An Enticing Marketing Story, Theory Without Measurement?

[RayDunzl]

I heard MBLs at the recent show.

Visited the (big) room briefly, 3 times.

Liked them one time out of the three.

 

[b1daly]

Interesting discussion, and as I commonly experience in such threads, there was something that surprised me...such that I think I might have not understood.

When it comes to listening environments, I’m fully of the “if it sounds good it is good” philosophy. I mean for a personal space, not in professional setting.

I’m a fan of using room furnishings as treatment because, aside from the obvious convenience and asthetics, I find the the combination of shapes and absorbing materials allows for a kind of diffuse yet lively sound that I find hard to get in studios, with treatment.

What surprised me was that my subjective experience is that hard reflective surfaces can be very problematic in how they “mess up” the clarity and image from the speaker. They can create a “phasey” sound, which throws off my ability to hear where the sound is coming from, it’s annoying and makes the image seem unstable.

This is the high frequencies so it’s usually not hard to deal with by putting something there to diffuse or absorb.

What I’m saying only works in smallish rooms. But by using a variety of home furnishings, you can tune the room to a large degree, but keep a natural sounding space.

In our studio control room, which is used for critical listening, I’ve never been able to do this, and over the years we’ve added more and more absorption.

It’s now a very “clear” sounding room, which makes it easy to work in, but lacks the “liveness” which I think makes for the most fun listening experience.

The biggest challenge in my little house is being limited by openings and lack of symmetrical set ups.

 

[Juhazi]

I heard MBLs at the recent show. Visited the (big) room briefly, 3 times. Liked them one time out of the three.

Wow, that's statistically significant, but not enough to be generalized!

 

[Juhazi]

b1daly, having an open and asymmetric room has some benefits too, regarding modes. Sure listening triangle setup is a compromise with normal living, but have you tried a diagonal setup? I have my dipoles on the long wall, but speakers and my ears are too close to front/back walls. However good distance to side walls helps for stereo image definition and balance and I can easily enhace imaging by dragging my chair a bit closer to speakers!

 

[Hephaestus]

I decided to pull the trigger with Acourate for my secondary system which consists of ROON, streamer and active studio monitors.
Results are truly amazing what comes to bass performance. I slightly reduced the treble as well which led to more "balanced sound" in my room.

What is clear is that some reading is recommended before doing Room EQ. I would have not succeeded without Floyd´s latest edition of Sound Reproduction and Mitch´s guide @ audiophile style.

 

[Bentoronto]

In the traditional AES point of view sound reproduction is like trying to make a realistic painting on canvas using physical descriptions of the applied paint. Or think of it this way, the tools in Photoshop manipulate certain visual features but these are only features which can be characterized by physical parameters not by the necessary perceptual parameters. As with hearing, it is true to say the physical parameters can fully describe and even perfectly reproduce the painting, as for example by making a digital copy.

When viewed through a knothole in a fence, the painting exhibits many of the cues of the real thing or a least as many as you’d get looking through a knothole. But as soon as you move your head or go closer to the painting or change the light, etc, etc. it is clear the “virtual image” of the pixels is certainly not the real thing.


Granted, some day there can be an artificial intelligence powered Photoshop, but that would be no different than a human brain and so would not prove anything.


But the AES approach is quite literally seeing the trees but missing the forest. Audio reproduction with stereo speakers is comparable to viewing a painting through a knothole. A handful of the cues are present and these support the percept. But as soon as you move or turn your head, the discrepant cues destroy the image. What is left really is, as psychologists say, “veridical”; it sounds just like two speakers playing a recording. But it doesn’t sound like Carnegie Hall.


At the highest level of perceptual processing, the effective cues in hearing are analogous to the “laws” of visual perception explored by the Gestalt school of psychology. Today, psychologist would call those laws “heuristics”. For example, in visual perception the laws include “good contours”, “familiar size”, and “similar fate”. It is easy to imagine the analogous heuristics in hearing.


The Gestalt heuristics should not be confused with the mechanism at the more peripheral or sensory level that do a kind of pre-reduction of stimuli. The Haas precedence effect, masking of various sorts, “deafness” for early reflections, and tools analogous to contrast enhancement which is so dominant in vision would possibly be examples.


