Hi. I’ve just been trying out SL One in Cubase 11 on an old 4-track recording and I love what it’s allowing me to do. The only thing is I’m getting some artefacts appearing on the vocal track, presumably stolen from the backing (as there are equivalent ones there) and it sounds fine when I play them together. I had a play around with FFT size but couldn’t hear much difference tbh so stuck with the defaults in the end. But I wish it was doing a better job. Can anyone advise?

Cheers, C

Which processes are you applying?

Just a basic extract vocal then drag the layers into the project window. I bounced and tuned the vocal but the artefacts were there before. This is the first time I’ve used SL so all I know about it is a couple of introductory videos.

Stems extraction does not depend on the FFT Size (cf end of this page to see what processes depends or not on the FFT Size: The Importance of FFT Size )

If you want to further improve the results from the AI, you can use selection tools and cut/paste frequencies that were assigned to the wrong layer back to their proper layer. That being said, SpectraLayers One is pretty limited in term of selection tools, you’ll have better control with SpectraLayers Elements or SpectraLayers Pro.

Thanks, Robin, I’ll give that a go.