"Official" Audyssey thread (FAQ in post #1) - Page 1640

I seem to remember it being said that it was a false assumption that our hearing sensitivity drops faster behind us than in front of us. Hence, the thinking is, that lowering the surrounds slower than the fronts is not needed and distorts what the artist intended.

Personally, I don't know what to think as the algorithm for it was supposedly based on tests with film mixers manually moving the faders on the surrounds to where they thought they should be for each given volume level of the fronts.

Jeff

Correct.


Also correct. But when the mixers did their tests, I think it is safe to say that they were not listening to any loudness compensation processing (the variable EQ part of DEQ). Plus, as I mentioned in my post about Peter Tribeman, the effect that mixers heard as the volume is reduced (diminished surround effect) is real. However, had they done the test I posed, with equal loudness signals in every channel, the result would have been different, thus calling into question the conclusion about hearing sensitivity being different in the rear.

Phase shifts are integral to Butterworth filters and their cascaded Linkwitz-Riley form.

Thank you for that! If you have a THX receiver, do you think you could run the same test with THX Cinema & Loudness Plus?

Maybe your perception of loudness doesn't decrease faster in the back than it does in the front? That would make you human, not mutant (sorry, no X-Men plane ride for you).
That's too bad, since that leaves you without potentially useful aspects of DEQ (frequency response correction) because you don't like the surround boost. Don't make perfect the enemy of good. Consider trying some of the offset options before leaving DEQ off completely.

So DEQ would have needed to been applied .. dynamically .. to each surround depending on where the mixer placed the fader? (OK, I guess that's rhetorical as that is what you said; I needed to type it to fix it in my head.) Your thinking is, if that had been part of the test, the faders would have been positioned differently.

Gosh, doesn't that seem like such a simple thing, i.e. error?

Jeff

Pepar, ...please, what do you mean by "false assumption"? Assumption of whom? Why is this DEQ discussion going downhill without proper reasoning by ones having enough knowledge on the issue, or at least they are thinking they have? Why is it that denying patented solutions is no longer accepted in this thread coming up in form of bashing a product of a maker licencing it to the most reputed hardware makers know the world over? Where is this attitude gonna lead us to? I think to nowhere!

Just as BCJ reported.


You can think about the human hearing mechanism like a pair of microphones inside a funny enclosure. In fact they model it that way, it's called a dummy head. The outer ear, head, torso, all of that "funny enclosure" stuff alters the frequency response of the sound depending on the angle of incidence, and we call that alteration the HRTF, Head Related Transfer Function.

If we use an audio signal of a narrow spectrum, one that is not so strongly affected by the HRTF, and play it at, say, 75 dB SPL measured at the head from each of the 5/7 main speakers, and reduce it by x, y, or z dB, then the actual energy delivered to the "microphones" (the eardrums of the listener) for each x, y, and z step will be identical regardless of which speaker plays the sound. As a result, it will have the same perceived loudness from each speaker.

Now it turns out we don't have to set such narrow conditions on the audio signal to achieve these results, but I started by eliminating the HRTFs as a factor just to isolate a first principle: whatever sound or wherever it emanates, the same "mic" is used and it has the same nonlinear properties all the time. It cannot use one law for front and another for rear sounds.

If the sounds are materially affected by the HRTFs, or the loudspeakers are not identical in frequency response, then we can expect that those response differences will take increasing perceptual effect as the volume is dropped. That's where the variable EQ part of DEQ comes in to mitigate the effect. If it does so properly, why do we need overall gain shifting in the surrounds?

The false assumption is on Audyssey's part. The test they performed was not relevant to how they implemented DEQ.


I see no downhill trajectory.


Patents do not validate technology, they only protect it.


Our explorations are not bashing. If you fear knowledge, this may not be the place to hang out.

I'm sorry Feri, but I just love it how angry you get when people question Audyssey technologies

Roger was the one that made the assertion about "false assumption" and now we are discussing it. Nobody is bashing. I think it's a valid discussion to have. Frankly, I also dislike the surround boost of Dynamic EQ and personally use the offsets in my setup to tone it down a little bit. And that's with film.... with video games and music it's a major problem. Chris even tacitly admitted as much as he has made comments about a potential "game/music mode" for DEQ that has the loudness compensation without the surround boost.

Re surround level: I tend to listen to most everything at a pretty constant level (-20dBFS, except late at night) and I found long ago that simply lowering the surround channels 2dB made them not stand out again.

Honestly Roger. How long did it take you to google in on the subject while you could come up with abbreviations like HRTF and the like.

OK, I'm finished here on this thread. Have a nice day!

