Just leave the Strings

[Arubin]

Hi, I'm trying to extract the strings section off of Damien Rice's Amie.
I'm using the Audacity 1.3B and a stereo wav file.
The track has 3 components, vocals, accoustic rhythm guitar and strings.
I've managed to remove vocals using the "vocal removal" effect.
Now there's the accoustic guitar left which plays throughout the track.
Any ideas how to remove it? or reduce it?
I actually need the strings track to play at my wedding.
I'm gonna play guitar and sing live, with the strings background on the track.
Any help appreciated.
If you'd like to help me remove it from the track, i could send the track to your email.
Thanks!

[PGA]

It is not possible to remove the "voice" of one instrument, or class of instruments, from a mixed track. Our brains are smart enough to work out which sounds come from which instruments. So far nobody has come up with a computer (hardware or software) that can do that - at least, not a price level that us consumers would tolerate. Unfortunately, you're trying to achieve the impossible. The only reason the vocal was removed successfully was because it had been placed absolute dead center on the stereo sound stage. If you listen carefully to your new copy you might be able to hear that some of the backing has gone too. Your only real hope is if the guitar is panned hard to one side and the strings panned hard to the other. But that would produce such an unnatural stereo effect you probably wouldn't have wanted to listen to the piece in the first place.

[Arubin]

Thanks for the reply PGA.
If it's not removable, any chance to reduce the volume for that component?
Maybe by adjusting the frequency?

[PGA]

The strings will be producing frequencies, some of which will lie in the same frequency range as the vocals. It is unlikely you will be able to manipulate any frequency band without having a detrimental effect on the rest of the recording. What you are up against is the fact that your brain is a damned sight smarter than any computer system. You can work out which parts of the sound come from which instrument (including the different human voices) - by tone, timbre, harmonics, etc. To the computer a given frequency is just a pattern of 0s and 1s. When there are several instruments all playing a note at the same frequency, we can distinguish each instrument. To the computer it is just a jumble of mixed frequencies which it sees as a string of as 0s and 1s. The computer doesn't know which 0 goes with a violin, which goes with a trumpet and which with a female vocalist. The only way to work with each separate instrument is to have access to the indivdual tracks that the recording engineer had access to in the studio.