Separating musical / non-musical content?

Hello, new owner of SpectraLayers Pro 11 here. I’ve spent the day working with the tool and it’s very impressive - I will certainly be getting a lot of use from it.

One task I’m attempting without too much luck, however, is isolating background music from other content. Unmix Crowd Noise is perhaps the closest single operation, but seems to leave “foley” audio intact. I believe a module specifically for intentionally separating into musical (at least instrumental music) and non-musical layers would be very useful.

In my case I’m trying to isolate some background instrumental music for reproduction, but I can imagine the more common desire may be to remove it; for instance to remove copyrighted music from the background of a video.

Of course it’s possible I’ve missed something, but I believe I’ve tried all of the existing modules - there are likely enough tools to brute-force a workable solution with enough effort, though, I’m sure.

A difficult task. A combination of Unmix Song and DeNoise might help but highly dependent upon the nature and levels of the music and noise.

How to define music and non-music?

Is spoken italien music in comparison to spoken german language?

Well it’s really down to AI training, right? But that’s also why I suggested “instrumental music” as the more likely target. Separating singing & speech I imagine would be a separate task, but who knows with AI separation!

plenty of teething…I was a newbie in June 2024

automatic results are not necessarily forthcoming for musical stem separation- SL is a re-balancing tool as it stands right now

I still spend weeks on film audio sections as I have logged plenty here in this forum

music separation is a full blown rabbit hole…maybe someday :slight_smile:

Really hard for AI to know what´s musical or what is not? Like Stockhausen, some Avantgarde or primitive music, that´s pretty hard task for separation…

If you or anyone needs assistance with separations please don’t hesitate to reach out to me. I am confident that I can achieve 2-3 hour long film separations in 1-3 days (depending on what content you would like separated). It definitely doesn’t take me weeks.

1 Like

It’s my documentary we are making…in French…I recorded the audio…I’m not separating a pre-existing film

It takes me weeks to run NR, separate voices where required (on two mics) then remix envelopes for each mic and then export…I’m talking about 5 min parts…I appreciate the offer…are you offering help for free?..there is no budget

It depends, I am on break for the holidays for Thanksgiving so I have time. If it’s something that is reasonable I may do it for free, if it’s something complex then I’ll discuss what I’ll offer to do.

1 Like

The Unmix Song feature doesn’t work for all genres either. I’m not asking for magically perfect one-click solutions (despite how close some of the SpectraLayers tools already get to that), just a tool which covers a target I don’t believe is covered.

Anyway I’ve already isolated the music I was after - I still say it could have been an easier task with a more targeted module, but whatever.