Audyssey results - why do they look so poor?

[sting004]

I'm running a 2.1 setup right now with two KEF Q350's on bookshelf stands. They are roughly at ear level, about 7-8 ft from the listening position.

Room is rectangular, no sound treatment. Speakers are about 7 inches from the wall, toed in to the listening position. No port plugs.

I have run Audyssey multiple times but seem to end up with a similar result as the one here. My expectation was the after being much more of a "flat" response.

What am I missing / not understanding here? Clearly I'm out to lunch.

 

[raindance]

Looks pretty good to me. Audyssey aims for a downwards slope.

 

[peng]

I'm running a 2.1 setup right now with two KEF Q350's on bookshelf stands. They are roughly at ear level, about 7-8 ft from the listening position.

Room is rectangular, no sound treatment. Speakers are about 7 inches from the wall, toed in to the listening position. No port plugs.

I have run Audyssey multiple times but seem to end up with a similar result as the one here. My expectation was the after being much more of a "flat" response.

What am I missing / not understanding here? Clearly I'm out to lunch.

It doesn't look too bad actually, especially if you have the XT version. Is it XT or XT32 that you have? Audyssey can do very good in flattening the peaks with cuts but for the big dips due to room mode, there is only so much it can do. Regardless, you can probably do better if you try the following:

- Follow instructions exactly, to the letter.
- Pull the speakers out to say at 18 inches, or at least 12 inches.

You are running 2.1 so what happens to the subwoofer's graph.

The "after" curve is not the actual FR curve, but one that is predicted by the Audyssey app based on the measurements and the filters created.

 

[flipflop]

What am I missing / not understanding here?

The most important thing that you are not understanding is unrelated to your question: by equalizing the in-room response above the transition frequency, Audyssey is also equalizing the direct sound. In this case, it's creating a massive resonance in the treble, likely degrading the sound quality.
You should only let Audyssey correct frequencies below the transition frequency of your room, or better yet, do the EQ manually.

 

[sting004]

It doesn't look too bad actually, especially if you have the XT version. Is it XT or XT32 that you have? Audyssey can do very good in flattening the peaks with cuts but for the big dips due to room mode, there is only so much it can do. Regardless, you can probably do better if you try the following:

- Follow instructions exactly, to the letter.
- Pull the speakers out to say at 18 inches, or at least 12 inches.

You are running 2.1 so what happens to the subwoofer's graph.

The "after" curve is not the actual FR curve, but one that is predicted by the Audyssey app based on the measurements and the filters created.

I actually just have the base MultiEQ through the Denon AVR-s750h.

I had figured that pulling them out may help a bit but I don't think it's possibly with our room setup.

 

[tecnogadget]

The most important thing that you are not understanding is unrelated to your question: by equalizing the in-room response above the transition frequency, Audyssey is also equalizing the direct sound. In this case, it's creating a massive resonance in the treble, likely degrading the sound quality.
You should only let Audyssey correct frequencies below the transition frequency of your room, or better yet, do the EQ manually.

Ignoring the fact that the AFTER graph is meaningless (less reliable than a true REW measurement) and mostly a simulation...what are you talking about? I can't spot any "massive resonance" on the midrange or treble. The down slope is actually very smooth above transition frequency.

 

[flipflop]

Ignoring the fact that the AFTER graph is meaningless (less reliable than a true REW measurement) and mostly a simulation...what are you talking about? I can't spot any "massive resonance" on the midrange or treble. The down slope is actually very smooth above transition frequency.

The graphs are showing the in-room measurements, not the direct sound, which is where you are going to find the resonance.
The treble in the pre-EQ measurement is down by 1-4 dB up to 10 kHz. Audyssey then fully corrects it in the post-EQ measurement.
Here's the spinoroma of the KEF Q350:

With the EQ by Audyssey, the on-axis (blue) changes to something that approximates the curve I've drawn in black.

 

[peng]

I actually just have the base MultiEQ through the Denon AVR-s750h.

I had figured that pulling them out may help a bit but I don't think it's possibly with our room setup.

If it is just the bas MultEQ version, then I think you have done a very good job already. Even the base version has higher resolution for the subwoofer channel EQ so you may see a smoother curve.

Two things you can try and see if you like the result better, I bet you will.

1) As suggested, since you have the app, limit the EQ freq range up to about 300 Hz.
2) Adjust the target curve (draw it with a touch pen, or finger) such that it would gradually boost the low frequencies.

Something like the following will do:

Audyssey Room EQ Review

This is a review and detailed measurements of the Audyssey Room Equalization. Audyssey is standard technology in a number of Audio/Video Receivers such as the Denon AVR-X3600H I recently reviewed. The standard version in the AVR is mostly all or nothing so I downloaded the Marantz/Denon...

www.audiosciencereview.com

 

[GalZohar]

Audyssey XT32 predicts a pretty flat response, but doesn't achieve it in reality according to measurements shown here (and also according to common sense).
Seems like you have Audyssey XT, which is very limited at the bass frequencies of the speakers (and a somewhat limited for the subwoofer), and does poor corrections for the high frequcies that you should probably just disable by limiting the correction frequency by using the app.

While you should verify with REW and not rely on the "after" curve, you can see that it somewhat reduced your extra low frequency energy but couldn't really do much with the local peaks (XT limitation compared to XT32), and applied many filters to the high frequencies which might look good on the graph but in practice may sound bad, so you should verify it with a listening test comparing full-range correction and just low-frequency correction (max correction frequencies that may be worth trying. according to your results. are 250-700). If you like the full-range correction (unlikely), then you can play with the curve like disabling the mid-range compensation (which seems like your speakers don't need) or changing the high-frequency roll-off. But if you limit the correction frequency then all those high-frequency curve changes will affect nothing.

Note that with correction on your bass will be a bit light. If you use Dynamic EQ that should usually help fix this. Set its offset to 0 (default) for movies, and for other content you should experiment although offset of 10 seems to mostly work fine for me. If it's not enough bass for you (it might take some getting used to and making sure you use the appropriate Dynamic EQ settings), then you can try the above suggested changes to the curve, although it would be tricky to find a curve that will fit all content for all volume levels.

Also note that if you use the app, you should only use the "reference" curve, as the "flat" curve ignores anything you've done in the app (it's always flat curve and full-range correction).