True Mid from Stereo

[Lpettro]

Hello, I have been scouring for an answer to this but have yet to find one. Does anyone know of a way to get "true mid" from a stereo track? I understand that people sometimes use mid to mean L+R but that is not what I am looking for.

"True mid" would be only what is common to both track. For example: if guitar is only on the left and bass is only on the right then "true mid" would be silent, whereas L+R would contain both. Anyway, that is what I want. Let me know if any of you know a way or if you would like me to clarify.

Thanks,

[kozikowski]

So that's Center Pan Isolation, the exact opposite of Center Pan Removal or Audacity Vocal Removal. You would think that the effect is a simple revision of the arithmetic used in Vocal Removal, but no matter what you do, you're missing one vital term in the calculations.

http://manual.audacityteam.org/o/man/tu ... ation.html

I don't know that I would call it "True Mid." Mid in a surround presentation carries dialog, not directional sound.

Koz

[kozikowski]

Oh, and the more compressed and MP3-ised your presentation is, the less likely the tools are to work. So if you're doing this on a ratty download, success may elude you, even if you do get the tools to work.

Koz

[steve]

There is no simple algebraic way to do that.

Have a look at the experimental "rjh-stereo-tool.ny" http://forum.audacityteam.org/viewtopic ... 89#p216689
As that plugin is not part of Audacity, any questions regarding the plugin should be addressed to Robert JH (the plugin author) by posting to that topic.

[Lpettro]

Thanks, that was kind of the impression I was getting. I understand that mathematically what you would need is the "complement" and there does not seem to be a way to get it easily.

It does seem like there should be a way to do it since it is basically phase cancellation where you keep what is cancelled rather than removing it. I also understand that it is easier said than done.

I used the term "true mid" because there is a lot of talk about mid/side processing and it makes sense that "side" is everything that is different between L and R so basically L - R. But then some people say that mid is (L+R)/2. It seems to me(and some others from the looks of things) that mid really should only be what is common between the two tracks. In any case, center isolation is another way of thinking about what I am looking for.

Thanks,

[kozikowski]

If Mid was really "pure," then you'd never be able to get back to Left and Right simple stereo from a recorder that shoots Mid-Side. MID says: "I have a flute." SIDE says: "put it over there." FM Stereo is Mid-Side. It's very efficient.

There is no simple algebraic way to do that.

True. You can get most of the way there with with tricky audio processing. It is required that the software understand the show, like a volume compressor, not just blindly add and subtract bits from each other. How well it understands the show defines the quality of the result.

Koz

[kozikowski]

there does not seem to be a way to get it easily.

At all. There is no direct relationship between the voiceless mono track and the original stereo show.
Koz

[steve]

It does seem like there should be a way to do it since it is basically phase cancellation where you keep what is cancelled rather than removing it.

No, it can't be done with phase cancellation, and that is why it is not easy to do.