This really helps clarify Toole’s re-examination of room equalization (and which, by the way, applies to misguided efforts at room treatment). The DSP “Photoshop” approach is to sit a mic where the listener would be and to diddle the knobs until there is a correspondence between the recording mic and one of the listener’s ears.


What Toole is struggling to say is that we hear by cues. The speaker is emerging as a hearing percept that our auditory system is trying to bring into consciousness as a coherent object. For stereo, we are beset with a peculiar synthetic set of cues that belong to a virtual image although those cues can never be all together too coherent because, well, it is just virtual.


There is no real difference between the emergence of a virtual stereo image and a real cowbell ringing in front of you in the same sense as there is no such thing as an “optical illusion”. The cowbell, the virtual image, and the percept arising from an “optical illusion” can theoretically be correct percepts. But unlike the cowbell, the stereo image and the “optical illusion” in a book can never have ALL the cues mutually supportive and cohering together. It is the role of audio professional to get enough of the cues working together to create a strong “illusion”.


Here’s an example. There is something that might be called “pinna perception” whereby the characteristics of the pinna and ear canal provide cues to the audible object but will be discrepant with a virtual image. A virtual sound object can’t possibly provide coherent pinna and canal stimuli either with speakers or even less so with headphones. The Bauer headphone correction is another example of cues that likewise contradict virtual percepts. Kunstkopf recording is a strategy to address these cues.


To the psychologist (as with the wannabee-psychologists in Toole’s lab), the speakers provide certain cues to their corporeal selves. They are embedded in a room which provides identify cues to its corporeal self. As Toole is trying to say, in human hearing perception, the room does not blend with the speakers - even if it does blend at the mic as a physical format. In the same way, my hand resting on the keyboard does not blend with the keyboard I am looking at although the image per se can be captured to a satisfactory degree in a digital still image.


I bet there are people wondering what is the difference between what the eye sees coming through the pupil and what the digital camera sees and the colour printer prints?


The answer is: nothing is different. The camera depicts my hand. So if your speaker can bring to your ears a satisfactory copy of the sound at the recording mic, aren’t you then enjoying perfect sound reproduction?


Think of it this way: can you see a camouflaged octopus on the sea floor? If I showed you a picture you would not see the octopus even though it really is plainly before your eyes. A psychologist could attempt a technical explanation that the Gestalt cues have been manipulated by the smart octopus in such a way as to make it imperceptible. Got that?


Similarly, the oboe in the orchestra is poorly perceptible on a reproduction system because of (a) the absence of all the cues you’d hear in real-life that support the percept of an oboe and (b) the presence of other cues that are at very obvious variance with thinking there’s an oboe ten feet from your nose.


Take a second to imagine how stupid it is to talk about bringing Carnegie Hall into your living room. Or for that matter, the “Down the Hall” criterion that I’ve talked about many times: have you ever thought a singer was really in your room EVEN when listening from down the hall. Even just a flute? Even just an old telephone with brass ringer? The failure of the Down the Hall Test is good evidence for the viewpoint Toole is trying to explain.


Wrapping up the outlook that Toole is struggling to clarify, the hearing system takes in a great profusion of stimuli. As Helmholtz taught, we produce an “unconscious inference” from that material using peripheral sensory mechanisms and central heuristics. In sound reproduction, the auditory system is trying to form perceptions of objects. These objects include the speakers, the room, and confusingly, some emerging percept of the recorded artist... but with no corporeal reality.

Audio professionals like to think they are creating the soundscape by introducing a virtual image into the room. But that doesn’t eliminate the existing landscape of the room or the corporeal reality of the speakers which are making the sound, no matter what new material is introduced.



(I leave it to somebody else to examine how the visuals in movies interact with the sound cues. Always amazes me how sound editors can produce sounds that are discrepant with the scene but that are rendered plausible and acceptable by the content of the screen image.)

Bentoronto

 

[GrimSurfer]

LOL

 

[Cosmik]

The DSP “Photoshop” approach is to sit a mic where the listener would be and to diddle the knobs until there is a correspondence between the recording mic and one of the listener’s ears.

