Differentiable Wavetable Synthesis
This addresses audio synthesis and manipulation for applications in real-time and interactive audio, offering a novel approach with practical efficiency gains.
The paper tackles neural audio synthesis by learning a dictionary of wavetables through end-to-end training, achieving high-fidelity synthesis with only 10 to 20 wavetables and enabling one-shot learning on short audio clips for tasks like pitch-shifting.
Differentiable Wavetable Synthesis (DWTS) is a technique for neural audio synthesis which learns a dictionary of one-period waveforms i.e. wavetables, through end-to-end training. We achieve high-fidelity audio synthesis with as little as 10 to 20 wavetables and demonstrate how a data-driven dictionary of waveforms opens up unprecedented one-shot learning paradigms on short audio clips. Notably, we show audio manipulations, such as high quality pitch-shifting, using only a few seconds of input audio. Lastly, we investigate performance gains from using learned wavetables for realtime and interactive audio synthesis.