This has more to do with overloading the signal path than it being a 2 or 3 way design. Obviously - the mid/hi energy of the combined voice and music is being divided in a 3 way design - but with enough volume it too will sound just as distorted as the 2 way. It's just a matter of threshold.You can definitely tell. Most common situation that dj's will refer to is when you're playing at full dance volumes and you wanna get on the mic, talk over the music and say something, especially in a hyping up kinda way. A 2 way system will not make the voice sound as distinct and appear as if it's more of a muddied sound and less sharp in tone compared to a 3 way.
The way to solve this problem is to make "space" for the emcee's mic. You do this with a ducker acting specifically upon a notch corresponding to key frequencies where his voice resides. This keeps to total sum of the energy consistent despite the added source.
To the audience the overall volume remains constant, yet the new voice (much like the lead vocals) will stand out better in the mix. I'm presuming of course that the emcee is not talking over the vocals. Though even in that case, the voice will be heard clearly over any portion of the original vocal that would otherwise intrude.
While some of the latest high end DJ controllers/mixers have improved digital processing in the mic section this type of ducking might not be part of it. Where it is - it's a complicated adjustment that should be worked out well in advance of the gig for a specific mic and emcee.
A full spectrum ducker would work for anyone (what you hear at the airport for example.) When applied to a specific range however, the result would be different for a male versus female voice.
