I've been in that room when I visited and those "main" speakers have subwoofers connected to them. The "jacked up" bass is based on the natural response within the room.
While that may be true, it's out of the scope of my point - he was adjusting the main channel
well below the presumptive bass management. So basically either that twiddling is either going to be EQ'ed out to fit the highpass (i.e. pointless) or it's going to fight with the AVR bass management rolloff and affect the in-room bass response who-knows-how (i.e. terrible).
This would not be a problem if bass management were properly integrated into the room correction. Then one could draw a target curve for the whole channel.
Alternately, is it set up as "double bass?" If so, that's even worse for the combined response! Audyssey will EQ the main to whatever target curve and separately EQ the subwoofer(s) to whatever subwoofer target curve is chosen for the subs, and likely do a great job with both independent parts. But then the mains and subs will interact, and the end result will be nothing like whatever target curve was chosen.
A lot of people use something similar to a Harman target which has a 9dB-10dB bass rise.
The current Harman target is actually more like an ~8dB total tilt with a ~3dB rise to 20Hz (re 1kHz), but yes there are older ones with more of a "ski jump" profile that may concentrate more rise in the bass.
However, here my point is that because you mess with integrated parts separately you're apt to introduce interactions at cross purposes with whatever you want to do. So what you draw isn't likely to be what you get through the mains/subs crossover region.
If you really want to be sure there's a blend, that would require a measurement after the filters are set and make sure the speakers align at the crossover points.
I'm well aware of that because I spent a long and tedious afternoon figuring it out...it doesn't always work the way you'd expect and in the end it takes
a lot of trial and error to get something as simple as a constant tilt to stitch correctly between bass and bass-managed fronts. Maybe the new app improves on that, as there's a tilt target built in?