4 ms·
I think we still need the step where the AI learns what a high quality sound library sounds like and then applies the previously learned abilities by triggering
by qwertox 3y ago
I think we still need the step where the AI learns what a high quality sound library sounds like and then applies the previously learned abilities by triggering sounds of that library via MIDI.
That way you'd get perfect audio quality with the creativity of a musical AI.
- eru 3y agoHow would MIDI get you eg a guitar being played dirty? Or some subtle echo that comes from recording in a bathroom?
- arrakeen 3y agothe AI designs and controls the effects chain and mastering too
- qwertox 3y agoIt would use a sampler and for the subtle echo effect add a reverb to the bus. https://www.youtube.com/watch?v=EQdp2QLiSYQ&t=187s https://www.youtube.com/watch?v=EQdp2QLiSYQ&t=187s
- sebzim4500 3y agoYou could have AI do some postprocessing. I think a similaar approach is the future for image generation, you have a model output a 3D scene, use a classical raytracer to do rendering and then have a final model apply corrections to achieve photorealism.
- jchw 3y agoI've always wished for something like that for image generation AI. It'd be much cooler/more interesting to watch AI try to draw/paint pictures with strokes rather than just magically iterate into a fully-rendered image. I dunno what kind of dataset or architecture you could possibly apply to accomplish this, but it would be very interesting.
- Auracle 3y agoI get what you’re saying, but if you watch Stable Diffusion do each step it’s at least kind of similar. If you keep the same seed but change a detail, often the broad “strokes” are completely the same.
- 3ds 3y agoIsn‘t that what suno.ai does?