This website contains supplementary material for the paper „Gaussian Process Synthesis of Artificial Sounds“.

Sound Selection Interface

The sound selection interface helps the user to find interesting sounds. The visualization is based on a t-SNE analysis of the sounds by the similarity of their MFCC features. When the user clicks on one of the dots, the corresponding sound is played. The kernel and its parameters are shown in the upper left corner. It is also indicated whether the wavetable or waveshaping synthesis is used. Note that a kernel can also be a multiplicative or additive combination of two base kernels.

Real-Time Control

The sound selection interface sends information about the selected sound via Open Sound Control (OSC). Together with the wavetables and shaping functions, which are precomputed, the interface can be use for live sound selection. In the video, the user controls the sound, which is generated within the multimedia programming environment Max/MSP.