I have recently moved to the North West on the Wirral, Greasby to be precise any local members around.
 
I need some help. I'm not an experienced Sonar user more a cameraman, I recently videoed a friend singing in cabaret, he like most of are not the people we were years ago. I need to extract his voice from my master audio alter some of the keys in Melodyne then reintroduce the vocals back to the backing track. What is a good acapella app to extract the vocals. Any local guys would be a bonus. Or is there anyone that could do the job for me. I have 23 tracks to edit. For obvious reason mainly cost I'd rather do it myself.

I don't really think there is any program that can completely rip out the vocals from a mono or stereo track, so you can be able to work on it as a separate track. 

Isolating vocals from a full band recording is at best hit and miss because other things will be occupying the same frequency band as the vocals so gets isolated along with them. Pitch-changing will also pitch-change any of the band that's still in the "isolated" vocals and when put back together the end result might not be good. It's a little easier to remove and throw away vocals than it is to keep the vocals and remove the rest, but not much.
 
Audacity (a useful and free application) might be able to do what you want. At least, do it as well as anything else. Audacity's help on removing and isolating vocals is here - http://manual.audacityteam.org/man/tutorial_vocal_removal_and_isolation.html
 
Software aimed at the forensic acoustics side of things may also be useful, but it tends to be expensive and not much better than anything else if you want to preserve the tone and pitch of the voice rather than just increase the legibility of speech in a recording with a lot of background noise.
 
In Sonar I'd suggest isolating the vocal frequencies by trial and error and seeing how much of the rest of the track still persists. If the vocals are panned dead centre then a combination of mid/side processing to isolate the centre followed by eq to try to remove as much of what remains other than vocals would probably be the way to go. Sonar's Channel Tools can convert stereo in to mid/side and back again.
 
If the recording is effectively mono, things get a lot more difficult.

I agree its a crap shoot at best. Vocal removing programs use phase manipulation.
What was the format recorded in ? Video obviously but what audio ? Stereo, mono ?

One little trick I can think of is get your friend to re sing the part in key and layer it over the stereo recording. Mixed just right it might mask the original off key parts. 
Or just re sing each little off key part and just overlayer those little short bursts. 
 
As said the software that may or may not fix this is going to be much more than your going to want to pay.