• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Great series on Audessey failings and how to correct

DubbyMcDubs

Active Member
Joined
Jul 2, 2019
Messages
143
Likes
132
Just came across this new series on YouTube covering the issues with Audessey and ways to manually correct for them. Its a multi part series he is gradually releasing. He has done a fantastic job!

Part 1:

Part 2:

FR of various Audessey settings:

Updates/revisions to the techniques:
 
Last edited:

NoFuzz

Member
Joined
Sep 11, 2019
Messages
15
Likes
14
The first video is very thorough, quite a lot of information which is good of course if you want to know about the why's and how's.

I think if you "shortcut" what is the method here, it's exchanging the audyssey "reference curve" with the Harman curve.

There seems to be 2 ways of doing this: with an optimization for crossover (at 80 Hz and 120 Hz, found in the youtube description) and a "universal" curve, found at the link posted here (and same link in the pinned comment on that first youtube video), although the last one seems to be updated, thus "better"?

For me, using the univerersla Harman curve does the trick and gives me a more "cinematic"experience.
 
OP
D

DubbyMcDubs

Active Member
Joined
Jul 2, 2019
Messages
143
Likes
132
Part 3:

 

Steve Dallas

Major Contributor
Joined
May 28, 2020
Messages
1,217
Likes
2,926
Location
A Whole Other Country
Interesting. I like his methodology, but I can't help but wonder if he is overthinking some things, and I question some of the basic premises in the first video.

Disclaimer: I have a Denon 4700 with XT32, which may be different than previous versions and what he is using.

1. We need to care about the low and high frequencies being butchered by the Audyssey curve and therefore need to replace it with the Harman curve.

I know Audyssey presents 2 curves with 2 different high frequency roll-off profiles. That would seem to suggest much attention is paid to that area of the curve. But, I have not casually found Audyssey to use many taps up there. For example, in these plots, the orange line has correction limited to 1100Hz, and the blue line lets Audyssey correct the entire spectrum according to curve 1 (Reference). There is no appreciable, audible-to-an-adult-human difference in the rate of high frequency roll off between letting the speaker roll off naturally in the room and letting Audyssey correct above 10KHz. I certainly do not want Audyssey to try to make that region flat by boosting the tweeter there.

Audyssey to 1100Hz vs Audyssey Full.png


(These measurements were taken during initial setup, while I was experimenting with correction limit, crossover, sub phase, and sub level settings.)

Overall, Audyssey makes the FR above 1000Hz worse across the spectrum, but there is no reason to change the target curve in the HF range--especially since I do not let it correct my front 3 above some multiple of Schroeder below 2KHz. It also performs worse in bass frequencies when it is using HF taps. In my system, I limit Audyssey to some frequency that blends well with the front 3 speakers and let it have its way with the rest of the speakers.

Also, I am not sure about needing to adjust the target curve in the bass region to correct for crossover profile. I have found that simply adjusting crossover frequency, sub phase, and sub level is enough to achieve good results. I don't see how changing the stock curve to a steeper roll off in bass frequencies will help this plot, where the crossover is 80Hz, and the main problem is in the 200 to 350Hz region. It may actually hurt.

(This plot shows stock levels vs. +3dB bass tilt from 500Hz down added to the stock Audyssey curve using the curve editor in the MultiEQ app. The red line is stock with adjustments made to the sub phase and level to force an approximation of my target curve, which is the blue line. The orange line is the tilted curve. We are not interested in the orange line here; it is merely a bonus. I prefer the sound of the red line vs. the orange line.)

F206 Left Audyssey 1100Hz Limit +0 vs +3dB Tilt Sub+.png


The red line, which is what I actually use, looks and sounds very good to me. Forcing Audyssey to use something like the curve in the video will not improve that result in the low frequency octaves.


2. Audyssey boosts height speakers by 3dB.

I have not found this to be case. In this table, all my surrounds and tops have the same sensitivity rating (Polk RT series in-walls and in-ceilings). Audyssey set the levels differently according to distance from the MLP, and the levels look appropriate to me for balanced sound. More importantly, the system SOUNDS balanced.

Screenshot_20221113_100822_MultEQ.jpg



3. The Audyssey mic is inadequate and needs to be calibrated.

This may be true in some cases, but I have never found this to be true in practice with multiple Sound United receivers and processors. He specifically talks about bass frequencies, but as you can see from my corrected plots above, which are measured using a calibrated UMIK-1, the included mic is perfectly usable. I use it to take initial measurements, then use the UMIK and REW to make manual adjustments in the MultiEQ app until I am happy with the actual measured results.

(I seem to remember the SU guys saying the capsule in the Audyssey mic is the same as the capsule in the UMIK-1, and they test and select them for being within a certain accuracy range.)

--------------------------------------------------------------------------------

Anyway, this is all interesting information, and I applaud him for putting it out there. I just find I do not need to put that much effort into Audyssey optimization.
 
Last edited:

GalZohar

Senior Member
Joined
Aug 8, 2021
Messages
449
Likes
218
While I understand it's probably not a great idea to change the rolloff slope of the target curve, to my understanding it might be a better idea to set it to start rolling off at the same frequency as the crossover (which it often won't do by default). Have you had any experiences changing the rolloff to 80Hz (asuming 80Hz crossover) over using the default rolloff (often 40Hz or 60Hz rather than 80Hz). There were hints in some interview with Jeff from Audyssey that this might be a good idea. I wonder if anyone made measurements to show if and how much it improves the result, and whether that justifies the existence of the 200$ app (as this can't be done in the 20$ app).
 
Top Bottom