs into a single stereo signal.Both types support the disconnecUgen(ugen) method to disconnect a specific unit generator from the bus.
Both effect and filter prototypes include a defaults object containing a bypass property. Setting bypass: true completely removes the effect or filter from the audio callback, effectively bypassing it.
// Example of bypassing an effect
myEffect.defaults.bypass = true;The PolyKarplus object is a polyphonic implementation of the Karplus instrument. It acts as a bus (Gibberish.Bus2) using the Gibberish.mixins.polyinstrument mixin.
Behavior:
maxVoices is set during instantiation.PolyKarplus instance affect all child voices.a = PolyKarplus({ maxVoices:3 }).connect()
a.chord([ 330,440,550 ])The PolyFM object is a polyphonic implementation of the FM instrument. It uses the Gibberish.mixins.polyinstrument mixin and acts as a bus (Gibberish.Bus2).
Behavior:
PolyFM ugen.PolyFM instance (like cmRatio or index) updates all child voices simultaneously.maxVoices option must be set during instantiation to determine the number of available voices.a = PolyFM({ maxVoices:3 }).connect()
a.chord([ 330,440,550 ])The PolyMono object is a polyphonic implementation of the Monosynth instrument. It acts as a bus (Gibberish.Bus2) using the Gibberish.mixins.polyinstrument mixin.
Behavior:
maxVoices is set during instantiation.PolyMono instance affect all child voices.a = PolyMono({ maxVoices:3 }).connect()
a.chord([ 330,440,550 ])The SSD (Single-Sample Delay) is a special unit generator used to create feedback loops within the audio graph. It records a single sample from a target ugen and makes it available for use in the next calculation cycle.
To implement a feedback loop:
SSD instance.ssd.listen(targetUgen) to specify which ugen to sample.ssd.out property to modulate other parameters (like frequency) in the current cycle.Note: ssd.isStereo can be set during initialization to determine if it listens to a stereo or mono signal.
ssd = SSD()
sin = Sine({
frequency: Add( 440, Mul( ssd.out, 100 ) )
}).connect()
// sample our sine oscillator
ssd.listen( sin )To use Gibberish in a web project, include the dist/gibberish.js script. You must set the Gibberish.workletPath to point to the location of gibberish_worklet.js before calling Gibberish.init(). Once initialized, use Gibberish.export(window) to make the API available on the global window object. All synthesis calls should occur within the .then() block of the initialization promise.
<!doctype>
<html lang='en'>
<head>
<script src='dist/gibberish.js'></script>
</head>
<body>
<script>
window.onload = function() {
// Set the path to the AudioWorklet file
Gibberish.workletPath = 'dist/gibberish_worklet.js'
Gibberish.init().then(() => {
// Export the API to the window object
Gibberish.export(window)
// Start using the API
Sine().connect()
})
}
</script>
</body>
</html>Gibberish components (Ugens) can be instantiated and connected to the master output using the .connect() method. If no argument is passed to .connect(), it connects to the master output by default. You can also chain connections to route audio through multiple components (e.g., syn.connect(fx).connect()).
// Connects to master output by default
kik = Kick().connect()
// Connects to a chorus effect, which then connects to master output
chr = Chorus().connect()
syn = PolySynth({ maxVoices:4, attack:44, decay:22050, gain:.1 })
syn.connect(chr).connect()To start a Gibberish audio session, call gibberish.init(). This method creates an AudioContext, a ScriptProcessor Node, and connects the output to the audio destination.
Options:
memorySize (int): Determines the size of the memory block used by all ugens. The default is 44100 * 60 * 20 (approximately 20 minutes of audio at 44.1 kHz). Increase this if you use a large amount of samples.Gibberish.init();To build the library, you need node.js and gulp installed.
npm installgulpgulp minifyThe output files will be located in the dist folder.
npm install
gulp
gulp minify