Load SuperCollider plugins into python.
The package is motivated by two distinct goals:
PyCollider makes it possible to run SuperCollider plugins with numpy arrays as inputs and outputs. This simplifies the development of new synthesis algorithms as numpy/scipy provide a well established framework for prototyping DSP techniques and tools to plot and analyze results. As such the package can be used complementary to SuperCollider as a development and debugging tool for new plugins. Because PyCollider loads the plugin binaries, you can be sure that you are executing the exact same code as when you run the plugins in SuperCollider.
Teaching the fundamentals of computer-music to composers can be complicated by the amount of different specialized (and often ancient) languages and software systems. To seriously learn about computer-music, some programming knowledge is necessary. While this comes naturally to some students it is an obstacle to others. By creating an infrastructure for sound synthesis and basic algorithmic composition techniques (e.g. the python pbind package) students can focus on concepts instead of syntactic and semantic idiosyncrasies of multiple different programming languages. For DSP and machine learning python has become one of the most wide-spread languages. Python, which was not developed to be a responsive low-latency real-time language will not replace systems like SuperCollider for real-time electronic performances, but for pedagogical purposes and offline composition work, a general purpose language, designed to have a shallow learning curve has some clear advantages.
The package can be installed with
python setup.py installYou need to have SuperCollider plugins installed and load them. To load a plugin (the binary) use the load_plugin function:
pycollider.load_plugin("/path/to/plugins/SomePlugin.so")
You must make sure to load the plugin files before you create any UGen object.
The following code generates 44100 samples of a 100 Hz sine-wave (returns a numpy array).
sig = SinOsc(100).render(44100)
Numpy arrays can also be used a arguments. In this example the previously generated sine-wave is used as frequency argument to another Oscillator resulting in a frequency modulated sound.
sig2 = SinOsc(sig).render(44100)
Additional to numpy arrays, unit generators can be used as inputs to other generators as well to create more complex modular synthesis networks. The following code generates a 220 hertz sine-wave multiplied by a 10 Hz Sawtooth oscillator (some form of amplitude modulation).
sig3 = SinOsc(220, mul=Saw(10).range(0, 1.0)).render(44100)
The unit generators can also be combined by arithmetic operators. To generate the sum of two oscillators you can just use '+' to add the outputs together.
sig4 = (SinOsc(100) + SinOsc(110)).render(44100)