Actually part of the problem is the insanely low dialog levels on cinema mixes in 5.1. Compared to tech specs for broadcast these levels may be 15-20 db lower than typical stereo programming.
They do this to allow for a wide dynamic range and tons of head room for big explosions. Apparent loudness is a very tricky thing to technically quantify, so it will not be easy to police.
Devices like the Dolby analysers also are fraught as they try to equate volume over time. A show with short sharp loud bits and then quiet will register as less loud that a show with a lot of regular but not excessive levels.