Audio AI: isolating vocals from stereo music using Convolutional Neural Networks

Re:Audio AI: isolating vocals from stereo music using Convol... by thibauts on Saturday February 16, 2019 @03:09PM (#19180667)

Of stereo tracks L and R, you invert R and add it to L, effectively canceling anything centered. This usually removes voices. If you subtract (invert then add) the result from the original L and R tracks you get centered sounds only. Results range from perfect to not effective at all depending on the songs you apply it to.

[ Parent | on HN ]

Re:Audio AI: isolating vocals from stereo music using Convol... by lightedman on Saturday February 16, 2019 @10:53PM (#19182583)

This is correct. It is a shame there are no music players that can do this natively. I have to do every track I wish to play along with by hand in Audacity (or if I'm using my Win2K laptop, Cool Edit Pro 2000.)

Alternatively (in theory, anyways,) you can do a spectral analysis, create a selection/range of frequencies to invert, and that can usually handle isolating individual instruments regardless of stereo isolation as long as the other instrument octave ranges do not overlap with the one you're isolating. I'd love a music player that could do that.

[ Parent | on HN ]