Clearly, if we can apply some absorption and kill a reflection or move woofers to boundaries and prevent reflections, then that is a preferred correction. Still this whole discussion has been how to best measure a system in a room for the sake of adding electrical EQ. If we believe that Kates and Bech are correct in that the aberrations are audible, then we don't want to adjust our measuring approach to one that tends to hide the audible problems.
Note that Kates is not just saying that the aberrations are audible, but more "this is the curve as you perceive it". If you accept that, then EQing that perceptual curve to flat must be the answer. You should read the paper as critical band smoothing is just a small part of it.
We can try the cautious approaches to EQ, such as fully correcting peaks and only partly correcting the dips, etc., but saying that we made the system flat in the anechoic chamber, and then doing nothing, seems like the suboptimum choice.
Regards,
David
I'm not really following the argument either. I think you are still claiming that your MMM technique is pulling out the direct response or at least showing more than other measurements at that distance would. This isn't true. MMM is simply a handy way to do spatial averaging. The direct to reflected ratio is set by the listening distance and spatial averaging doesn't change that, other than to degrade it more towards the reverberant field as the sampling area increases.
Regarding the Toole and Olive tests, they have shown that they can predict fairly well (above the Schroeder frequency) the room curve. They have not shown that it is important. As I mentioned a few posts back they have done recent tests that show that a small but significant benefit came from biasing the EQ towards the direct response, rather than EQing for the smoothest power response (smooth power, not flat, which is never good).
As to your measured aberrations being inaudible, I guess I am shocked that that is true. Those are very significant dips and balance shifts. You are saying you can put in pink noise, move those distances, and not hear character changes to pink noise? Makes me wonder if we are putting too much effort into perfecting our speaker designs.
Regards,
David
A Multiple Regression Model for Predicting Loudspeaker Preference Using Objective Measurements: Part II, in figure 7, you clearly see that ER and PIR curves are very near one from another.
To the point that in later papers, the PIR curve seems generally to be dropped out (see the M2 responses in Pos's post)
Toole, in "Loudspeakers and Rooms for Sound Reproduction—A Scientific Review", p454
"In thinking about what may happen in the small rooms of interest to us, assuming no other differences, the critical distance will be larger because these rooms have proportionally more sound-absorptive material and the sound sources have significant directivity, and are aimed at the listener. As a result, we may find that we are not listening in the reverberant sound field."
I didn't say that MMM is different to other spatial averaging methods, I wanted to say that the differences with Harman responses is that spatial averaging measurements I generally get are not strictly PIR (or ER, see above) but between ER and LW, so showing more from direct field and less from diffuse field than expected.
But measurements with log sine sweep at multiple positions or MMM give same results.
Yes, why would people not bother about those reflections : the level is high, the timing is short !
Maybe because most film mixers have two ears and because they are able to move their head while listening ?
I would hypothesize that :
- The spectrum of the console reflections is very quickly changing with even small movements of the head (much more than floor or ceiling reflections which origins are farther from listener and so change less with head movements)
- the spectrum of the reflections may be different at each ear
Sorry for another sentence from Toole, p459 of same ref as above :
"It is as if we can separate the sound of a spectrum that is changing (the sounds from the different loudspeakers) from that which is fixed (the colorations added by the room itself for the specific listener and loudspeaker locations within it)"
Which paper are you referring to ? Kates, A Perceptual Criterion for Loudspeaker Evaluation ? I have to read it.
Last edited:
There are two separate questions:
1. how to measure the various elements of sound from loudspeakers in rooms, and
2. what are the elements of sound that are humanly meaningful and how do these create perceived sound; how do listeners create or ignore certain streams
Question #1 can get a reasonably direct tech answer and this thread delves into that nicely. But Q #2 has only hints in the published body of knowledge (and that includes psychology).
All along, each writer in this lively debate has slipped in crypto-perceptual terms like "inaudible" and so on. But the listener is the "elephant in the room" and needs consideration beyond the fine-grain of measurement technique.
My own 2-cents-worth is that many experiments are flawed because sound inherent in room modes and recurrent resonances are learned within an hour and forward are treated as a perceptual stream that can be ignored like yellow light from certain light bulbs (and hence fall below the listeners consciousness). The time-course of this learning needs to be taken into account in studies because you'd get really different results sooner or later (just as listening to your home music room through headphones makes the room sound really different). I suspect in familiar rooms, the room modes "go silent" after a while and therefore it isn't important to address these with EQ since they aren't noticed (as several posters above have also said).
That leaves direct, early reflections, and late reflections as perceptually meaningful elements.
Ben
1. how to measure the various elements of sound from loudspeakers in rooms, and
2. what are the elements of sound that are humanly meaningful and how do these create perceived sound; how do listeners create or ignore certain streams
Question #1 can get a reasonably direct tech answer and this thread delves into that nicely. But Q #2 has only hints in the published body of knowledge (and that includes psychology).