But when you say correspondence, it is a correspondence in one dimension only: the steady state sine wave magnitude response. If they were really getting a strong correspondence in all dimensions it would be like listening to headphones but without having to wear any. No, what they strive for is the worst of all worlds: neither the sound of a neutral transducer relaying the recording into a real room, nor the sound of headphones.

I disagree with you about the futility of audio: stereo does something extraordinary, providing a stable image that, if done by the book (which includes the humble panpot), produces an authentic 'scene'. By listening to it in a room, the magic trick occurs: blending the listener's room with Carnegie Hall to produce a natural hybrid of the two. I think it is only because sticking two speakers at the end of a room is so obvious and easy, that people don't realise just what a brilliant result it gives.

People are determined that they know about the defects of stereo audio over speakers, just as they understand the impossibility of CD reproducing frequency content up to 20 kHz. But they're wrong.

 

[jazzendapus]

But the AES approach is quite literally seeing the trees but missing the forest. Audio reproduction with stereo speakers is comparable to viewing a painting through a knothole.

Depends on what you're trying to reproduce exactly. It is known to many here, and Toole wrote about this as well, that reproducing a whole orchestra playing in a particular 3D space using just two speakers is futile. But recorded audio is an artform in itself, musicians and their crew ultimately create a product that is to be played on two speakers. So it's not their actual performance that many audiophiles are trying to reproduce, but the recording of their performance that the musicians and their crew approved.

 

[Bentoronto]

stereo does something extraordinary, providing a stable image that, if done by the book (which includes the humble panpot), produces an authentic 'scene'.

Respectfully, I believe you are thinking of a stream of bits flowing out of a speaker but aren't thinking in terms of cues which are the building blocks of perceptual consciousness.

What in stereo-at-home is comparable to looking at a perspective Renaissance painting through a knot-hole? For the painting there are various discrepant cues such as your lens accommodation, surface texture, reflections, lots of stuff I can't remember, and not to mention the saints can hold their breath interminably and have halos.

Perhaps you have comparable minimized false cuing with an in-ear bud in one ear and your head held stationary with a dental bite-board. Even so constrained, perceptive readers can probably recognize a bunch of cues that are discrepant with putting your ear on a door to Carnegie Hall. And in any more a natural listener position would have a whole lot more discrepant cues.

There is no difference except by degree between the body of discrepant cues for a large orchestra and the body of discrepant cues for a cowbell. (I use a cowbell for a Live-vs-Recorded demo. No professional recording engineer should be without his or her own.)

It is within that narrow body of cues that Cosmik at the control panel can simulate (surrounded by revealing discrepant cues that diminish the stereo illusion) that the audio specialists work their magic. I take for granted that the great recording teams know which cues to bank on in order to make the DGG recordings of Michael Tilson Thomas luminous.... if never entirely veridical. Call it an art form.

 

[GrimSurfer]

Setting aside live trios, quartets, and symphony orchestras for a moment, has anyone considered the following:

At a live venue, the listener focuses intently on the sound (which, outside of the aforementioned fora, is also reproduced and played through an audio chain that has nowhere near the fidelity of most home systems). Why? Because if the listener misses something, it's gone forever. So they focus intently on the moment.

In a private setting, the intensity of focus is usually absent. The listener no longer focuses on the moment. At best, they sample the broader atmosphere... but not the moment. Why? Because recorded music can be played back.
This changes the listener interpretation of the sounds they're hearing. One listener adopts a form of hyper vigilance. The other is merely listening.

It's like the difference between a midnight stroll in your local park and a midnight patrol in a dangerous place. The sounds are comparable, perhaps even identical. The experiences are quite different.

All of this is to say that the perceived level of risk affects how sound is perceived and remembered.

 

[Bentoronto]

Although I haven't brought this up yet, Toole's critique of room EQ applies to the sound treatment of rooms the way I've applied it to speakers. You can take a lot of the sound out of a room but you can't take the room out of the sound... only re-shape it.

 

[GrimSurfer]

Although I haven't brought this up yet, Toole's critique of room EQ applies to the sound treatment of rooms the way I've applied it to speakers. You can take a lot of the sound out of a room but you can't take the room out of the sound... only re-shape it.

What does this mean?

 

[Bentoronto]

What does this mean?

