Remove Vocals from mp3 with Sonar?

Most lead vocals are in the middle of the stereo field (panned center). You can use a "mid/side" tool like Channel Tools to turn down the middle while retaining the left/right edges. This is generally how vocals are removed from premixed material (and is all Audacity does AFAIK).
 
So you'd load the mp3 into an empty track in Sonar (this will convert it to a wave), insert Channel Tools in the track's FX bin, in Channel Tools click the "Mid" button (to put it in Mid/Side mode) then turn the "Mid Gain" dial all the way down.
 
If you own R-Mix it can do this type of thing as well but a little more accurately/graphically (figure out which blobs/areas the vocals are in, surround that area with the shape thingie and turn down the "Inside" level).
 
Of course nothing is guaranteed to completely remove specific sounds from already mixed material.
 
BTW... if using Channel Tools (or other stereo tools) make sure your track ouptut/interleave is set to stereo.
 
Cheers.

Thx for the quick reply.  I've attempted to do it with R-Mix but there is some instrumentation close to the middle and to get the vocal out I'm having to remove too much of the instrumentation.  Here are the audacity instructions which may end up being the same: http://www.howtogeek.com/56335/how-to-remove-vocals-from-music-tracks-using-audacity/
 
Mike F.

Unfortunately when removing individual parts of a fully mixed file you are also going to yank out sounds that occupy the same pan postiion and (when using things like R-Mix) frequency range. Usually the bass and snare are panned center as well so they get affected. With bass it's a little easier to avoid with something like R-Mix because you can leave the lower freqs out of the removal. Snare is a little harder because a lot of the crack will likely be around where your vocals are.
 
What you can try is automating the "Inside" level in R-Mix so it only removes the target sound as it occurs. In this case automate the fader to only turn down when the there is singing. It will of course still yank out competing frequencies at those points but might be a little better.
 
To compensate (if you want to get really crazy) you could use synths/samplers/MIDI to match the original parts as best as possible and blend them in to the full mix. That way when R-Mix yanks those sounds out it gets masked a little better by the synth sounds.
 
eg: Dial in a snare sound that closely matches the snare in the song and program it in to match the original performance. Then when the removal happens you don't lose your snare. Might want to automate the volume of that too so it gets turned up exactly when the removal happens. Drum Replacer would actually be a good tool to do this type of thing.
 
 
However... as noted, this type of thing rarely works perfectly. It's just the nature of mixed audio. Once it's mixed down and all smooshed together it's next to impossible to completely isolate. It's like trying to pull the eggs out of a cake after it's been baked. lol

Yanno... and you may already be aware of this, if you want to REALLY learn about this stuff I'd recommend googling for information about producing Karaoke mixes professionally. Firebird brought up Soundforge which I've seen referenced when perusing the subject but there are of course many ways to skin an audio feline. The pro Karaoke production places get some amazing results and that's really what you seem to be going for. I'm guessing it's a combo of programs/techniques.
 
Cheers.