All along, each writer in this lively debate has slipped in crypto-perceptual terms like "inaudible" and so on. But the listener is the "elephant in the room" and needs consideration beyond the fine-grain of measurement technique.
My own 2-cents-worth is that many experiments are flawed because sound inherent in room modes and recurrent resonances are learned within an hour and forward are treated as a perceptual stream that can be ignored like yellow light from certain light bulbs (and hence fall below the listeners consciousness). The time-course of this learning needs to be taken into account in studies because you'd get really different results sooner or later (just as listening to your home music room through headphones makes the room sound really different). I suspect in familiar rooms, the room modes "go silent" after a while and therefore it isn't important to address these with EQ since they aren't noticed (as several posters above have also said).
That leaves direct, early reflections, and late reflections as perceptually meaningful elements.
Ben
Last edited:
That's your conclusion? I don't think that anyone is disagreeing with that. It's kind of obvious isn't it?
jlo - I didn't connect ER that you were mentioning to these measurements. How is ER defined in the Toole graphs? To me Early Reflections are a pattern of sound not an averaging method. They would not include the direct sound for example. Different definitions I guess.
This is precisely what I was alluding to at the end of my previous post, but it is simply stated as a hypothesis without any real support. Kind of a guess really. But that still noes not change the fact that if one is beyond the critical distance - defined as where the direct field and the reverberant field are equal - then the MMM will measure dominantly the reverberant field even if this distance is different than that predicted by a 6 dB / DD. It is not going to reduce the reverberant ratio by simple averaging as I keep understanding you to imply.
This is precisely what I was alluding to at the end of my previous post, but it is simply stated as a hypothesis without any real support. Kind of a guess really. But that still noes not change the fact that if one is beyond the critical distance - defined as where the direct field and the reverberant field are equal - then the MMM will measure dominantly the reverberant field even if this distance is different than that predicted by a 6 dB / DD. It is not going to reduce the reverberant ratio by simple averaging as I keep understanding you to imply.
No, that isn't my conclusion and it isn't obvious or agreed that room modes are not important in "hearing" in a room after a person has "learned" the room sound, as I guess.
My core thought is to ask what "are the elements of sound that are humanly meaningful [as opposed to physically meaningful to engineers]; how do these create perceived sound; [and] how do listeners create or ignore certain streams [in order to make sense of the sound]"?
If you have knowledge to add to what has already been brought up on those topics (for example, Kates, Toole, Olive), very glad to hear it. Otherwise, prefer not to be the mis-aimed target of your sarcasm.
Ben
Last edited:
Are we talking about the same curve?
I am talking about the green one here, the one that represents the averaged anechoic direct response over a 30° horizontal and 10° vertical angles, so-called "listening window". It is not room dependent as it is solely the direct field.
As it is a spatially averaged response curve I think it is safe to say that any localization specific accident like diffraction and comb filtering effects will be blurred and minimized. IMHO this is a good thing and that curve is probably the best one to EQ, and that is exactly what JBL does these days.
As I have mentioned before, they never showed evidence that the predicted in-room curve was perceptually important. More recently they have given curves that, again, show that the direct or early sound is key.
But no more Listening Window than the room allows. Your technique can not drill down to the direct sound by any amount.
Absolutely not.
Toole likes to make a distinction between reflections and resonances and I think he is totally wrong in this.
Try a few tests. Get a good sized piece of foamcore or plywood. Put pink noise into a speaker and place the plywood behind the speaker. Move it near and away from the speaker and you will hear the constantly varying comb filter effect.
Place the same piece in front of and below the system and move it about. (Simulates a floor bounce.) You will hear a comb filter effect.
Place a ruler on edge on the system baffle to the side of a tweeter. Vary its angle and distance to the tweeter and you will hear a comb filter effect.
All of these effects, once heard on pink noise can also be detected on music. They are all reflections rather than resonances. You and Toole would say the are inconsequential when they are not. Your measuring approach will tend to average them out when they are clearly audible.
But your arguement was that the mixers are moving and so they can ignore the reflection. Toole is saying that the stationary effects are ignored (room) and the variable effects stand out (loudspeaker), quite the opposite. In any case we can expect some accomodation (getting used to our pink sunglasses) but that doesn't mean that the stationary effects are inaudible or shouldn't be fixed. Otherwise we would just listen/monitor over the worst speakers and say "we'll get used to it".
I've seen a lot of studio designers put considerable effort into reducing console bounce. Are they wasting their time?
Yes, pay attention to his reference to the floor bounce that makes it through his measurement filters.
First, your experiment establishes (even without anybody on this forum needing to try it besides as a "thought experiment") that you sure can A-B instantaneous changes. The rest of your remarks about combs is gratuitous.
