A granular delay effect using randomization of input parameters and pitch detection to alter the grain size.
https://github.com/schlegelflegel/rtapic-project/blob/main/Demovideo.mp4?raw=true
Move your soundfile (.wav) in the same folder as the PureData file.
Type the name of your soundfile in the message box or use the three examples.
There is small variety of presets given. Each one in a seperate textfile (.txt). In the upper left corner you can choose them, the first slot resets all values to default.
There are the parameters of the granular engine (orange) and three kinds of modulators, 4 LFOs (green), 3 random number generators with normal distribution and pitch detector. Every parameter with a slider can be controlled and modulated. For modulation click first on the button next to the modulator name and second on the vertical slider next to parameter which shall be modulated. The modulator is now attached to the parameter. The amount of modulation can be set with the vertical slider. The resulting value for the parameter is given by:
value = parameter + modulation * amount
One modulator can be attached to several parameters and even the parameters of the modulators can be modulated. To detach the modulation from a parameter click first on the detach button and second on the amount slider of the respective parameter. Click the post_parameter button to post the current values of the parameters to the PureData console.
Use the wet/dry slider to mix the altered signal with the original sound.
| Parameter | Description |
|---|---|
| grainsize | length of the grain in seconds |
| graindistance | distance between the grains relative to the grainsize |
| graindelay | delay of the grains relative to the original signal in seconds |
| grainpitch | added pitch in semitones |
| use relative pitch | uses the pitch detector to keep the grain at a constant frequency |
| grain envelope | choose between four envelopes to shape the grains |
| attack and release time | in seconds, can be altered for trapezoidal and cosine bell envelopes |
| Parameter | Description |
|---|---|
| frequency | rate of the LFO |
| curve | choose between sine, triangle, square and two different saw waves |
| Parameter | Description |
|---|---|
| frequency | rate of newly generated random numbers |
| value | expectation value of the normal distribution |
| variation | standard deviation of the normal distribution |
Returns the current frequency multiplied by the factor which can be set with the slider. Use negative values for inverted modulation.
G.O.A.T is separated into multiple separate components. The file src/goat.c bundles all these components into a single struct. src/goat_tilde defines a pure data object wrapper around the main goat struct.
src/util contains a few utility constants, definitions and functions. This is also where the circbuf resides, a generic purpose circular buffer with one writetap and a variable number of readtaps. This buffer is used by the main algorithm and the pitch detection.
src/control defines three classes for the control system: A control_parameter which can be seen as a single knob in the final audio application. Each parameter has slots for three control_modulator. These can be used to alter the parameters as described in the formula above. The control_manager keeps track of all parameters and modulators and calls their update methods.
src/modulators contains all the predefined modulators. Each struct "extends" the control_modulator struct.
src/scheduler contains all parameter definitions for the granular core. The main purpose of the scheduler class is to mark the position when a new grain should be created by the granular class.
src/graintable buffers all the grains before they are synthesized. This way we can have grains with different speeds overlapping each other.
src/synthesizer generates the audio output by creating active grains and reading their audio data from the main buffer
src/envelopbuf stores envelope information used by the grains.
src/granular contains references to the scheduler, graintable, envelopbuf and synthesizer. This also contains the main buffer with around 5 seconds of audio and the pitch buffer with the corresponding frequency information.
src/pitch contains classes regarding the pitch and vocal detection. Implemented is the bitstream autocorrelation algorithm as described here: https://github.com/cycfi/bitstream_autocorrelation .