or Connect
AVS › AVS Forum › Audio › Receivers, Amps, and Processors › "Official" Audyssey thread (FAQ in post #51779)
New Posts  All Forums:Forum Nav:

"Official" Audyssey thread (FAQ in post #51779) - Page 1935

post #58021 of 70895
Quote:
Originally Posted by gurkey View Post

Dynamic means, it varies with the signal contents.
I can understand equalizing in order to minimize some of the room's unwanted contributions to the frequency response. I can understand changing equal loudness curves based on volume level in order to compensate for our non-linear human hearing.

But what is in the contents of the program that would need varying equalization on the fly? What is that equalization correcting or compensating for? Is it because the program contents have loud and soft sounds? Or is it something else?
post #58022 of 70895
Quote:
Originally Posted by IgorZep View Post

"Changing frequency response dynamically" and "compressing selectively depending on frequency" are technically exactly the same things, it just different wording.

OK, just what are you referring to with "compressing selectively depending on frequency" ?
post #58023 of 70895
Quote:
Originally Posted by batpig View Post

"While a person with infinitely quick hands, and iinfinitely quick mind and audible capacity far beyond the human could theoretically do exactly what a compressor does by riding the fader, compression is not the same as volume control."
I think I know this person, and his name is Dynamic Volume biggrin.gif

Dynamic Volume also uses "The Equal Loudness Contours", and according to the Audyssey thread religion, if a compression is based on it then it is not a compression biggrin.gif It is something else... the name of it is "Magic".
post #58024 of 70895
Quote:
Originally Posted by sdurani View Post

I can understand equalizing in order to minimize some of the room's unwanted contributions to the frequency response. I can understand changing equal loudness curves based on volume level in order to compensate for our non-linear human hearing.
But what is in the contents of the program that would need varying equalization on the fly? What is that equalization correcting or compensating for? Is it because the program contents have loud and soft sounds? Or is it something else?

It is because we not only hear low and high frequencies worse than middle frequencies, but also because our sensitivity to those frequencies fall down faster with decreasing volume.
post #58025 of 70895
Quote:
Originally Posted by pepar View Post

OK, just what are you referring to with "compressing selectively depending on frequency" ?
Multiband compression as an example. Different old Dolby and the like noise reduction systems for recording on tapes compressed only high frequencies (when recording, then expanding at playback) with pretty complicated compression rules.
post #58026 of 70895
Quote:
Originally Posted by sdurani View Post

I can understand equalizing in order to minimize some of the room's unwanted contributions to the frequency response. I can understand changing equal loudness curves based on volume level in order to compensate for our non-linear human hearing.
But what is in the contents of the program that would need varying equalization on the fly? What is that equalization correcting or compensating for? Is it because the program contents have loud and soft sounds? Or is it something else?

Hi Sanjay,

Here's an interesting Q&A with Chris K. where he explains the dynamic part of DynEQ:

Q:
So now I'm confused - what is Audyssey actually trying to achieve? Does it try to make a violin played at reference level sound like a violin played at the level corresponding to the master volume setting? Isn't the goal of equal-loudness compensation "just" to maintain the spectral characteristics of the whole recording, so a forte played violin still sounds forte even when the recording is played at -30 dB from reference?

A:
One goal is to maintain the same perceived spectral balance when listening at levels lower than those used during the mix. A violin, for example, is playing a wide range of notes and it is mixed so that every passage is at a given balance with the other instruments. When you turn the volume down the lower notes of the violin will start to be perceived softer in level than the higher notes for that same passage. The overall perception of the violin relative to other instruments playing along with it will also be perceived differently. So, with the static part of Dynamic EQ we are trying to make spectral adjustments that follow a set of static curves. The dynamic part of Dynamic EQ adds one more level of detail: it looks at the moment-by-moment variations in the content loudness and compares them to the perceptual model. Based on that information it determines how "loud" that passage will be perceived and then, by knowing how loud it was perceived during the mix, it makes a secondary adjustment. This secondary adjustment is, well... secondary and changes continuously as the signal is playing.
__________________
Chris

It works so smoothly that it would only become apparent when taken away. I wish that secondary adjustment of DynEQ could be toogled on/off so as to be able to verify it by our own ears. smile.gif
post #58027 of 70895
I have another probably silly question about subwoofers and XT32.

