Just a quick question...

  • Thread starter Thread starter TelePaul
  • Start date Start date
TelePaul

TelePaul

J to the R O C
Was talking t a mate of mine about this...he says once a song has been mastered, it can still be broken down into its component parts and have the levels adjusted....we're not talking about adding an EQ filter, i mean removing an aspect of the sog without the 'master tapes'. Thanks guys.
 
I've heard there's a program that does it by recognising the frequency range o each instrument and separating them into individual tracks, but I've never seen/used/know what it's called, and I'm a little sceptical myself...
 
that'd be pretty much impossible to do without truly destroying the music...
although different instruments may have different resonant frequencies, they still ring across a wide band of them...

to remove them would be to kill the frequency spectrum
 
That's exactly why I can't quite bring myself to believe that it can be done...
 
I feel the same, it was understanding that the music would be fundamnetally changed to the point where its a new song. Thanks guys.
 
Just recently came across the new iPod accessory from Griffin. The iKaraoke. It claims to remove the vocals from your songs so you can sing along. Not actually out yet, but it would be interesting to see if it really works well.

You can mark me down as a skeptic.

http://www.griffintechnology.com/
 
RAK said:
Just recently came across the new iPod accessory from Griffin. The iKaraoke. It claims to remove the vocals from your songs so you can sing along.
I would imagine that ill simply work the same way as other vocal removers; it probably assumes that the vocal is down the center of the mix. It then attacks the vocal by doing some MS-style matrixing to synthesize a "center channel" out of the stereo signal, and then mutes the center channel either completly or partially. It then re-matrixes the signal back to normal stereo with the center pan information muted.

Dissassembling a mix into it's component tracks can be done very roughly via some heavy-duty analysis software, but the information loss in the resulting individual tracks is heavy; if you have only two or three clean tracks (say bass, vocal and piano recorded and mixed fairly dry) from which the mix was made and if the software is good (NSA-intelligence level ;) ), you can get some listenable individual tracks. Probably not good enough quality to remix with, but enough to get fairly clean seperation (after applying some intelligent filtering and modeling).

But start raising the number of component tracks, start adding some non-conventional sounds as some of those tracks, and start adding things like convolution verb to the two mix, and the quality of the results degrades quite rapidly from passible to useless.

This is all based upon labratory-level software for eventual use in stuff like intelligence gathering and information filtering - being able to pick out one conversation in a crowd, for example - and is just the non-classified stuff that they are allowed to talk about in science magazines and the like. As far as I know, there are no such commercial products available to the public though. Not that Iv'e heard, anyway.

G.
 
i always figured stereos and what have you with a karaoke option took a stereo mix and would make it out of phase in certain frequencies, where vocals reside, and kill them...

seems reasonable considering you'll sometimes get some vocals left, or kill the snare sound and other things like that a bit
 
Back
Top