I believe there are some products in this space. It might be made simpler by integrating dac/dsp and amplifier in the speakers (and the just stream lossless digital audio for playback) -- rather than try and adapt to traditional speakers. It might be the kind of thing that the radical commoditization and continued price drop in hw will make feasible.

Cinemas (and concert venues) do calibrate their sound, here's a sample link that touches on what's involved:

http://education.lenardaudio.com/en/17_cinema_7.html

I'm not sure a combined mic/speaker system makes much sense -- it would appear easier to just model what's important: how the sound sounds where the listener is -- which is typically anywhere other than on top of the speaker.

Model a human head w/ears, but mics in each -- and hook that up to a system that plays/records/compares audio. While dynamically adapting (noise-cancelling when there's construction next door, the garbage truck rolls by?) -- I'm not sure that makes much sense. We're used to acoustics changing a bit -- but having a half-decent base-line would probably be great.