"Official" Audyssey thread (FAQ in post #1) - Page 202

I hate to keep repeating myself, but it seems that the obvious escapes most everyone where these =loudness curves are concerned.

First, those curves are to 20Hz. I'm sure 1956 was a great year for playback hardware that offered a clean 20Hz at 130dB, never mind lower frequencies.

If you're a live venue sound mixer, the =loudness curves have value in properly adjusting the in-house FR of the PA. If you're a sound mixer of recorded music or movie sound tracks, the =loudness curves have value in adjusting the relative levels of various events by frequency.

If you're talking instead of reproduction of recorded source in a home environment, the =loudness curves are irrelevant.

A flat in-room FR is all that's required.

Bosso

Apparently, Dynamic Volume acts as a sort of compressor/expander to equalize the disparity in volume between a TV program and the commercials. It also is designed to act in concert with its other enhancement programs to attempt to make a poor playback system sound less annoying.

It's said to be based on terabytes of data involving people changing the volume by remote control, blah, blah.

The words 'TVs', 'commercials', 'HTIB' should be a dead giveaway that it's a mass market approach and has little to do with a capable playback system.

Dynamic EQ is a loudness button, on-the-fly. It distorts the FR for very low volume listening. Again, this is a mass market gadget aimed at poor playback systems. It's a variation on a theme that's as old as audio.

What I'm talking about is the fact that recorded sound has the =loudness curve built in, based on an average listening level of 75-85dB. If your listening level varies from that, adjustments may be made accordingly, which Mr. Holman is attempting to do for you through software.

One most certainly does not have to boost 20Hz by 40dB when watching a movie or listening to music. I think that should be painfully obvious, but the subject just keeps popping up.

Bosso

Originally Posted by audyssey

Hi JJ,

The correction resolution varies continuously as a function of frequency. It starts with the highest resolution in the low frequencies where most of the problems occur and then gradually relaxes as it gets to higher frequencies.

Chris

Originally Posted by a1sy

OK, I am on 61st page of this thread, and probably it will take me another week to read all. Meanwhile I would like to get some advice.

Originally Posted by audyssey

Hi Rob,

MultEQ uses fuzzy logic in the following way: Multiple room measurements are taken and then a set of pattern recognition rules is applied to group them into clusters. After each cluster is formed the membership functions are calculated (these assign a probability of membership to each measurement and how "well" it fits within each cluster) and then a representative response is created for each cluster. The benefit of this approach is that the representative responses in each cluster contain the appropriate weighting based on the similarity of the acoustical problems--the assumption being that more common problems are more serious. The representative responses are then clustered again to produce a single room representation for each speaker and then an FIR filter is created that inverts that weighted representation.

Coming up with this method took about 5 years of research at my lab in the USC Electrical Engineering Dept. What we found very early in this work was that a single position measurement is hopeless in providing enough information to a room correction system. There is a common misconception that if you have one listener (e.g. a mixer) then you should put the mic next to them. Our work showed that this may provide some smoothness in the high frequencies (some times), but often makes things worse in the low frequencies. So we focused on methods that allow us to combine multiple measurements and our algorithms are based on these findings.

Simple averaging of measurements can offer some improvement over the single position case, but can also get you into trouble for two reasons: (1) it is an RMS average of the magnitude responses and thus ignores any time domain effects and (2) it assumes, by definition, that each response is weighted the same when in fact certain problems are clearly more important than others.

After figuring out how to combine measurements, we worked on the measurement process itself. Pink noise and other stationary signals are limited to frequency response results only. They are basically time blind. So, we decided to focus on an impulse response method using a deterministic signal that has the same spectrum as pink noise, but is based on a time domain representation. Once the signal is collected from each measurement, an analysis is performed to separate the minimum phase components from the excess phase components and each are treated differently for the exact reasons you mention in your message. Mourjopoulos and others have shown that the contribution of the excess phase is directly related to the reverberant field and requires different methods from what the minimum phase components need.

Chris

PS. I have read your AES papers (and JJ's) and I think there is great overlap in our research interests.

Originally Posted by aaranddeeman

Thanks.

Thought, it would be helpful for folks getting their mic lined up correctly.

Originally Posted by a1sy

What's the proper way to got SW adjustment?
As far I undestand Audissey doesn't care how I did setup for the speakers.
So I finished Audissey, set up SW distance manually to 13 feet, changed Front and Center speakers to small, crossover to 80Hz for the front. Should I go and change SW level from "-12db" to "0" ?

Originally Posted by pepar

Audyssey is not measuring distance, it is measuring elapsed time including any delays from processing. Does your sub have some sort of filtering?

Originally Posted by bgarner

Well, I purchased the Denon 4308ci a few weeks ago, ran the Audyssey setup and after running it a few things, my first experience is that I am not impressed. I had a Pioneer 56txi previously that was calibrated using Sencore equipment (RTA's and stuff) and the difference is almost night and day.

