Im using cubase 11 pro to extract some vocal stems from an ancient track I wrote many years ago. Its a copy of a copy of a copy so the quality isn’t great but the vocals are pretty solid. Ive run it through once and the results are ok but not great there’s a lot of bleed through of the underlying music. So I thought what if I used the processed vocal stems and ran that through spectral layers, surely it would have less music and more vocals so running it through a second time would give even better results, right ? no such luck, the result is exactly the same. How could this be ? surely if the source audio contains way less of the background music it should be able to extract a better quality vocal ? am I missing something here ? I did export the first set of stems to an audio file then imported it and ran it through spectral layers, but the result is the same, anyone got any pointers ?
Can you upload a sample somewhere ?