LinuxMusicians

Is there any way in linux (Ubuntu Studio) to remove the voice from a song and vice versa to keep the voice and remove the music ? IN windows are many ways but since i have forevere "divorced" windows i want to learn in Linux,Ubuntu.The music filles i would be processing would be *.Mp3 and maybe *.Ogg and *.Wav !
Thank you !

To begin with, it’s not advisable to use lossy-compression source material (e.g. mp3) when you plan to do further processing.

Next, it depends on whether you want to process the files in real time or offline. The swh plugins available in Ubuntu (both LADSPA and LV2) list karaoke that might do that dynamically. There’s also Kn0ck0ut available at http://www.freewebs.com/st3pan0va, which is a bit more complex, but that is a Windows VST. There is an LV2 port in Arch, but it’s not in the Ubuntu repos.

If you want to process the files offline, the Audacity audio editor has a ‘Vocal remover’ plugin that can work both ways (either keep only vocals or ditch them). You can also set a frequency band for the plugin to operate in. It will load just about any audio file format you can think of.

I haven’t tried any of those, so I don’t have any experience in the field, sorry.

As always, extracting or removing the vocals from a track is a tricky operation and it depends on how they are panned in the original, so your mileage may vary.

By the way, if you’re a DJ, you might want to take a look at Mixxx (http://www.mixxx.org/). It’s also available in the Ubuntu repos. Never used it, though, I’m not a DJ.

It's pretty standard for the main vocal track to be the only perfectly centered track on CD. Everything else is generally panned one way or another. So you can split a stereo track to two mono tracks. Invert one side, and mix them together to get just the instruments. Or a resulting track that you could invert and mix with the starting track to get just the vocal. With various inverts and mixing to taste. You can do this in audacity or other audio editors. This is how a lot of karaoke tracks are made. Unfortunately with digital, they are not always perfect matches and you can end up with a faint vocal track. And if your starting point is more lossy, the process is less and less perfect. And it's not a law or anything to make CDs with centered vocal tracks, it's just a common practice.