Sorry if I seemed obscure. Let's say you have a belief about comb filters and so put up panels alongside the speakers to reduce that bounce. You've made a minor change in the room and so the room is little changed as a sound percept even though you're done something big to the output of the speakers.

If Toole is right, you've possibly confused your brain about the room and about the speaker output rather than improved the situation. I'm not dogmatic and the evidence of the effect of different room treatments on the listener's ability to create stable perceptions has to speak for itself.

 

[Cosmik]

Respectfully, I believe you are thinking of a stream of bits flowing out of a speaker but aren't thinking in terms of cues which are the building blocks of perceptual consciousness.

What in stereo-at-home is comparable to looking at a perspective Renaissance painting through a knot-hole? For the painting there are various discrepant cues such as your lens accommodation, surface texture, reflections, lots of stuff I can't remember, and not to mention the saints can hold their breath interminably and have halos.

Perhaps you have comparable minimized false cuing with an in-ear bud in one ear and your head held stationary with a dental bite-board. Even so constrained, perceptive readers can probably recognize a bunch of cues that are discrepant with putting your ear on a door to Carnegie Hall. And in any more a natural listener position would have a whole lot more discrepant cues.

There is no difference except by degree between the body of discrepant cues for a large orchestra and the body of discrepant cues for a cowbell. (I use a cowbell for a Live-vs-Recorded demo. No professional recording engineer should be without his or her own.)

It is within that narrow body of cues that Cosmik at the control panel can simulate (surrounded by revealing discrepant cues that diminish the stereo illusion) that the audio specialists work their magic. I take for granted that the great recording teams know which cues to bank on in order to make the DGG recordings of Michael Tilson Thomas luminous.... if never entirely veridical. Call it an art form.

On the one hand you say it is ridiculous to consider having Carnegie Hall in your living room and on the other you bemoan the lack of perfect simulation of the live event. The home audio system gives us the perfect hybrid.

Do you know what a timing discrepancy between the ears tells you about the location of a sound? It tells you only that it lies along a certain line radiating out from the listener. Do you know what stereo over speakers does with a volume difference between the channels, but no timing difference? It produces a timing difference between the ears. And do you know what it does when the listener moves or turns their head? It maintains a remarkably stable 'bearing' for the source.

Throw in some cues to do with reverberation related to distance, plus some extra 'real' room acoustics and you've got yourself a beautiful illusion that isn't static like a binaural recording (you do accept that a binaural recording is related to 'reality' but is unfortunately static?) and perhaps resembles the sound you would get if your room was transported to the recording venue with its end wall open.

I can only think you haven't heard a very good system!

 

[GrimSurfer]

Those changes are either measurable or not, regardless of whether they are perceived. Conversely, anything perceived can be measured (provided we know what to look for).

Now if we're talking about comb filtering, the sound panels would need to be between active speakers. This is where the effects of comb filtering mostly occurs, which explains why multiple driver arrays (and, in particular, multiple midrange drivers or tweeters) are so susceptible to it.

 

[GrimSurfer]

On the one hand you say it is ridiculous to consider having Carnegie Hall in your living room and on the other you bemoan the lack of perfect simulation of the live event. The home audio system gives us the perfect hybrid.

Do you know what a timing discrepancy between the ears tells you about the location of a sound? It tells you only that it lies along a certain line radiating out from the listener. Do you know what stereo over speakers does with a volume difference between the channels, but no timing difference? It produces a timing difference between the ears. And do you know what it does when the listener moves or turns their head? It maintains a remarkably stable 'bearing' for the source.

Throw in some cues to do with reverberation related to distance, plus some extra 'real' room acoustics and you've got yourself a beautiful illusion that isn't static like a binaural recording (you do accept that a binaural recording is related to 'reality' but is unfortunately static?) and perhaps resembles the sound you would get if your room was transported to the recording venue with its end wall open.

I can only think you haven't heard a very good system!

I agree. Good sound is good sound.

It might sound impressive to regale you with listening to a live performance at the Sydney Opera House but the acoustics there were (and still are) unimpressive. I've been in living rooms that, with the help of a good sound system, sound better.

