Anyone here working with Stable Diffusion to create sound?

MakePatchesNotWar's icon

Hello,

If so, i'd love to see/hear the results!

Two months ago when i saw the release of Stable Diffusion and learned of the ability to train it locally i had this idea of using spectral images of sound to train it and was wondering what it would come up with.

Is anyone here by any chance experimenting with that? Or does anyone know of research being done in this field? Because it seems the "mainstream" focus mostly appears to be on images but surely someone is working on the integration of AI into sound?

etenduskunstide osakond's icon

any news with SD implementation?

MakePatchesNotWar's icon

No, i haven't explored it myself, i was just curious to see if others were.

Jason Charney's icon

Look at this stuff from a few months ago -- yes, images of spectra as the training! It's such a cool idea.