Thanks for the 'review'! :)
No low-end freqs was redirected during the upmix.
I don't know the technicalities of active matrix decoding, so I don't know if it favors some frequencies 'by default', I mean maybe it puts more of the 'voice freqs' in the center channel, making them seem louder?
One question though: Did you listen to the 2.0 as stereo or did you run it through a hardware dpl/dpl2/neo6/etc decoder?