I have XT32, then do I even bother with subwoofers such as JL Fathom or B&W DB1? I know that the correct procedure is to let the subwoofer run its own EQ and then run XT32, but considering how good XT32 is supposed to be, why not save money and get a sub that does not have the built in EQ and let XT32 do the work? Or is there going to be an even greater benefit if one runs both the sub EQ and XT32? Thanks.
post #58028 of 70895
Quote:
Originally Posted by IgorZep View Post

It is because we not only hear low and high frequencies worse than middle frequencies, but also because our sensitivity to those frequencies fall down faster with decreasing volume.
Understood, but switching equal loudness curves already compensates for our non-linear hearing and our changing sensitivity as volume level decreases. Once you've compensated for that, what is in the program material itself that needs further correction? Do you need to change the relationship between loud and soft sounds, or something else?
post #58029 of 70895
Quote:
Originally Posted by IgorZep View Post

Dynamic Volume also uses "The Equal Loudness Contours", and according to the Audyssey thread religion, if a compression is based on it then it is not a compression biggrin.gif It is something else... the name of it is "Magic".

The ELC is supposed to be used by DynEQ not DynVol. When all three Audyssey features (MultEQ+DynEQ+DynVol) are engaged I think (?) DynVol is sandwiched in between MultEQ and DynEQ as reagards the signal flow. Errr,... I'm not really sure about that I have to admit, so time to ask Chris. Hold on please.cool.gif
Edited by mogorf - 12/11/12 at 2:01pm
post #58030 of 70895
Quote:
Originally Posted by sdurani View Post

Quote:
Originally Posted by IgorZep View Post

It is because we not only hear low and high frequencies worse than middle frequencies, but also because our sensitivity to those frequencies fall down faster with decreasing volume.
Understood, but switching equal loudness curves already compensates for our non-linear hearing and our changing sensitivity as volume level decreases. Once you've compensated for that, what is in the program material itself that needs further correction? Do you need to change the relationship between loud and soft sounds, or something else?

I'm surprised to hear you asking about this, Sanjay, as you've been around this thread forever. Surely you've read responses from Chris like the one pasted above?
post #58031 of 70895
Quote:
Originally Posted by IgorZep View Post

If I pass 50Hz signal through processor, then get something like this (numbers are arbitrary): for 0dB in +5db out, for -20db in -10db out, so instead of 20db range I get 15db range - it is a compression, and it doesn't matter if it is because "the frequency response is changed" and 1kHz signal would not be compressed. It just "excuse" that is there only to not use the word "compression" so users would not be scared, as they are learned to think compression is a bad thing.

Igor, aren't we just arguing semantics at this point? It seems nobody is disagreeing about what Dynamic EQ actually DOES, you are just saying it should be called "compression" whereas others are saying it's "frequency response modulation".
post #58032 of 70895
Quote:
Originally Posted by BVLDARI View Post

I have another probably silly question about subwoofers and XT32.

I have XT32, then do I even bother with subwoofers such as JL Fathom or B&W DB1? I know that the correct procedure is to let the subwoofer run its own EQ and then run XT32, but considering how good XT32 is supposed to be, why not save money and get a sub that does not have the built in EQ and let XT32 do the work? Or is there going to be an even greater benefit if one runs both the sub EQ and XT32? Thanks.

Basically, it depends on how bad the bass situation is before you start applying electronic room correction. XT32 has a ton of processing power, but it doesn't have infinite processing power...

So, it's possible (maybe even probable in most rooms) that XT32 alone will have enough processing power to equalize the bass as good as it will get..... but, let's say you have a couple of really nasty modes. First applying some broad correction, like AntiMode or the built in PEQ of some subwoofers, to tackle the nastiest "big bumps", will give XT32 a better starting point from which to start working, which means the final response could be even flatter.

So you can look at is as the first layer (a simple one or two band PEQ correction) as the "hatchet" to hack off the biggest bumps, and then XT32 is the scalpel that goes in and does really fine correction on all the little dips and bumps to get that clean final response.
post #58033 of 70895
Quote:
Originally Posted by BVLDARI View Post

I have XT32, then do I even bother with subwoofers such as JL Fathom or B&W DB1? I know that the correct procedure is to let the subwoofer run its own EQ and then run XT32, but considering how good XT32 is supposed to be, why not save money and get a sub that does not have the built in EQ and let XT32 do the work? Or is there going to be an even greater benefit if one runs both the sub EQ and XT32?
Depends on number of subs you are using and how many seats you want to correct for.

If you're using a single subwoofer and primarily concerned with the main listening position, then use the EQ built into your subwoofer to get a flatter/smoother response at the sweet spot. This way when you measure with XT32, doing a tight circle around the sweet spot, Audyssey will see a flatter/smoother subwoofer response and correct whatever the subwoofer's built-in EQ couldn't.

