How to isolate vocals in a sample?

  • Thread starter Thread starter barefoot
  • Start date Start date
barefoot

barefoot

barefootsound.com
I'm mixing electronic music and would like to isolate just the vocals from some of my samples. There are some 'vocal remover' plug-ins out there, but I can't find anything which does the opposite - removes the music and leaves the vocals.

Does anyone know of a trick for doing this?

Thanks,
barefoot
 
deja vu

then use this vocal removal plug-in, remove vocal
now you have music without vocal, right?
now invert phase of that signal ( "invert phase" in wavelab edit options )
mix that inverted-phase-without-vocal-music with original music ( in cubase or something ), and music will cancel itself, leaving more or less usable vocal
 
Igormeister,

Thanks, but I tried this and it doesn't work.

All the vocal removal techniques I know of use some variation of L-R, relying on the fact that most vocals are placed in the center. This makes sense if you think of it as follows:

L = 1/2vocals + X
R = 1/2vocals + Y
then
L - R = 1/2vocals + X - (1/2vocals + Y) = X - Y

Usually this is done with just the mids and highs to prevent the bass from getting removed as well. You're left with a mono music signal and almost no vocals.

However, no matter how you use the result, you still can't isolate just the vocals.

L - (L-R) = 1/2vocals + X - (X - Y) = 1/2vocals + Y = R
R + (L-R) = 1/2vocals + Y + (X - Y) = 1/2vocals + X = L
L + R - (L-R) = vocals + X + Y - (X - Y) = vocals + 2Y = 2R
Etc..........

I've tried every way imaginable but this simple phase inverting technique doesn't work. In fact, just looking at the math, it's clear why. You can't separate three variables having only two initial equations.

Surely someone has come up with a clever alternative method!

Thanks,
barefoot
 
well, i isolated vocals from some hip-hop song four weeks ago, but material was very simple in arrangement and sounds, very dry
- we need DSP-hungry-frequency-envelope-artificial-inteligency-follower with built-in thesaurus plug-in (tm)
:D hey Waves, do you hear me ?

( beta version comming soon, TDM & DirectX ) ;)
 
What I do (and suggest that you try) is EQ the music out.
When I record the passage, I pull the bass out, turn the treble up, and adjust my mid's (on my DJ mixer).

Then, I try to EQ more of the music out once I get it into the computer.

Then I use Multi-Dynamic Compression to squeeze the rest of the music out.

And then, if I have to, I'll send the vocals through the one of the computer EQ's again.

I hope that helps.

spin
 
You must log in or register to reply here.
Back
Top