Feri, can I make you a margarita? I just squeezed some fresh lime juice.

"I remember it being said" is what I said. Personally, I don't know. I am only trying to learn.

I actually thought about doing that too. But for me, the problem is I would have a nagging in the back of my head that said "my system isn't properly calibrated" and then I would wonder if what I was hearing is correct or not, and I would be mentally taken out of the film or show......... I'm weird I know.

It's not the case on my side about questioning, but if you followed the discussion, its about the verdict that Audyssey DEQ is based on flawed premise. We all need to belive in something, eh? Now I believe I'm gonna have a beer, ...maybe two!

Some things translate in any language, and are universal in all parts of the world!

See my reply to batpig, I'm a beerer. Is there sucha word? Draft me a beer, a cold one, please, let it shave my tonsils!!

Like I said, Feri, I have the utmost respect for Chris. But he does have to toe his company's marketing line for obvious reasons. It just seems that it is not correct that human hearing is less sensitive to sounds emanating from behind than to sounds emanating from in front - and given that, it follows that if DEQ was based on that premise and that the design of the system was built around that, and that in reality they designed it with real-world material in which the surround content is lower to start with, instead of using equally loud content as their starting point, then we do have to, in this one instance at least, take Chris's comments with a pinch of salt. I don't think this denigrates Chris in any way or detracts from the numerous good things he has done for the members here. Certainly it is not my intention to do so.

It's interesting that THX Loudness Plus does the same thing, so are both in error?

http://www.thx.com/consumer/thx-tech...loudness-plus/

Obviously not everyone believes it is a flawed premise.

Good report. I'm glad you've done that because it saves me a job I am sure your results are the expected ones because it is exactly what one would expect based on Roger's input.


The conclusion I came to as well, albeit for slightly different reasons. And I never questioned DEQ before either. I just took it on face value that it was always "a good thing".


I have spent (wasted?) this evening doing DEQ tests. I have compared DEQ on vs off, compared DEQ vs THX Loudness Plus and compared both with DEQ with a 5dB reference level offset. I have done this at my typical MV of -9dB and also at MV of -25dB.

I notice something sort of similar to you with DEQ off. With DEQ on at MV -9dB, the sound is slightly 'thicker'. I can't pinpoint it any better than that. With DEQ on at -25dB the effect of DEQ is obviously much more noticeable and yes, the surrounds seem to be a bit too hot.

Compared with THX Loudness Plus using the same circumstances, at MV -9dB I preferred THX LP, mainly because the bass seemed 'thicker' and a little overblown with DEQ on. At -25dB I actually preferred DEQ to THX LP.

With DEQ at RLO of 5dB, at MV of -9dB I could hear only a very slight difference between DEQ and THX LP. I could not say which was preferable - just slightly different. This perhaps can be explained by the fact that using a RLO of 5dB and a MV of -9dB means that DEQ is not doing very much at all.

My conclusion from all this is uncertain I think that using a RLO offset of 5dB gives me a fairly nice result at MV -9dB and also would benefit me if I ever listened at lower MVs (which I don't usually). Using this RLO means that for most material I could just leave DEQ on all the time. I considered also testing a RLO of 10dB but at my usual MV of -9dB, DEQ would be doing nothing at all then which would have made that test no different than just running with DEQ off.

With DEQ and a RLO of 5dB, it is a toss-up as to whether I prefer that or THX LP. I need to listen to more material to be able to come to a conclusion there.

I may be peculiar in that I rarely ever (like almost never) use a MV of lower than -9dB, so these results may not apply to a lot of people.

So for me it is either THX LP or DEQ doing very little with the RLO of 5dB. If I decide to use a RLO of 10dB this, for me, would effectively mean DEQ is off at my usual listening level but I would benefit from it at any lower MV levels, without having to remember to turn it back on. I’ll give that some further thought after more prolonged listening to the current two favourites.

Let's be fair about the "obvious reasons": it's not because he has to say the emperor has clothes when naked. It's because Audyssey did scientific research that we dont have access to, but could conceivably re-create, that when analysed fed into their product design. The "obvious reasons" for Chris' statements are that given all the data available, he believes his statements are correct.


Part of the problem is whether we are talking about volume, clarity and detail, what frequencies, and what kind of noise floor in the room, just to start.


If the content had low surround levels at reference level, and they reproduced the same relative sense of surround at a lower master volume level, that's key. And this appears to be what they claim to have done. Clearly not all of us agree they achieved that parity.

Of course my favorite part of DEQ is maintaining a sense of bass presence.