However, if you're EQing for multiple seats, then I wouldn't use the EQ built-into the subwoofer, since it will likely correct for a single seat, which could make things worse in other seats. Why unnecessarily saddle Audyssey with that problem. So I would forego the subwoofer's built-in EQ and only use XT32.

Likewise, if you're using multiple subs, the EQ built into each sub will likely be correcting for a single subwoofer, not the interaction for both subs together. This isn't very helpful, since you won't be listening to each subwoofer individually. By comparison, XT32 (with or without Sub EQ HT) will measure and correct for the interaction of both subs together, which is how you'll be hearing them anyway. So, again, forego the EQ built into both subs and just use XT32.
post #58034 of 70895
Quote:
Originally Posted by pepar View Post

Quote:
Originally Posted by IgorZep View Post

"Changing frequency response dynamically" and "compressing selectively depending on frequency" are technically exactly the same things, it just different wording.

OK, just what are you referring to with "compressing selectively depending on frequency" ?

I think this is what he is saying (Igor, forgive me if I am butchering your meaning):

Picking arbitrary numbers for the sake of argument.....

Let's say you have the MV turned down to some level such that Dynamic EQ is active. Let's say that, based on DEQ's equal loudness contours, a 50Hz sound will require boosting, but a 2kHz tone will be left flat.

So then, a particular passage of average volume comes along, and the 50Hz portion is being boosted by +5dB, whereas that 2kHz portion is still flat (no boost).

Then, the passage becomes much softer, let's say it drops by 10dB globally. Dynamic EQ reacts to the softer content, and is now boosting that 50Hz portion by +8dB, whereas that 2kHz portion is still getting 0 boost.

So the "selective depending on frequency" part is that a different adjustment is being applied to the 50Hz part than the 2kHz part of the frequency spectrum.

And what he is arguing is "compression" is the fact that the 50Hz part was only allowed to drop by 7dB (the net of the 10dB drop minus the 3dB increase in DEQ's boosting, from +5dB to +8dB). The "dynamic" part of DEQ prevented that 50Hz portion from dropping the full 10dB as called for in the soundtrack, thus was "compressed" because the full dynamic range was restricted *at that selective frequency*.
post #58035 of 70895
Quote:
Originally Posted by IgorZep View Post

Multiband compression as an example. Different old Dolby and the like noise reduction systems for recording on tapes compressed only high frequencies (when recording, then expanding at playback) with pretty complicated compression rules.

No, no, no, we are (were?) discussing Dynamic EQ and you opined that it was "compression" and vigorous disagreement ensued. So what part of DEQ is "compressing selectively depending on frequency" ?
post #58036 of 70895
Quote:
Originally Posted by batpig View Post

I think this is what he is saying (Igor, forgive me if I am butchering your meaning):
Picking arbitrary numbers for the sake of argument.....
Let's say you have the MV turned down to some level such that Dynamic EQ is active. Let's say that, based on DEQ's equal loudness contours, a 50Hz sound will require boosting, but a 2kHz tone will be left flat.
So then, a particular passage of average volume comes along, and the 50Hz portion is being boosted by +5dB, whereas that 2kHz portion is still flat (no boost).
Then, the passage becomes much softer, let's say it drops by 10dB globally. Dynamic EQ reacts to the softer content, and is now boosting that 50Hz portion by +8dB, whereas that 2kHz portion is still getting 0 boost.
So the "selective depending on frequency" part is that a different adjustment is being applied to the 50Hz part than the 2kHz part of the frequency spectrum.
And what he is arguing is "compression" is the fact that the 50Hz part was only allowed to drop by 7dB (the net of the 10dB drop minus the 3dB increase in DEQ's boosting, from +5dB to +8dB). The "dynamic" part of DEQ prevented that 50Hz portion from dropping the full 10dB as called for in the soundtrack, thus was "compressed" because the full dynamic range was restricted *at that selective frequency*.

Perhaps that is what he is saying, but it just seems like a real stretch ... a tortuous stretch ... to make DEQ into compression. I will never see how boosting the high frequencies and low frequencies to match our hearing is compressing the signal.
post #58037 of 70895
If listening robots would be put into the room with microphones on their heads that have a flat FR thoroughout the whole MV range then DynEQ would not be needed. We humans need the research behind DynEQ to be able to enjoy percieved loudness at any MV setting. Kudos to Audyssey for making that possible. smile.gif
post #58038 of 70895
Quote:
Originally Posted by batpig View Post

I'm surprised to hear you asking about this, Sanjay, as you've been around this thread forever. Surely you've read responses from Chris like the one pasted above?
I have, but his response was a non-response: "This secondary adjustment is, well... secondary".

