Isolating voice from background music

[wcstarks]

I have a single channel voice recording with music in the background. Is there a way to isolate the voice from the music and keep just the voice? The descriptions of the built-in methods for isolating the voice seem to all depend on stereo recordings.

[DVDdoug]

There is something new you can try.

[kozikowski]

And just to illustrate the seriousness of what you're doing. #4.

Screen Shot 2021-06-05 at 9.04.50 AM.png (42.98 KiB) Viewed 121 times

So if that AI App fails, that's the end.

Koz

[kozikowski]

There is something new you can try.

I wish they had picked a simple voice with music as illustration rather than a massively processed song. I agree it's impressive and we'll be watching.

Koz

[wcstarks]

DVD Doug,

Thanks. This got most of the music out--not all, but much better that the original.

[kozikowski]

This got most of the music out--not all, but much better that the original.

"This" being the AI sound splitter? Most people who ask this are trying to make a studio performance in software.

Splitter's example rang alarm bells.

Closer, but not yet.

Koz

[wcstarks]

kozikowski,
I just needed to extract the voices from the background music recorded nearly 50 years ago, so I could better understand what was being said.

[steve]

I have a single channel voice recording with music in the background.

Audacity's tools for voice separation require a true stereo recording with the voice in the centre of the stereo field.
A single channel voice recording won't work with Audacity's tools.

As DVDdoug wrote, there's an on-line tool called "Spleeter" that may be able to help. You can find it here: https://splitter.ai/