Um...put the compressor on the track to be ducked, route the vocal to the key (sidechain?) input on the compressor. The threshold control now sets the trigger level coming from the vocal, attack and release are in relation to the vocal, everything else affects the ducked track.

I hate "ducking", automated or otherwise.

Get a nice level to the music, a nice level to the vocals and maybe a bit of(And I mean a BIT of!) compression "overall", for a much better "effect"...

Listen to the mix at v-e-r-y low volume to make sure you still get all the words(Or all of whatever you don't want to get "lost") - idea is to turn the Vu down so low that almost all you DO hear is the most "neccessary" part of the mix(In the case of a commercial, for instance the "sales pitch") and that no "words" get "covered" by a "brass stab" or something... The vocals NEED to "sit" in the mix with NO conscious "ducking"!

Leave the massive, horribly noticable level changes to the novice disk jockeys...

TG

Ducking is really best for doing voiceovers for commercials and such.

appreciate the input, have been lowering the gain of instruments that mask the vocals, in little bits and pieces, before and after the words, using fades to "sometimes", help distinguish the english