Second, a world of difference between reflections and room modes. Room modes are present, so to speak, even when the room is empty. Like with poorly constructed bass reflex boxes, they are an enduring and fairly stable characteristic of a room. I would guess that our hearing would be as bad as listening to somebody talking to you on a speakerphone if our minds were unable to subtract away room modes.*
I guess that Toole is more right here.
About your comment about what professionals call "adaptation" to pink sunglasses, the plain truth is that you do adapt and pinkness goes away (as do some other discriminations). But you can't take well established mechanisms about perceptual elements (color, motion, temperature) and verbally transpose that to any word-concept you feel like (like getting "adapted" to lousy speakers).
Ben
* I bet somebody will want to ask why you can't "subtract away" room modes when listening to somebody talking to you on a speakerphone.
Last edited:
Room modes and early reflections and comb filtering modify perception of direct sound and mask perception of low level recorded sound.
That we might "get used to them" is not a good point to make cuz it ain't hifi - it might be a pleasant sound we hear but it's not the reproduction of the signal that came to the speakers.
I personally find them annoying and therefore distracting because I keep trying to hear what I know ought to be sounding and is not.
That we might "get used to them" is not a good point to make cuz it ain't hifi - it might be a pleasant sound we hear but it's not the reproduction of the signal that came to the speakers.
I personally find them annoying and therefore distracting because I keep trying to hear what I know ought to be sounding and is not.
Adaption to sound is a fact of life. It happens and is well known. But it is difficult to determine the good aspects from the bad. If it were not for adaption people would not come to judge any new loudspeakers against their current loudspeakers and the marketplace would change. But the persistence of bad loudspeakers is a direct result of adaption. Should we adapt to bad sound? Shouldn't we seek out that that is the purest, without the need for adaption?
"Adaption" is also a clear problem when discussing "perception". If we have adapted to not hear a problem, i.e. not "perceive" it, is it no longer a problem?
I once had this exact conversation with Dave Clark of DBX fame and Behringer vs. Orion. He agreed that adaption is a big part of loudspeaker evaluation. I then asked him: "Then wouldn't it make the most sense to adapt yourself to the speakers that measured the best?" He had to agree that would be true.
Ben - I didn't mean to be sarcastic, I just didn't see the line that I quoted as saying anything that hadn't already been said or was well understood. As far as adapting to room modes, I suppose that we can adapt to anything, but I prefer not to do that. I think that listening is less fatiguing when we don't have to twist our brains to ignore some obvious flaws.
"Adaption" is also a clear problem when discussing "perception". If we have adapted to not hear a problem, i.e. not "perceive" it, is it no longer a problem?
I once had this exact conversation with Dave Clark of DBX fame and Behringer vs. Orion. He agreed that adaption is a big part of loudspeaker evaluation. I then asked him: "Then wouldn't it make the most sense to adapt yourself to the speakers that measured the best?" He had to agree that would be true.
Ben - I didn't mean to be sarcastic, I just didn't see the line that I quoted as saying anything that hadn't already been said or was well understood. As far as adapting to room modes, I suppose that we can adapt to anything, but I prefer not to do that. I think that listening is less fatiguing when we don't have to twist our brains to ignore some obvious flaws.
I see gedlee's sense pretty much. Which is happily often the case. Thanks.
But... just as an engineer would be apoplectic about people misusing terms like mass and weight.... let me say this. Many aspects of perceptual systems show adaptation and its cousin, negative after-effect. Depending on the issue, there are specific time-constants (a thought that should sound comfortingly scientific to some people).
When you hike to the HiFi store to hear speakers, you are well out of time range of any adaptation (or negative after-effects) that I know of*. You might have some loose memory of the sound you are at-home with ("great soup, burnt like my mother used to make it"). But that's called memory, not adaptation.
But that's nothing like sensory adaptation, except just resembling with words.
I don't know if FrankWW is stating a basic credo of his music reproduction faith, math ideals, or widely accepted facts or, as seems to me, a mix of all if a bit light on the last one. I am puzzled to learn that reflections are somehow getting louder in FrankWW's music room when the direct music is softer and obscuring the direct sound.
Ben
*unless like some hifi fans, you live in a store
But... just as an engineer would be apoplectic about people misusing terms like mass and weight.... let me say this. Many aspects of perceptual systems show adaptation and its cousin, negative after-effect. Depending on the issue, there are specific time-constants (a thought that should sound comfortingly scientific to some people).
When you hike to the HiFi store to hear speakers, you are well out of time range of any adaptation (or negative after-effects) that I know of*. You might have some loose memory of the sound you are at-home with ("great soup, burnt like my mother used to make it"). But that's called memory, not adaptation.
But that's nothing like sensory adaptation, except just resembling with words.