I don't begrudge Chris not wanting to reveal proprietary techniques, but in light of the 'explanation' quoted above, I don't think my question was unreasonable. MultEQ corrects for the room. The first step of DEQ compensates for our hearing. What is the "secondary adjustment" compensating for? What is in the source material that needs further correcting?
post #58039 of 70895
batpig, exactly! You get the point smile.gif
post #58040 of 70895
Quote:
Originally Posted by sdurani View Post

I have, but his response was a non-response: "This secondary adjustment is, well... secondary".
I don't begrudge Chris not wanting to reveal proprietary techniques, but in light of the 'explanation' quoted above, I don't think my question was unreasonable. MultEQ corrects for the room. The first step of DEQ compensates for our hearing. What is the "secondary adjustment" compensating for? What is in the source material that needs further correcting?

Don't forget that once we move away from calibrated 0 dB reference level MV setting things start to change in our perception of the same source material. smile.gif Techniques may remain proprietary, but the phenomenon is still there.
post #58041 of 70895
Quote:
Originally Posted by mogorf View Post

The ELC is supposed to be used by DynEQ not DynVol. When all three Audyssey features (MultEQ+DynEQ+DynVol) are engaged I think (?) DynVol is sandwiched in between MultEQ and DynEQ as reagards the signal flow. Errr,... I'm not really sure about that I have to admit, so time to ask Chris. Hold on please.cool.gif

Didn't have to wait too long for the answer, so here it is:

Me:

Hiya Chris, please help me out with my thinking. As we all know by now DynEQ does a secondary adjustment: it looks at the moment-by-moment variations in the content loudness and makes changes continuously as the signal is playing. Now, when DynVol is engaged there is a kinda dynamic range compression involved (shortly speaking). Since the secondary adjustment of DynEQ looks into the soft and loud parts of a passage on-the-fly, is my thinking correct that DynVol as regards the signal flow is sandwiched in between MultEQ and DynEQ? If not, what is the "truth"? Thanks in advance as always. Cheers, Feri

Chris:

Hi Feri, your thinking is correct.
post #58042 of 70895
Quote:
Originally Posted by mogorf View Post

Don't forget that once we move away from calibrated 0 dB reference level MV setting things start to change in our perception of the same source material.
That 'change in our perception' is what equal loudness curves compensate for. After the room and our hearing are addressed, why do we need to correct the source material? What is in the source material than needs correcting?
post #58043 of 70895
Quote:
Originally Posted by kbarnes701 View Post

Like the guys I quote below, I have reached the age where the difference between the two curves is difficult to hear. I love DEQ and think it is very important.

Sorry about reverting back to an earlier topic, but good grief this was posted this morning. Lots of posts today.

First off:
Old = current age + 10
remove glasses before looking in mirror

I was a bemused by the posts where several of those much younger than me (65 and can't distinguish anything above 14K) were saying they couldn't detect much difference between the Audyssey and Audyssey Flat curves. As I've always thought there was a pretty distinguishable difference I got curious.

I captured the filter graphs on my Denon 3808 (these show just the LF, but LR is similar - if images are saved and displayed in a viewer toggling back and forth reveals changes in the curves)

Audyssey curve:


Audyssey Flat Curve:


It appears to me that there are some significant differences in the curves even below 10K which I didn't expect given the explanations of what the rolloff was designed to do. It doesn't surpirse me that even I can hear these differences. I guess my question is whether these differences are "as designed" or if there is some error in the processing in my receiver?
post #58044 of 70895
Quote:
Originally Posted by sdurani View Post

That 'change in our perception' is what equal loudness curves compensate for. After the room and our hearing are addressed, why do we need to correct the source material? What is in the source material than needs correcting?

Let me put it this way Sanjay (seed for thought). Imagine that we could have a movie on a Blu-Ray disk that was recorded and mixed at not only reference level, but the same disc would have a user selectable option via menu to playback the material also recorded at the mixing desk at -5 dB, -10 dB, -15 dB, -20 dB,-25 dB, etc. You follow, eh? Would be a hekuva job from the sound engineers, but we're just playing with the idea. smile.gif

Then the whole secondary adjustment of DynEQ would be pointless, coz that secondary real-time adjusment at any MV setting as per the above 5 dB increments would negate the need for that secondary adjustment in the first place.