I will be spending the next few days calibrating my Audio with my Sencore equipment, but I must admit, with all the talk about Audyssey, I thought the results would of been a bit better. Maybe I was expecting too much.

Also, I noticed that I can't EQ my Sub. I also believe that there is a way to use the pro version of Audyssey with the 4308? Anyone know how this can be done? I would love to get rid of my parametric EQ in my system and have everything running from the Denon.

Any help is appreciated.

Brian

Originally Posted by a1sy

Well, when Audissey cal's finished I can review the measurements.
It does show distance to all my speakers - and show it very accurate.
Since I didn't enter any distance manually, how come I got distance right for all 5 speakers (besides SW). So it does measure it.
Regarding SW - I don't know if it has any filter inside. Paradigm 12 is 6 year old SW and I don't have any reference material for it.

Originally Posted by pepar

The Paradigm website says it has a "Subwoofer Cut-Off" that is Variable 50 Hz - 150 Hz. Where do you have that set?

Originally Posted by a1sy

Per 5308ci manual SW Crossover should be set to Max and Sensitivity to 1/2 scale. On SW it's called Crossover Dial.
Tried that - getting 24 feet distance with level check for SW at "-12db".

Originally Posted by pepar

Yep, that is the correct setup. Nothing alarming there with the "distance" measurement in and of itself. How does it sound?

Originally Posted by a1sy

Here are alarming list
1. SW distance is wrong
2. SW level shows "-12db"
3. Front and Center Speaker set to 40Hz crossover as Large, when Freq is 44-22khz.
4. SW almost not noticible.
When I change crossover to 80Hz have a lot of SW.
I just want to undestand what the right way to acompish cal's and make sure it's right.

Originally Posted by aaranddeeman

Thanks Tocaje..

So if I understand it right, what you have done is first ran Audyssey so as to get sub trim as "0 db" on Denon. Marked this position on the sub for its volume knob. Then you turned back (to lower side) this physical volume knob on the sub to your liking.

Could you please confirm this? I am about to re-run the Audyssey and would like to follow the steps towards "Audio Heaven"

Originally Posted by EricScott

Curious as well. I re-ran Auto Setup on my 1909 to get the sub as close to 0 as possible. Then I thought people said the best way to reduce the bass is to manually adjust the sub trim through the Denon (which you can set separately for each input). However, whenever I get really boomy bass, even tweaking the SW trim to -12.0 db (the minimum) doesn't seem to help much (if at all). So I settled on basically leaving the SW at -4.0 (vs. the Auto Setup level of +1.0), which seemed somewhere in the middle of the range. But it definitely hasn't fixed the boomy bass.

Otherwise I love the sound - amazing surround sound details.

Originally Posted by aaranddeeman

Yes. You do not need the boom.

The steps I mentioned are mainly useful for the couch/sofa seating. Because it's almost impossible to keep the tripod on couch and make it leveled at exact ear position (not height-wise, but horizontally). It invariably be in the middel of the seat.

Originally Posted by tocaje

I tried the -12db thing and it was still too much bass. The solution, as posted above, was to grab hold of the sub's volume knob and physically turn it down (I marked it though, just in case.)

Originally Posted by saprano

Also can somebody give a link to a simple tripod from radio shack or best buy.

Originally Posted by pepar

OK, I should have said "nothing alarming to me."

Read the "document" I linked here.

Originally Posted by a1sy

Ok, I will read this, it looks like a lot of info there.
I am still in page 68 of this thread, you gave me link to page 191

Originally Posted by jj_0001

So, your filters are, effectively, shorter at high frequencies.


How is this different that the Vista (tm) room correction system, I wonder.

(you knew I was going to say that, didn't you?)

Originally Posted by craig john

Chris,

In another thread on the Subwoofer Forum, there is a discussion about the "Equal Loudness Curve(s)). One poster stated that the Equal Loudness Curves are "irrelevant" in sound reproduction:

Do DV and DEQ use the Fletcher/Munson Curve, the Robinson-Dadson Curve, ISO Standard 226 or some other "in-house" Equal Loudness Curve to compensate for volume and FR? If it's an "in-house" curve, how did you determine the curve you use?

Originally Posted by a1sy

Well, when Audissey cal's finished I can review the measurements.
It does show distance to all my speakers - and show it very accurate.
Since I didn't enter any distance manually, how come I got distance right for all 5 speakers (besides SW). So it does measure it.
Regarding SW - I don't know if it has any filter inside. Paradigm 12 is 6 year old SW and I don't have any reference material for it.
Thank you

Originally Posted by n737nc

Chris,

I'm having a similar problem as sterryo, except I'm using HSU MBM and HSU HO Turbo.

MBM nearfield, and HO farfield. Any more advice on running Audyssey with a multiple sub placement like this?

Nick

Originally Posted by bgarner

Also, I noticed that I can't EQ my Sub. I also believe that there is a way to use the pro version of Audyssey with the 4308? Anyone know how this can be done? I would love to get rid of my parametric EQ in my system and have everything running from the Denon.

Any help is appreciated.

Brian