I don't know if FrankWW is stating a basic credo of his music reproduction faith, math ideals, or widely accepted facts or, as seems to me, a mix of all if a bit light on the last one. I am puzzled to learn that reflections are somehow getting louder in FrankWW's music room when the direct music is softer and obscuring the direct sound.
Ben
*unless like some hifi fans, you live in a store
I agree all around.
Due to adaption we prefer quick changeover tests. Differences are most obvious at the point of change. If flaws aren't too bad we can get used to them and overlook them to a point. (break in?) Still, a bad speaker is a bad speaker and there are absolutes in listening tests. Clearly, in the case of monitoring there is also a danger of getting used to a misbalanced speaker and putting complimentary errors into the program, to flatten the total sound.
One thing we used to ponder at one of my companies as we did comparison tests between multiple speakers. There was always a loser in the group as well as a winner. But I am convinced that you could take the loser home, open a bottle of wine, put on a disc and enjoy the music. Outside of the direct comparison the standards relaxed.
Back to Toole and the room tests, he was concerned that listener preference could be tied to room. That is, what if everyone preferred speaker A in room one but speaker C in room two? There would be no absolutes and we couldn't design the best speaker, only the best speaker for a particular room.
He was relieved through his testing to find that preference did not vary from room to room. We either got used to or compartmentalized the room sound (primarily LF modes) and concentrated on the differences the speakers provided within that room. Afterwards he went back and did dummy head recordings in the rooms, so that rooms alone could be A B'ed. In those tests the rooms were highly audible and easily differentiated. The sound of the room was no longer stationary. Luckily no speaker was an exact inverse filter for the non-linear room.
Again, we might get used to stationary errors, but that doesn't make them desirable. With our Rose colored glasses on we quickly get used to the tint and forget that any color is added at all, but notice that pink looks like white and white looks like pink, Not a good thing!
David
OK... I sense some agreement that some aspects of rooms - called "stationary" by Speaker Dave - are susceptible to disappearing as we learn the room*. And the other aspects, have to be identified as separate entities, named, and measured. Back to this thread.
Ben
*I admit to making up the notion of "learning" the room some years ago, although I am sure smarter others have done so long before me. But it is about time to incorporate room-learning in our understanding of reproducing music at home (and in the experiments which can be contaminated by learning and adaptation effects). Too bad it doesn't fit nicely in the world of math truths (like comb filtering) that some take as the whole of psycho-acoustics.
Ben
*I admit to making up the notion of "learning" the room some years ago, although I am sure smarter others have done so long before me. But it is about time to incorporate room-learning in our understanding of reproducing music at home (and in the experiments which can be contaminated by learning and adaptation effects). Too bad it doesn't fit nicely in the world of math truths (like comb filtering) that some take as the whole of psycho-acoustics.
Last edited:
Not exactly, as I already said, spatial averaging near listening place will average out more the lateral or close reflections (ie from the console) than the reflections or diffraction coming from near loudspeakers.
I didn't say that the mixers were moving, I said that while listening, everybody does small movements of the head (this is well known mainly for front/back localization). But it may also remove a part of coloration from unstable comb filtering.
Empirical. 1. Get close to the speaker and there's all sorts of detail heard which disappears as you move away. 2. Listen to music on head phones while at the normal listening position in the room. Lift them off and you can hear the loss of low level details.
I see no point in listening to grunge in place of music, even if it's nice sounding grunge.
Edit. Musical grunge is like the light.We don't see the light; we see things. Bright light, for instance, can mask detail in chiaroscuro effect.
Your last sentence doesn't cohere after the word 'when'.
I see no point in listening to grunge in place of music, even if it's nice sounding grunge.
Edit. Musical grunge is like the light.We don't see the light; we see things. Bright light, for instance, can mask detail in chiaroscuro effect.
Your last sentence doesn't cohere after the word 'when'.
Last edited:
To me the adaption is not rapid at all. You listen to a particular set of speakers for a long time and they become your reference, right or wrong. Then everything is compared to that reference and "better" speakers may not actually be perceived as better (unless you re-acclimate to them which takes a long time.). If you acclimate to a know set of high quality speakers - based on measurements not reviewers - then you acclimated to a standard that you should have, not a false one.
Message from the Department of Down the Rabbit Hole:
I suppose you are correct that "adaption" is a slow process, if this process exists at all. It could opeerate long enough to get to the HiFi store as you say. In as much as you invented the term it can mean exactly what you choose.
But to cognitive psychologists, "adaptation" is faster. l prefer to use that concept, long established in perception studies.
There are all kinds of factors operating in comparing speakers leading to false results (with short-term adaptation being just one). The Toole group mention this although I can't recall specifics. Is that their proprietary Harman Int. information? Link?
Ben
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.