But since the sound engineers record to only one single level called reference level, the secondary adjustment of DynEQ comes into play since it is us who depart from 0 dB reference level when listening in our home environments while wishing to be able to listen at the same perceptual quality as it was done and heard during the studio recording.
post #58045 of 70895
Quote:
Originally Posted by mogorf View Post

But since the sound engineers record to only one single level called reference level, the secondary adjustment of DynEQ comes into play since it is us who depart from 0 dB reference level when listening in our home environments while wishing to be able to listen at the same perceptual quality as it was done and heard during the studio recording.
You're still avoiding the question of what that "secondary adjustment" is. Once you've compensated for listening at lower than reference levels, what are you correcting in the source material itself? Boosting quieter sounds?
post #58046 of 70895
Yes, I think it's pretty clear from Chris' responses on this matter that the "secondary" part is based on moment to moment changes in actual program content.
post #58047 of 70895
Quote:
Originally Posted by pepar View Post

Quote:
Originally Posted by batpig View Post

And what he is arguing is "compression" is the fact that the 50Hz part was only allowed to drop by 7dB (the net of the 10dB drop minus the 3dB increase in DEQ's boosting, from +5dB to +8dB). The "dynamic" part of DEQ prevented that 50Hz portion from dropping the full 10dB as called for in the soundtrack, thus was "compressed" because the full dynamic range was restricted *at that selective frequency*.
Perhaps that is what he is saying, but it just seems like a real stretch ... a tortuous stretch ... to make DEQ into compression. I will never see how boosting the high frequencies and low frequencies to match our hearing is compressing the signal.
Using Batpig's example: if you've limited how quiet a sound can get, by boosting it when it gets too soft, have you not squeezed its dynamic range?
post #58048 of 70895
Quote:
Originally Posted by batpig View Post

Yes, I think it's pretty clear from Chris' responses on this matter that the "secondary" part is based on moment to moment changes in actual program content.
Moment to moment changes in...volume level, frequency response, or something else? What do you think is changing that needs correcting on a constant basis?
post #58049 of 70895
I think it's volume level -- the frequency response part is taken care of by the loudness contour right? The point of the "secondary" part as I understand it as that shifts in LEVEL of actual program content will mean that Dynamic EQ is "dynamically" adjusting the loudness contour, moment to moment, based on the level of the content, as opposed to just picking a fixed loudness contour based on the MV setting.
post #58050 of 70895
Quote:
Originally Posted by batpig View Post

I think it's volume level -- the frequency response part is taken care of by the loudness contour right? The point of the "secondary" part as I understand it as that shifts in LEVEL of actual program content will mean that Dynamic EQ is "dynamically" adjusting the loudness contour, moment to moment, based on the level of the content, as opposed to just picking a fixed loudness contour based on the MV setting.
OK, that makes sense. Aside from the bass and treble dropping faster than the mid-frequencies, quieter sounds disappear faster than louder sounds when you lower the volume level. By 'disappear' I mean fall below the noise floor of your system or fall below your hearing threshold.

So, for example, if you're listening to a simple mono recording of a singer in a jazz club, you can hear her voice clearly in the foreground and subtle room ambience in the background. As you lower the volume level, not only does her voice get quieter but it also gets drier (room ambience becomes inaudible). It's almost like she goes from singing in a club to singing in an absorbtive recording studio, changing what the artists intended you to hear.

If you can boost only those sounds that are about to disappear, by doing nothing more than keeping them above the noise floor or threshold of hearing, then you still get to hear the singer in the intended context of the club, with the subtle recorded ambience in the background. Yes, the dynamic range has been squeezed a bit (since sounds that were boosted when quiet weren't similarly boosted when they are loud), but it is worth it in order to hear the original intent of the recording. Besides, if you don't like what it's doing or you hear artifacts (pumping), you always have the option to turn it off.

IF this is what DEQ is doing, then it is not exclusive to Audyssey. Dolby Volume, for example, does the same thing. In fact, depending on manufacturer's implementation, you have the option to turn off the Leveler (boosts quieter sounds) while still keeping the Modeler (applies equal loudness contours). THX Loudnes Plus does the same, though I don't know if it has the flexibility to use one part (loudness contour) without the other (level boost).
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Receivers, Amps, and Processors
AVS › AVS Forum › Audio › Receivers, Amps, and Processors › "Official" Audyssey thread (FAQ in post #51779)