I was very curious about your last post, so thank you for the update. I have been through with two Audyssey XT and one Audyssey XT32 and found the best results were obtained by following Chris's (Audyssey website) instructions to the letter. He also recommended not to use Audyssey flat unless in a small and not very reflective room.
By the way, I take exception to the Marantz warm sound beliefs, sorry, no such thing and I am quite sure it can be proved easily in simple level matched blind tests using pure direct mode. Sound signatures such as warm and bright (however loosely defined) have existed years ago, but audio amplifiers of the popular classes especially AB, are a highly mature product now. Once you get pass the entry levels, they are basically transparent, with virtually flat freq response. I have used multiple Marantz and Denon products and a few other brands so I am speaking not only in theory but by experience as well. If anyone insist on believing in sound signatures of well designed amps that have been verified time after time to be transparent to human hearing, I respect their opinions but just want to point out it is not to be taken as facts, that's all.
I think the reference setting might not work with certain types of speakers either, notably speakers with wide, uniform dispersion. The whole idea behind the target curve on the reference setting is the assumption that your speaker's off axis response follows a downard sloping curve beginning at about 4khz-8khz. Since a majority of the reflected sound is significantly off axis, most of the spectral content of those reflections will be limited to below 4-8khz.
Our brains are excellent at sorting out the difference of the direct sound in reflective rooms from those reflections, but a microphone simply measures a combined response of both the reflections and the direct sound from the speakers. Eqing it flat would actually just be adding a massive treble boost to the direct sound from the speaker, which would sound awful.
Speakers with waveguides and well controlled HF directivity are likely to have a much more equal amount of energy across the high end off axis, so the spectral content of the reflected sound should be nearly identical in response to the direct sound, only lower in level.
I don't know what would qualify as a small room, but in a 2000 cu ft room seated 11' from the speakers, with Klipsch RP series, which have a 90x90 dispersion pattern from 1khz-14khz, the reference curve does nothing but make the speakers sound like somebody put a blanket over them. This was initially in an untreated, highly reflective room with hardwood floors too.
You could make an argument that films are mixed to the xcurve, and therefore a HF rolloff is appropriate for proper playback the way it was heard in the studio, but almost all movies are remixed for bluray in a small room with either no xcurve or a modified xcurve.
Equal EQ target curves and equal spl calibration levels do not translate the same in a small room as they do in a large theater. Modifications must be made in order to maintain the same spectral balance and dynamic range that is heard in a large theatrical dub stage when playing back in a small room. This is one reason why studios remix blurays in a 3000-5000 cu ft room. If the theatrical mix was simply plopped unaltered onto a bluray, you wouldn't be able to understand the dialogue without constantly riding the volume knob or blowing yourself out of the room.
For rooms less than 10k cubic feet, dolby recommends a modified xcurve, which rolls off 1.5dB/octave beginning at 2khz. If you have a denon avr, their cinema filter does exactly this, so the use of audyssey flat along with the cinema filter is much more appropriate for movies than audyssey's curve, which starts at 4khz and rolls off a bit faster.
The other issue i have with both audyssey and thx, is they both calibrate to 85dB with -20dBfs tones. If you read through dolby's recommendations for spl calibration in small rooms and home theaters, they specify a calibrated level of 79-82dB, nearly all bluray home mixes are remixed at this lower level too, because 85dB sounds much different in a small room vs a large theater. I suspect this is one of the reasons many people find 0dB too loud at home, because you're actually playing the movie about 6dB above what it was actually mixed at.
If you want a "reference" presentation that matches the mix heard in the studio at home, id suggest lowering audyssey's trim levels 3-6dB and utilizing cinema filter with the flat setting.
