The field is moving fast. Three major trends are emerging:
To use it as a true vocoder, you need two files: A (your voice) and a carrier (a synth or different voice). ddsp vocoder
Because DDSP leverages the mathematical efficiency of oscillators and DSP, it doesn't need to generate audio sample-by-sample like WaveNet. This makes DDSP vocoders capable of on standard CPUs, a massive win for live musicians. 3. Infinite Scalability The field is moving fast
If you have an old, damaged recording (e.g., a 1920s jazz singer with heavy crackle), you can use DDSP to infer the "Harmonic Distribution" of the voice and re-synthesize it cleanly, removing the noise while preserving the vocal identity. damaged recording (e.g.