Wow, 4 AVRs in the room simultaneously? I might have had 2 AVR and 1 AVP at the same time for a short period and I thought that was crazy.
It is so interesting that you actually could compare the effects on "sound quality" of 4 different REQ on 4 different AVRs under the same conditions (assuming...) in the same room, so I would like to give it shot and comment as follow.
First of all, I don't think anyone can tell why you are having such bad experience with Audyssey, relative to the others, other than the simple answer could be that Audyssey XT32 Sub EQ HT is just as bad as you think it is, or there could be other reasons or factors such as any of, or combinations of the following that I can think of:
a) Your did not follow instructions to the letter. (I guess that's not likely based on what you are saying, e.g. "no matter how I adjust....")
b) Something went wrong with you specific hardware.
c) Many forum chats might have given you the preconception that Dirac and AARC are superior.
d) You just don't like the bass being flattened (relatively speaking) in the low bass region, that Audyssey typically would do by default.
My guess is that c) and d) are the likely reasons, at least to a large extent. d) can be alleviated by using the $20 Mult EQ editor App to customize your target curve for the 20 to say 100 Hz range so that you can have a tilt in that range similar to the default Dirac or AARC target curve. It seems that most people would not like what Audyssey did in the low bass range, that it would make their system sounding "thin, bright.." etc. The MultEQ editor app also allow you to defeat the MRC, and limit the EQ range to below the room transition frequency that could be between 200 to 300 Hz in a small to medium size rooms. Dirac Live and AARC don't have the MRC thing, and both allows you to limit the EQ range without any add ons.
If c) is the main reason then there is no fix, once you expect it to be worst then it will be worst. Unless you can do a real apples to apples AB instant comparison following the strict protocol for conduct a double blind listening test session and if such a DBT did change your view, then your bias might be fixed/removed.
Such DBT listening comparison would be extremely difficult, if even possible for most people to do in their room.
Did you find the Denon the worst too, in terms of "sound quality" that you perceived in direct mode (that is no EQ no nothing..)? Good thing you have 4, and you only have to like 1 of them lol.. The AVR-X4300H is still quite up to date so it should be easy to get rid of it.