Controllable Music Production with Diffusion Models and Guidance Gradients
This addresses controllable music generation for audio producers, but it is incremental as it applies existing guidance techniques to music.
The paper tackles realistic music production tasks like continuation, inpainting, and style transfer using conditional generation from diffusion models with sampling-time guidance, achieving results in 44.1kHz stereo audio.
We demonstrate how conditional generation from diffusion models can be used to tackle a variety of realistic tasks in the production of music in 44.1kHz stereo audio with sampling-time guidance. The scenarios we consider include continuation, inpainting and regeneration of musical audio, the creation of smooth transitions between two different music tracks, and the transfer of desired stylistic characteristics to existing audio clips. We achieve this by applying guidance at sampling time in a simple framework that supports both reconstruction and classification losses, or any combination of the two. This approach ensures that generated audio can match its surrounding context, or conform to a class distribution or latent representation specified relative to any suitable pre-trained classifier or embedding model. Audio samples are available at https://machinelearning.apple.com/research/controllable-music