Audyssey MultEQ Test Report for Denon AVR-3806

[bpape]

Looks like it did a pretty decent job above the subwoofer range. Hard to tell for sure when it's got smoothing applied. I did notice that wherever the 'average' response had a dip before, it now has a peak and vice-versa. Also, it did almost nothing to the large null around 50Hz - maybe 1db.

However, if one looks at the post EQ responses, they're still similarly ragged and still vary considerably from seat to seat. Graphs are one thing - especially when averaged. But the listener sitting at each position isn't going to hear the average, they're going to hear the post EQ'd response at their seating position. Look at the post EQ'd response vs. the pre EQ'd respone at each seat to really see what it's doing.

I'd also have liked to have seen waterfalls before and after. In theory, if this is really going to work as advertised and EQ that considers both frequency and time is going to be able to reduce modal ringing, then the decay times at the modal frequencies should have also gone down accordingly. No way to tell from what I see there.

Lastly, if I may ask, why were the mics placed so low? Looks like they were basically on the seat for all practical purposes. Not only does that mask the usual height related modal issues, but it also places the mic 'behind' the back of the seating. That thick seating can absolutely help knock down the wave off the rear wall that generally causes one of the worst modal nulls in most rooms. Granted, they were the same for pre and post EQ so at least they were all masked equally for both parts of the test. It just would have been nice to see it attempt to really deal with a real room's issues. When was the last time you saw an unEQ'd room response only have 1 large null - especially when the speakers were large and the subs were in suboptimal (pardon the pun) position.

Please don't take this wrong. I'll admit that this is a large step in the right direction in comparison to more conventional EQ systems and there are absolutely some apparent benefits - especially for something that can be built into a $2k receiver. It appears to me to be superior to some of the other systems built into other brands' receivers. I'm just trying to see what it's really doing where listeners are sitting - both in terms of frequency and decay times.

Bryan

 

[Nick250]

I thought after all the high tech posts, I would write my brief low tech subjective observations of Audyssey and low frequencies using my Denon 3806.

You can see my equipment below, but I have been using a Rocket UFW-10 recently and my comments apply to that configuration.

My observations are, that the UFW-10 better integrates into the system after Auto EQ has run. I can tell the low frequencies are there, but they don't draw attention to themselves. I suspect it's the Auto EQing of all the speakers together as a package that plays a major roll in SQ improvement in my listening room. This of course may be the bleedin obvious to most, but there it is.

All that being said, I do not think Audyssey is a cure all for room acoustics issues. I think that proper speaker and sub placement and room treatments come first, if possible (I have bass traps on order as we speak). However, I see Audyssey Auto EQ is another solid tool in my sound quality tool kit.

Nick

 

[audyssey]

...Perhaps Chris can elaborate as to why all before/after graphs show little to no change below 80hz.
-Brian

Hi Brian,

I don't quite understand your statement about "all before/after graphs". I think that may be a slight overstatement. In any case, I have been pretty open about the resolution of MultEQ XT in the low frequencies. Given infinite DSP resources, or numerous DSP chips, anyone can correct with arbitrary precision. The trick is to do it in a way that allows the numerous other DSP algorithms that come with today's receivers to fit on one chip. To do that, we have had to work very hard to find ways to perform FIR filtering that goes beyond what is found in DSP textbooks.

MultEQ XT can correct most problems below 80 Hz and runs into limitations when these problems are very narrow. The benefit of correcting extremely narrow problems has been debated in the audio literature for years and there are trade-offs to consider. This is especially true when dealing with parametric EQ solutions that attempt to fix the magnitude response, but do untold things to the time domain.

Furthermore, many of "all the plots" out there are taken under unknown conditions. There are numerous conclusions drawn from an uncalibrated mic measurement in one position in the room showing "before" and "after". We have spent a considerable amount of time trying to not only find computationally-efficient ways to solve these problems so that they can appear in consumer products, but also methods of measuring room acoustics that truly represent what is audible and how it is weighted from seat-to-seat.

There is no doubt that throwing more computing power at this problem will have significant benefits. A well-designed, dedicated, sub equalizer that does not cause time domain problems and captures spatial information from the room will definitely enhance the subwoofer performance. No argument there.

Regards,
Chris