How to: stereo to three channels

Audacity 3 Help Forum
1

Hi
I got a stereo recording of an interview with 3 people talking.

Left: Alice 100% + Charlie 30%
Right: Bob 100% + Charlie 30%

To get a better transcription, I want to split this to three channels, with the ‘center’ Charlie on channel 3 (and normalize him).

I added a mono track and got the Nyquist Channel mixer plugin - that doesn’t seem to help in my case.

2

The old rule is, “You can’t un-mix sound”. But AI is getting smarter every day!

You can try Open VINO Music Separation or Vocal Reduction and Isolation

If that’s what you’ve got, Charlie should already be in the phantom center (the same in both channels).

3

I want to remove Charlie from the phantom center and put him on a third channel.

What is the correct way to:

  • eliminate everything that is in both channels?
    (to get Alice and Bob channels clean)
  • subtract the ‘clean Alice’ and the ‘clean Bob’ from the recording to be left with Charlie?

Based on a youtube-tutorial, I:

  • duplicated the stereo tracks
  • inverted the second pair
    = the phantom center voice ‘Charlie’ is mostly cancelled out.
    But:
  • the result does not sound as good.
  • now I’ve got 4 channels and do not know, how to merge them together.

Still don’t know how to subtract channels from each other.
Does anyone know add-ons to remix (more than two) (phantom) channels? Aren’t these just ‘arithmetic’ functions?

4

As far as I know the old “you can’t unmix audio” still applies. :frowning:

There is no simple mathematical solution. You can get various combinations of L, R, L+R, and L-R (or R-L). The problem is, L and R both contain Charlie and there is no way to get L-only or R-only without him.

…I know it seems like since you can completely eliminate the “center” you should be able to get the center alone. But no matter how you play-around with simultaneous equations, you can’t.

If start with 3 separate recordings and mix them into a stereo file with left, right, and phantom center, L-R will be a mono file without the center-audio, and what remains will sound perfect but without any center information.

But you can’t get center-only because L+R contains everything.

And, if you started with 3 separate channels/recordings you probably still have them and you don’t need to do this…

…Some commercial music has the vocals exactly in the center and they can be removed completely. But, that’s not the only thing in the center so you don’t get “professional” results. Vocal removal is usually just a “fun novelty effect” that’s not really useful.

The Vocal Reduction and Isolation effect uses advanced processing to try and isolate the center but it usually doesn’t work as “cleanly” like simple addition and subtraction.

5

This topic was automatically closed after 30 days. New replies are no longer allowed.