I have a Denon 4520CI and a Onkyo PR
Both have Audyssey multiEQ XT32.
In this post I'm going to compare two identically setup Audyssey XT32 calibration runs on the Onkyo PR
-SC5508 using two different Audyssey XT32 microphones. My Denon 4520CI is ready to go info repair (it's locking up), so all I can check is the two mics currently to compare the configs.
I ran two identical setups with just three simple placements on my main chair in the main listening position. The 1st mic position in the main seating position, middle of head placement, at ear height, the second placement on the right most seam of the same chair's back, the third placement on the left most seam of the same chair's back. I used a camera tripod to take the measurements points, and a mic stand to take the omnimic measurements in a pretty equivalent place to where the 1st position Audyssey capture point was.
to compare two different Audyssey XT32 microphones and see if they are different on the same processor, and to try to determine why I generally seem to like my Denon 4520CI's Audyssey calibrations better than the Onkyo PR
-SC5508 time after time in my subjective evaluations.
1) The Denon microphone (black in color but identical in shape and looks other than color to the Onkyo mic) - read both subwoofers about 2dB lower than the Onkyo mic (gray in color). 80dB on Onkyo, 78dB on Denon. No change in gain structure on the subs between runs.
2) The Onkyo mic set the crossover to 120hz on my back surrounds, while the Denon mic set the crossover to 150hz on my back surrounds. The rest of the speakers were set to the same crossover points.
3) Somehow - there is a half foot of measurement difference on most of the measurements - despite the mic being placed in the same spots for both calibrations. The Onkyo mic set my mains to 12.5' (L & R) the Denon Mic set my mains to 13' (L&R). I suspect this has less to do with the mics, and more to do with the general inaccuracy of Audyssey on repeated runs -- but none the less I'll record it here because it's unexpected - unless the distance is right on the middle of 12.5 and 13'. (since the Onkyo only sets to 1/2 foot increments in the auto calibration stage, but can be set to 1/5 foot increments post calibration manually.)
All measurements were taken with Dynamic EQ off.
FREQUENCY RESPONSE CAPTURES WITH OMNIMIC POST AUDYSSEY RUN:
I think it's pretty obvious why I like the Denon's setup better after capturing these measurements. I've always found the Onkyo's Audyssey calibration to be too bright --- too much treble. Well looking at the graphs here - that mic loads about 5dB more treble than the Denon. That alone explains my subjective evaluations of the two systems Audyssey calibrations. (since I've always used the 'correct' mic with each system.)
For these frequency response captures I was using Track 2 which is both left and right channel (and subwoofer) playing the sweep at the same time. The Omnimic was in the center position, very close to the intial position 1 of the Audyssey calibrations.
As a secondary note: I've always found that for whatever reason - Onkyo and now Denon's implementation of Audyssey always sets my crossover points too low too. I previously had JTR 228HT speakers ported to 70hz as the -3dB point and it would set the crossover to 50hz. I now have speakers ported to 90hz -3dB point, though they are loaded in a baffle wall, still the Denon and Onkyo both try to set the front speakers to 50hz. Look at the positive difference I get simply by setting the crossover to 90hz on my current speakers.
With the 90Hz crossover I can just turn up the subwoofer trim levels on the inuke amps about 6-10 dB using the input attenuator knobs and get the sound I like with a fairly smooth bass frequency response.
Yet another piece of evidence to conclude that there is a lot of variance, not only in auto EQ systems, but also in the mics that are responsible for capturing the readings, as well as each run potentially having variances that are somewhat unexplainable. (like the distance settings in my quick experiment).
Anyone else care to perform a similar test and compare data?