how do i take an mp3 that has vocals and instrumentals and take out the vocals so its karaoke? i tried a winamp plugin that sounded nasty, so is there any program that will do this for me, like sound forge or something? thanks!

The AnalogX DSP plug-in for WinAmp works about as good as any tool for removing vocals. There is really no magic way to do this. It's a bit like trying to remove the eggs from the cake batter after they've been mixed. There are two channels (often called tracks) in a stereo MP3 or any other file or track. The vocals are not isolated to one of these. When the music is produced say from 24 tracks, the vocals are isolated to some of these tracks but once mixed to stereo, they are in the mix.

No hardware or software can distinguish exactly what is vocal content from instrumental content. Therefore, most vocal removers use the same basic concept; that the lead vocals are centered in the mix. The guitar may be to the left and the keyboards to the right, even the background vocals may be off to one side (channel) or the other, but the lead is almost always dead center. This means they are recorded or mixed equally to both channels. The vocal remover uses phase cancellation, where it put's one channel 100% out of phase. This results in everything centered be cancelled out.

Unfortunately there are other things in the center, usually the bass and some of the drum parts, and there are other parts close to the center. This will also be cancelled out or greatly reduced in volume. On some songs this is a bigger problem than others and on these songs the vocal remover will work reasonably well. It will almost never work 100% unless for example it's a trio with one part extreme left the other extreme right and the vocal in the middle (this will virtually never happen since parts are hardly ever placed to the extreme left or right).

You can adjust the level of removal, trading off how much vocal is removed with how much of the rest of the mix is removed. You can also try adjusting the EQ by raising the frequencies that are outside the vocal range, the highs and the lows. You can of course manually edit the wave, but this would be very labor intensive and could take hours to edit one minute of audio, without the results ever being satisfactory.