Removing vocals from a track.

[sonikbliss]

Does anybody know step by step how this is done (without using vocal remover software)? I know that it doesn't completely remove the vocals and it works better with some tracks than others but I'm just curious what the technique is. I know that it involves flipping the phase on one channel in order to cancel out the center material. Any help would be appreciated.

[analogcat]

Take the two negative wires running to the two stereo speakers and instead of hooking them to the amp, hook them to each other. That will cancel out the center.

I'm sure someone will provide a better answer....

[greatmagnet]

I really hope someone does...I've been damn curious about this myself and too embarrased to ask!

[JGriffin]

If you reverse the polarity of one of the channels (hit the phase-invert switch, or process that channel in your DAW), then center-pan both chennels so they combine into mono, anything that is panned center in the mix will cancel out and disappear. This is because signals that have equal amplitude in the original mix (center pan=equal amplitude left and right) are now of exactly opposite amplitude and the resulting amplitude when they combine is zero (negative two plus two equals zero).

[greatmagnet]

WOW!

[sonikbliss]

I found a way to do it in ProTools that worked fairly well.

Once you have your track in ProTools, bus the L & R signals to a single Mono aux track that is panned dead center. Flip the phase on the aux track (the EQ II and Trim plugins have phase switches on them). Raise the level on the aux (mixing it with the hard panned L & R signals) until you find the best result. It also helps to roll off the frequencies on the aux that you don't want to get cancelled out (such as the low end so you don't lose your bass and kick which is usually panned to the middle).

I tried it how dwlb said too and the results were pretty much the same but this way you can keep it in stereo.

[Shawn Simmons]

I've done this before. It works. It just takes a little bit of time to line everything up. I mixed a song with some vocals going in the breakdown. Weeks later, the band came back and said they wanted the vocals out of the breakdown. As I didn't think we could re-create the mix I tried to just take the vocals out of the breakdown by feeding an out-of-phase vocal back into the 2-mix. I had to play with moving the out-of-phase vocal around (sample by sample) but once I got it lined up, that vocal was GONE. I was amazed at how perfect it worked. It can be done.

shawn

[psychicoctopus]

I've done this before. It works. It just takes a little bit of time to line everything up. I mixed a song with some vocals going in the breakdown. Weeks later, the band came back and said they wanted the vocals out of the breakdown. As I didn't think we could re-create the mix I tried to just take the vocals out of the breakdown by feeding an out-of-phase vocal back into the 2-mix. I had to play with moving the out-of-phase vocal around (sample by sample) but once I got it lined up, that vocal was GONE. I was amazed at how perfect it worked. It can be done.

shawn

that's really cool.

[Zeppelin4Life]

yep, math doesnt lie, it works perfectly everytime. sometimes though, the song will be weird and NOT have vocals in the center. if this is the case I have no idea

[NeglectedFred]

The phase thing is the general idea, but in many cases you might risk loosing your bass track. There is a remedy for this, maybe someone knows a better way. This will especially be effective if it was mixed for surround sound. The center channel uses a dedicated frequency response, sorry, don't remember the exact range, but here is what I personally would do.

Take the song (we'll call this 'A'), and copy it to another stereo track (we'll call this 'B'), apply a low pass filter to 'B', I would probably start with 100Hz and below.

Take the left channel of 'A' and invert it, mix to a mono track.

Mix 'A' and 'B'.

If you really wanted to go crazy, and keep some of the high end stereo imaging, you could do a little research to find the exact band for the center channel, and do the same thing with a high pass filter above the center's band. It would probably save some snare, too.

Haven't tried this, but it makes sense to me.. Let me know if you have any luck should you try this method.

Good luck!

[JGriffin]

The phase thing is the general idea, but in many cases you might risk loosing your bass track.

Or, in fact, anything also panned center, which may or may not include snare, kick, and solos. Good point.

[Milkmansound]

now if there only a way to take out the music and keep the vocals. There has to be some mathmatic algorithm somewhere to put something out of phase and save what was cancelled out somewhere for future use.

[Mark Alan Miller]

now if there only a way to take out the music and keep the vocals. There has to be some mathmatic algorithm somewhere to put something out of phase and save what was cancelled out somewhere for future use.

Never tried it, but couldn't one perhaps take the original difference signal and phase-cancel that against a mono-summed version of the original signal?

Kinda like enoding a standard L/R signal into a MS matrix, and only using the resulting M channel....

I think the Waves S1 can be flipped into MS mode, and on L/R signals the result would be the M on one channel and the S on the other.

[aishabag23]

Racism is Gay

... does this mean 'racism is happy and carefree' or 'racism is homosexual'? guess you're a white hetero male thinkin' you're pretty cool and witty and funny.

[I'm Painting Again]

yep, math doesnt lie, it works perfectly everytime.

especially when you take the sqaure root of a negative number..