Center Isolation in Flowstone

[chackl]

Hello!

Did anyone do some kind of center isolation in FS / SM?

In audacity they have an effect called Vocal Remover and Center Isolation that realy gets good results.
Source Code of the Effect:
https://github.com/audacity/audacity/bl ... lrediso.ny
But translating this into FS may be a little to tricky to me.

Thanks!, C.Hackl

EDIT: Since the Usage would be an own upscaling from stereo to 7.1 it is not that important to get that plain results.

[k brown]

Center removal in a stereo source ('Vocal Remover') is easy - one channel to non-inverted vca, the other to an inverted one; when these two are summed together, anything that's identical in both channels will be removed. Center isolation is much trickier and I assume would require DSP.

[chackl]

Hi!

Yes i did some experiments with mid/side conversation that you decribed. But this was not the result i expected.
Doing a Center Isolation and subtract this from the original source would give you a "wide" stereo without any center in it. I could not get this result by using a mid/side conversation.

Regards, C.Hackl

[k brown]

Not sure I follow you - I didn't mention anything about mid/side conversion.

Mid-Side hasn't anything to do with center removal or center isolation from a stereo signal, but is rather a circuit for extracting a stereo signal from two mono ones - an omni (or, occasionally forward-facing cardioid or subcardioid) microphone coincident with a side-facing bidirectional mic. The mid-side matrix converts these two signals into a left/right pair of coincident cardioid patterns ('X/Y') with variable angles to each other.

I'm guessing you're referring to software that simulates these two mono signals extracted from a stereo one, but even assuming these do what they say, it wouldn't get you what you're after; the signal equivilant to the center omni mic would contain all sound coming from every direction, and not just the center info from a stereo signal, such as isolating a vocal from a stereo mix.

[chackl]

Well the mid/side schematic i remember is quite old. Maybe it was just named unhappy.

I'll refere to this one and to some tests:
http://www.moitah.net/download/latest/C ... ut_GUI.zip

It also contains a commandline tool:

Code: Select all

Center Cut CL v1.6.0
Copyright 2006-2010 J.D. Purcell
Center Cut algorithm by Avery Lee
FFT library by Takuya Ooura
http://www.moitah.net/

Arguments: source_file [switches]

Switches:
  -s <file>     Write side channels to file.
  -c <file>     Write center channel to file.
  -a <percent>  Scale output amplitude by percent (default 100).
  -d            Dual-mono center channel.
  -b            Move bass to side channels.
  -o            Overwrite output files without prompting.

"stdin" or "stdout" may be used in place of input or output files.

I'm not shure if I will get the results i am expecting - i may have to test it within the next days.

[tulamide]

I'll lean far out the window here, so maybe it's nothing substantial for you. Anyway.

The basic thought behind vocal removement is that vocals tend to be mixed very centered. A vocal remover you describe (from Audacity) will not succeed if the vocals are not mixed centered.

To "remove" (actually it's more like suppressing) the vocals, you first would do a standard mid/sides, 0.5 * (left - right) for the sides, 0.5 * (left + right) for the mid, and then you can cancel out one of these by running it against the original via sides - original or mid - original. The former would amplify the vocals, the latter suppress them. But of course, it is highly dependend from the mix of the source, and it isn't fool proof.

[chackl]

Hi!

Yes but still "Center Isolation/Remover" - So i am not searching a vocal-remover.

The CenterCutCL tool did fit my needs at the moment. It is sounding a little different than the Audacity Center Isolation. (and sometimes a little metalic what can adjusted by remixing them with some gain adjustments)

So at the moment i can start up my developments on this tool here.

Thanks