DDSP
lets you combine the interpretable structure of classical DSP elements
(such as filters, oscillators, reverberation, etc.) with the expressivity
of deep learning
To start, you first have to initialize
SPICE so that you can use
SPICE to extract pitches from your audio.
const spice = new mm.SPICE(); await spice.initialize();
Extract Audio Features
spice.getAudioFeatures(audioBuffer)
Upload an audio file and call the `getAudioFeatures` method to get audio
features from your input audio.
Synthesize
To start synthesizing your audio, initialize mm.DDSP making sure to pass
in the URL of your own model or use one of our
4 models
(look for DDSP models).
const ddsp = new mm.DDSP(checkpointUrl);
ddsp.synthesize(audioFeatures)