I'm interested in hearing opinions (based on listening experiences) of the effectiveness of Clear Voice Detection in MPC. I just had a read of:

http://babelfish.altavista.com/babelfish/u...s.html&lp=de_en

and the feature seems quite an interesting one (and conceptually, not too hard to implement, assuming I'm reading it right). What I'm planning to do is introducing something similar into Vorbis but if it results in zilch improvement in quality, then I guess it's not really worth it.