To expand on what others have said:
What essentially you will need to do to extract the instrumental music is remove the vocals via phase cancelation. Without going into the complex nature of phase relations in waves what is happening is you are sending an equal an opposite signal into the audio every moment it is being generated.
(so if say the wave had an amplitude of 1dBSPL you are at the same time generating an signal of -1dBSPL, the resulting level being 0)
What DTK suggests is to use the right side of the stereo audio track to cancel out the left. In actuality to doesn't cancel out the left completely as there will information on the left that won't be on the right. As Vocals are nearly universally panned to the center of a stereo field you will notice that this will have the greatest reduction (unfortunately so will the bass, which suffers the most out of the frequency range)
In order to do this in your computer you will need a DAW that has sample accurate timing. This is a MUST (Logic and Pro Tools has this, maybe Audacity too. I'm not sure about Garage band.)
There is another way but it's a bit more long winded if you have vocal track on its own:
1. Set up both your stereo file and vocal track. Make sure vocals in the stereo track start at the EXACT time the individual vocal track does (down to the sample scale of the audio, failure to do this will produce incomplete cancellation)
2. In your vocal channel invert the phase of the signal, most DAW have a dedicated software plug in for this but a lot of eq plugins have them too.
3. Play the track, you will now notice the vocals are gone.
4. Bounce.
If not and the piece is repetitive with small sections of instrumental (even if it's just a bar of instruments without vocals but is the same as the backing track when the singer is on) cut this out. do this as much as you can, and line the cut out instrumental in a separate channel (you don't to move the original audio, you can just make a copy) when you have as much as you can, invert the phase. This will leave you with a bare vocal track (or as much as possible) and with this you can use it again on the stereo file as outlined above.
This is really good when you want to do remixes of tracks, especially hip hop or pop where the beat and music is pretty much the same throughout.
Hope this helps.