 

[oivavoi]

But the AES approach is quite literally seeing the trees but missing the forest. Audio reproduction with stereo speakers is comparable to viewing a painting through a knothole. A handful of the cues are present and these support the percept. But as soon as you move or turn your head, the discrepant cues destroy the image. What is left really is, as psychologists say, “veridical”; it sounds just like two speakers playing a recording. But it doesn’t sound like Carnegie Hall.

Bentoronto

I think this is fundamentally correct, even though it's expressed in a complicated way. I'm not sure I would label the opposing view as the "AES view" - I have seen authors who publish in the AES say exactly the same thing. The issue is simply that what we think we hear is always an integration of many things - different senses, what we see and what we touch, and what we expect.

When listening in the home, there will be many many cues that work in the opposing direction from the phantom image the speakers try to create in my brain. I find that speakers that are visually striking in themselves remind me that I'm listening to speakers and not musicians, for example. I also find that it helps positioning the speakers with a open space behind them, as this helps my brain to create the illusion that something is happening there (as opposed to imagining that there is an orchestra playing inside a wall!). Among studio people, rooms built by a guy called Thomas Northward are in high demand at the moment. One of his principles is to position the speakers in a way which makes the sound engineer look at a open space behind a large glass window (or door), in order to help create the stereo illusion.

And there are other sources of tension between what the stereo image is trying to trick us into believing, and what our surroundings tell us. I don't think this tension can be solved completely, but one can get some of the way at least.

I'm not sure though what conclusion this leads to for room eq/correction... One may claim that one of the things room correction can do, is to create less tension between what the stereo image is telling us and what our surroundings are telling us. OR it may be the other way round, that trying to removing the room from the equation becomes artificial. Better just to try it out and see.

 

[Bentoronto]

Those changes are either measurable or not, regardless of whether they are perceived. Conversely, anything perceived can be measured (provided we know what to look for).

Ah, spoken like a true engineer.

Attached is a picture of a triangle (the one on top without the border). Everybody see it? Anybody know how to tell Photoshop how to find it? Well, that's the difference between perceptual heuristics* and a bit stream.

* I think the Gestalt psychologists of yesteryear might have called that the law of "good contour". I know the smart audio authorities out there might be wondering about the "illusion" of filled-in bass fundamentals, eh.

 

[Cosmik]

Ah, spoken like a true engineer.

Attached is a picture of a triangle (the one on top without the border). Everybody see it? Anybody know how to tell Photoshop how to find it? Well, that's the difference between perceptual heuristics* and a bit stream.

View attachment 28511

* I think the Gestalt psychologists of yesteryear might have called that the law of "good contour". I know the smart audio authorities out there might be wondering about the "illusion" of filled-in bass fundamentals, eh.

But you were confident that the illusion would show up on all of our PCs and monitors of varying quality under all the ambient conditions we were viewing them in.

So you make a crucial point: that the illusion is 'robust'. This is akin to the point I was making in another thread about music comprising perceived 'objects', where an object could be a triangle or a trombone, a violin, a violin section, an implied fundamental that isn't there, an orchestra, a concert hall, a sad chord, a crescendo, a tempo, etc. Our brains are perceiving music in terms of those objects, not dB, Hz and microseconds.

Those quantities can, however, be used to ensure that we hear the objects at their best and crucially with maximum separation (a CRT might have shown your triangle with pincushion distortion, and if its resolution was poor it wouldn't have been as clear; if it had a strange colour cast or was set up too dim or offensively bright it would get in the way of seeing the pure illusion; there would be many ways to smear the objects together and lose the clean separation between them).

The fact that we have not recorded, and are not reproducing, anything that measures the same as the original 'sound field' of the performance doesn't matter, because our brain locks onto the objects, and adapts to the circumstances. And yes, tricks have been used to make the recording perceptually closer to what we would have heard at the performance. We know what those tricks are and why they are necessary.

Misplaced 'room correction' is the equivalent of the monitor with strange colours, 'blooms' around stationary objects and trails behind moving objects. You might still see the objects, but they're all being imbued with an unnatural characteristic that blurs the separation between them and destroys the illusion of a 'reality'. But overall the screen combined with the ambient room reflections and selected calibration sequences gives an average of 'grey' when measured with a single photocell at the viewing position, so it must be accurate, of course.