View Full Version : Vocal Eliminators?

10-22-1999, 10:05 AM
Any suggestions on this? Do they work or are they just glorified EQ which diminishes the overall quality of the original recording?

10-23-1999, 02:07 AM
I've tried the "Human Removal" plug, but I found it didn't do a very good job, so:

The obvious solution to vocal elimination is: If you assume that the lead vocal is centered, but that a fair share of the instruments are not, you can loose most of the vocal by reversing the phase on one channel in the stereo pair, and summing them up in mono. For this to work the levels and timing on both channels must be identical, and I mean identical. I've done this in protools, but I found that the system is not exact enough to do the trick well enough, so for that feat I had to go analogue. However, there's more to it if you want to retain an element of stereo imaging, and not loose too much of the sound.
What I did was:
1) setup your stereo source so that you have three parallell stereo pairs going in on the desk.
2) start with the "center pair", do the trick as described above. (Both pans in center.)
3) now, on the "left pair" (Both pans to the left), reverse the right input phase, and subtract eq from the left input to let some sonics thru from that channel (like the bass area and the high treble, also see what else you can subtract without letting the vocal thru)
4) Do as described in 3), for the "right pair" (Both pans to the right) but reverse the process to let sonics thru from the right channel)
5) You now have a left channel, a centre, and a right. Mix to taste.

...well?.. There's more to it, try using dynamic control creatively on the parts and the whole to optimize the result. You can set up key/side chains or a number of tricks, just try different feats to see what you get.

All this was done using analogue equipment, and I'm not sure you'll be able to do it that well in digital because of the infinitely small, but still, time lags, that can occur when processing.