Inverted phase - is there such a thing?

[Ervin Farkas]

Is there a way to extract something already digitally recorded? I only added the track in post, but I now hear that what the mikes recorded from the monitor running on location is too much (or I should say the kids were singing at only 50% steam)…

Is there a way I can invert the phase of the track and “subtract” it from the already recorded audio? Lowering the track’s volume (the one I added in post) only changes the way it sounds, will not lower it's general volume.

See/listen to attached. It's the mixdown from a multitrack recording (5 mikes plus track added in post) with Adobe Audition.

[Kit Hannah]

I'm not quite clear on what your final goal of this is. Do you want to just lower the overall volume of it? Because it sounds fine, the mix doesn't sound bad. Are you exporting this to cd or video?

[Ervin Farkas]

This is one of the better pieces, in most others the track (pre-recorded accompaniment) is too loud due to the mikes picking up the monitors used while recording. So I need to lower that somehow.

[Bill Ravens]

In some instances, inverting the phase will remove the vocal tracks, but, not always. It really depends on how the vocals were mixed in, in the first place. Most DAW's and NLE's that I know, have phase inverters for each track. If not, a simple VST plugin should provide that capability. I'm assuming, of course, you're on a PC.

[Kit Hannah]

But you only have a mixdown copy? Lowering the instrumetns without affecting the voices will be next to impossible to do. You may be able to adjust some frequencies to help lower the instruments or bring out the singing in a program like Sony's Sound Forge or use your NLE, but you'll have to play with it probably pretty extensively, or call your client and get an individual tracked mix.

[Ervin Farkas]

Originally Posted by Kit Hannah

But you only have a mixdown copy?

No, I did the recording myself, I have the multitrack Adobe Audition session with all the individual tracks.

[Abe Dolinger]

If you have a two-track recording of the accompaniment, try throwing that in there and reversing the phase on each channel. My gut feeling is that you're not going to get great results, since it bled onto your vocal tracks and is probably mostly echo and reflection. I think you would get best results if you recorded the vocals with two mics, and the left channel bled onto one and the right onto the other. In other words, try separating the tracks like that - with the mics on the left of the room panned more lefterly, etc. Hard to say if it will help much, could be worth a shot though.

[Ty Ford]

Originally Posted by Ervin Farkas

No, I did the recording myself, I have the multitrack Adobe Audition session with all the individual tracks.

Then remix it.

Regards,

Ty Ford

[Steve House]

It's really confusing what your issue is. Did you get the voice tracks you recorded and the separate instrumental accompanyment tracks you recorded out of balance in the mix? Then as Ty suggested just redo the mix. Or is it that when you recorded them, they were singing to pre-recorded playback and you accidently got too much of the playback bleeding into the mics on the voices?

[Ervin Farkas]

OK, let me try again. We have one choir of about 20 kids, one director, accompaniment playing from a CD player. Five mikes feed into an Alesis firewire mixer, firewire goes into my laptop running Adobe Audition. The five mikes are recorded onto five separate mono tracks in Audition. My mistake: I should have checked levels on location and figure out the CD player was playing too loud, got picked up by the mikes.

The plan was to add the accompaniment to a separate stereo track in post, which I did, but even if I mute this track, the accompaniment is STILL TOO LOUD - IT HAS BEEN RECORDED TOGETHER WITH THE KIDS VOICES, PICKED UP BY THE MIKES. I don't think I can discribe this any better... Probably the rookie audio engineer's mistake, I should have tested on location.

The question was: can I somehow lower the level of the accompaniment in post by subtracting an inverted phase signal? Theoretically if you add +5 and -5, you get zero, right?

I think Abe understood my situation - honestly I did not have any time to do any testing yet, I will try probably tonight, but as suggested, it probably won't come out clean.

[Bill Ravens]

Because of the existing phasing differences from mike to mike(distance from playback speakers to the mike plus room ambients) chances are you won't be able to get rid of the music background by adding inverted phase. Your vocals will be trimmed much more than the ambient music. You might try a noise reduction program, but, whether it works or not is doubtful.

In the future, the only way to really stop that kind of crosstalk is to output the accompaniment thru headphones to each singer. Even a shotgun mike will pick up background accompaniment.

[Greg Boston]

As I understand it, what he wants to do is take the CD player track that got picked up on the vocal mics out of the vocal track by using the actual music on separate tracks in post as a phase inversion reference.

IN THEORY, mixing those inverted music tracks with the vocal track would subtract out the music, leaving the desired vocal. That only works in a perfect world. As Ervin mentioned, +5 added to -5 equals 0, but in reality, you'll never have exactly equal and opposite amounts of signal present at any given time.

-gb-

[Ty Ford]

Originally Posted by Ervin Farkas

The question was: can I somehow lower the level of the accompaniment in post by subtracting an inverted phase signal?

Not in this instance.

Regards,

Ty Ford

[Kit Hannah]

I see what you're syaing now. Why not just lower the actual accompanying music volume? That way the overall volume isn't so bad. I would throw it into whatever NLE you are using and adjust things that way. You may have some luck as well with the bleed through track by trying some graphic EQ drops, meaning play with different frequiencies, drop them down to see if some of the music goes away without affecting the singing. There will inevetably be frequencies in there that the kids and music are not sharing too much, so you can at least reduce the effect. Good luck

[Andrew J Morin]

...it will bother me forever if I don't put this out here for you.

I'm just getting my feet wet in audio work, and I have also tried to get vocals from preshcoolers "singing" on stage with "guide tracks."

In your situation, I expect you could arrage your stage again just as it was for the performance, but obviously without the singers or audience. That done, you should be able to use the phantom-performers' mics to get a recording of the room's response from the CD track. Then you could try inverting that and sucking it out of your live track. Obviously synch would be a problem, not to mention the lack of the dampening effect of bodies in the room, but you sound desperate enough to try.