— turning shapes into sounds

2023 by Till Bovermann

A synthesis environment that utilises visual shapes as (2-channel) oscillator inputs such that the Lissajous curve of it resembles the original shape. Shapes were extracted by contours2wav, a custom python script made available on github.

an image turns into waveforms

an image turns into waveforms

The synthesis environment is implemented in SuperCollider. Currently, the following synthesis methods are implemented:

  • shapeDrone — a straight-forward drone synth with parameters for frequency, amplitude, and shape
  • shapeFM — an FM synth where the basic oscillators are shape-based; supports feedback modulation
  • shapeFMperc — similar to shapeFM but with a percussive envelope
  • shapeMorph — similar to shapeFM but with sound morphing capabilities

The instrument is controlled by a Sensel Morph controller. Each appearing touch point on the controller is mapped to a synth voice. Pressure and x/y position are mapped to synth parameters such as frequency, amplitude, and FM index.

The system allows to save synth states into four slots. Each saved slot can be globally controlled in frequency range, amplitude, and FM index. Additionally, parameters like frequency and amplitude of running synths drift over time which creates dynamics in an otherwise pretty static setting.

The combination of “parking” sounds in slots with their autonomous drift allows to layer synth voices to create evolving soundscapes.

The following video shows the synthesis environment in the context of a live performance for the open doors piece: