I like Dirac too. I have two systems with it (living room and 2 channel system). I have had the Dirac (NAD T758) in our theater room too. Switched back to Audyssey XT32 there. I'm limiting correction to about 500hz via the Audyssey xt32 app. I have dual subs (Rythmik L12) and I find Audyssey actually does really well with under 500hz. If a person is limiting correction from 300-500hz, I wonder if Dirac is actually better. I'm just not sure. If doing full correction, 20hz-20khz, I would think it probably is. However, I think it is worth the effort to eventually take a closer look comparing xt32 vs dirac (limited correction and also full correction). Maybe someday a more in depth look at the by
@amirm would be worth the effort (which would require a lot i would imagine).
It's interesting that you mention that you weren't sure if you could tell the difference under 500 Hz. I found that the bass region was actually where Dirac's benefits over Audyssey were most apparent in my experience. Dirac cleans up bass and makes it sound tight and have impact, while Audyssey sounded bloated and boomy to me in comparison. I suspect that this difference is due to Dirac's emphasis on timing and impulse response correction as opposed to Audyssey only correcting frequency response.
For that matter, exporting REW filters to match a target curve and correct frequency response using a miniDSP is something I've done before, and I've gotten the bass with it to sound as good as Audyssey, but not Dirac. Again, I think it's important to be careful not to get too fixated on only frequency response -- that's something I learned while working with Dirac for several years. If frequency response was the most important thing, you could take a single microphone measurement with Dirac, rather than taking several microphone measurements over a wide area of positions (as Dirac recommends), and the result of the single mic measurement Dirac calibration will definitely have smoother frequency response than the full calibration. But no one would say that a single mic measurement Dirac calibration sounds better than a properly done calibration taken with several mic positions over a wide area. Dirac performs best when you give it as much data as possible to give it sufficient information about your room, even though the resulting frequency response may not end up perfect.
It's important to note that these apparent differences in bass between Dirac and Audyssey were most noticeable to me in music, where a bass guitar may be constantly present throughout a song and have textures and layers of details to its sound. In movies, if there is an explosion occurring and my subwoofers are shaking my whole house, I am never thinking, I wish this explosion sounded more nuanced. Even a subwoofer that sounds boomy for music can sound just fine for LFE content in movies, as long as the subwoofer is not being pushed to its limits.
Still, optimizing for music always makes movies sound better too, since movies always contain soundtracks. But if movies are the focus rather than music then I think the sound quality of voices are a top priority, and this is another area where I have never achieved the level of clarity of voices I've gotten with Dirac while using Audyssey. This is also an area that showcases the benefits of running Dirac full-range. I'm aware of the position some have that EQ should be avoided above the room's Schroeder frequency, but human voice frequencies extend as high as 10 kHz (from 's's and 't's, for example). My speakers, Ascend Sierra RAAL Towers, were designed to sound neutral, but they sound slightly bright when placed in my room, so female voices can sound slightly harsh, and the sibiliance from these high frequencies can sound grating and unnatural. Customizing the Dirac target curve and giving it a rolloff in the high frequencies, designed specifically to what sounded correct in my room, was absolutely necessary to get male and female voices to sound crisp, clear, natural, and life-like.
Obviously these are just the experiences and opinion of one person, so folks are free to interpret them however they like. But unless a proper study with a large enough sample size is ever done, these personal experiences are about the best we can do for now.