Feri, surely Roger has already given his detailed reasoning? The suggestion of a 'home test' was to convince any doubting Thomases that his reasoning was correct.

That was the problem, Jeff. They based it on real world material where the surrounds were already much quieter than the fronts to begin with. As they lowered the overall volume this caused the surrounds to 'drop off' faster than the fronts. If the tests had been done using content where all channels were playing at the same level, then the algorithm would have been different to the one they ended up with.

Neither does Mr. Tribeman, and I agreed with is concept back when he invented it.

The devil's in the details. If it sounds right, it's all good.

Excactly the conclusion my tests brought me too (see reply to BCJ). Thanks for your input on this Sanjay. Thanks to you and Roger I have learned something that has resulted in an audible improvement here.

Feri - can you point to any reputable research (other than Audyssey's) that shows that human hearing sensitivity diminishes faster with sounds from behind than it does to sounds from in front? It does indeed seem to be a false assumption.

This isn't Audyssey-bashing - it is trying to better understand how DEQ works and as a result discovering that it does not actually work the way we have previously been told it does, or assumed it does.

I play games once in awhile on my set up and always felt that the surrounds were a bit hot. For example, in the game there's someone speaking from the front and when you turn around the voice gets louder from the surrounds. Turned off MultEQ and now it sounds about right.

^I was very impressed with DEQ from the start, to the point of upgrading from a Denon AVR 988 (MultEQXT) to a 989 solely for that new-at-the-time DEQ/DVol features which I felt enhanced my HT listening enjoyment tremendously. It certainly seemed worth the hundred bucks or so it cost me for what I consider a tweaking enhancement feature. It was actually offered to owners of the Denon 3808 for $100 as a FW upgrade. Back then I had a fairly average 5.1 HT setup including mid-quality Polk speakers and a basic Polk sub. I described the DEQ effect as uncanny, as the tonal balance was at such variance to what I (and most everyone) was used to when I turned down the MV from reference levels. And I loved it. And still do, mostly.

I listen below reference much of the time, as do most folks, and it semed to me that with DEQ on I was hearing something closer to what I'd hear in the the dubbing stage (at film reference level) or recording booth (at some unknown but doubtless higher level). And that, friends is my humble goal. Without DEQ I'd only had the bass control in the Denon, which I disliked, or I could of course simply turn the MV up louder. Since then I've upgraded all my gear and found better speaker/sub placements. And I've learned how to tweak DEQ according to source and my preferences, as I picked up tips from this thread and experimented with it.

Currently I use DEQ for film soundtracks off shiny disc all the time, usually listening at about MV -15 and often lower. DEQ works very well for film IME with no RLO tweaking needed. I am not a film soundtrack buff or expert but for me DEQ enhances the realism of the sounds and the surround bubble and hence my immersion in the film and suspension of disbelief. I don't watch much TV but DEQ needs a RLO of at least 10 as TV audio is not standardized to film reference. I usually turn DEQ off for sports games where I don't need the announcer's voices made any deeper or the crowd noises any louder in the surrounds.

Music is not standardized as is film either, as most of us here are keenly aware. For most of my casual 2 ch music listening in Stereo, DEQ is usually quite nice IME with RLO on 10. As a musician I am very attuned to what to listen for and I am particular about tight punchy bass. For finer tweaking the standard Audyssey recommended RLO settings for music listed here are quite useful IME.

I do my critical listening of good music source material (often stereo SACDs) close to what I think is "reference" level but it's really a guess. I usually listen at around 78-82 dB peak; often that's MV-15 or so. With RLO=10 DEQ is really not doing much at that level. However if the bass content is heavy in the recording I may shut DEQ off. If the bass is too light DEQ will surely be on.

MC music is even trickier irt DEQ. Almost all of my MC music is very well recorded (SACD/DVDA). In order to fill in the bass but not boost the surrounds excessively over what the mixer intended, I may try enhancing the RLO by use of the fader control to trim down the surrounds a few dBs. Often I feel there's too much guesswork and tweaking involved for little overall gain in SQ so for MC music near "reference" I usually just turn DEQ off.

Yes, I just turn off DEQ if/when I don't like it!

You say that my 'obvious reason' is an assumption and then you make an assumption that your own 'obvious reason' is correct

Any senior executive of any company who came out and said that they based their technology on a flawed assumption wouldn't last long in his job.

Audyssey did indeed do some scientific research. Unfortunately, it seems they started from a false premise.

Roger has already suggested a simple experiment one can do at home - if you try it and get different results, it would be very interesting to hear them.


That's the part that I disliked the most! Even more than what it does to the surrounds.