Why do some Objectivists fear Psychoacousitics?

[microstrip]

In following up some links, I came across this quote from Toole's paper
I'm not sure how to read the bolded statement - anyone help in understanding this?

John,

You have to read the F. Toole book to fully understand it. He also addresses "realistic reproduction" in several sections. IMHO one of the most difficult challenges in interpreting short quotes of Toole writings is separating what refers to stereo or what refers to multichannel. Remember he wrote these and similar thing many times:

Old habits die hard. The introduction of stereo in the 1950s gave us an improved left/right soundstage, but close microphone methods, multitracking, and pan-potting, did nothing for a sense of envelopment—of actually being there. The classical music repertoire generally set a higher standard, having the advantage of the re?ectivity of a large performance space, but a pair of loudspeakers deployed at ±30° or less is not an optimum arrangement for generating strong perceptions of envelopment (as will be explained later, this needs additional sounds arriving from further to the sides). Perhaps that is why audiophiles have for decades experimented with different loudspeaker directivities (to excite more listening room re?ections), with electronic add-ons and more loudspeakers (to generate delayed sounds arriving from the sides and rear), and with other trinkets that seem capable only of exciting the imagination. All have been intended to contribute more of “something that was missing” from the stereo reproduction experience. The solution to this is more channels.

For decades, society has been conditioned to derive pleasure from ?rst single-channel sound (mono) and then two channels (stereo). Only recently has music been offered in multichannel formats that permit a somewhat realistic directional and spatial panorama. Impressed by the novelty that music and movies were available on demand, society appeared to lower its expectations and adapted to the inadequate formats.


The limitations of stereo were evidently inspirational to creative minds. (next paragraph seems to be an attack on the unnamed ESL63 electrostatic dipole...)
(all quotes from Sound Reproduction - my bold)

As I have written before, some people criticize Toole of underestimating and even misrepresenting the real potential of stereo and follow different routes in their developments.
And yes, stereo is living difficult moments in audio forums because its incapability of being a precise and predictable system, but also because some of the best of it became part of the luxury industry. However, as in many luxuries, we can not discard some real and unique improvements just because they are attached to the luxury label.

 

[jkeny]

Thanks, Micro & that does add to the perspective but can you interpret Toole's statement for me "High-quality stereo reproduction is compatible with those loudspeakers yielding high sound quality; however, there appears to be an inherent trade-off between the illusions of specific image localization and the sense of spatial involvement." as it reads to me that those speakers yielding high sound quality are traded off against image localisation & spatial involvement but this can't be so what do you read it to mean?

 

[microstrip]

I do not have access to the full paper, but my guess is that he is not addressing the speakers, but the stereo format. He wrote similar comments on consumer and professional preferences and needs of absorption in listening rooms. The trade off is between precise image localisation (pro's) & spatial involvement (consumers, audiophiles included )

The book has more specific references, but this the one I could find now:
Memo for Listening room recommendations: for stereo listening, leave side walls reflective at first-reflection points. For multichannel listening it is optional. Audio professionals may have their own preferences—it’s all right, they are just different.


As far as I remember it comes from sections on apparent source width (ASW) and listener envelopment (LEV).

 

[jkeny]

Thanks Micro, I think I understand what Toole is saying now & as you say it's not about the speakers themselves:
- specific image localisation means eliminating first reflections - of more concern to the pro industry
- spatial involvement which requires reflections - is preferred by consumers

 

[amirm]

Cross posting this from similar thread on another forum:

-------

I'm not sure how to read the bolded statement - anyone help in understanding this?

Eyleron essentially answered this but to put in my two cents , the issue was whether speaker directivity could be quantified with respect to listener preference. Study cited showed that while recording engineers used wide dispersion speakers at home, half of them surveyed did not want to use the same for their work. Instead they preferred focused, narrow angle radiation.

For most listeners, this does not hold as a) we are not performing mix decisions and b) we have not developed the preference for point source reproduction (indeed for such applications as center channel for movies such point source reproduction is "wrong" as that channel needs to span the full video screen).

 

[amirm]

Indeed & that is important but is it the MOST or ONLY important aspect - that is my question!

It is the most important thing. Not the only one but definitely the most important one. As I explained colorations are very easy to hear, and are not subject to such things as masking.

 

[esldude]

Thinking a bit on this stereo vs mono auditioning of speakers question. Two speakers always causes some comb filtering. Miking technique may or may not. That alone would allow one to more accurately access mono speakers vs stereo. And might mean some blind tests of some parameters would be better in mono as well. Why would you demote a speaker when it was effected by comb filtering caused by two speakers playing that might obscure most other characteristics? And as someone else pointed out already, in real life most sounds, nearly all sounds are mono. There are many of them, but each is pretty much mono.

Then again, maybe it is like a car a high school friend purchased once. It had 18 speakers scattered within it. Stereo,but well before the days of any really good audio equipment for cars. It was a weird group of mismatched speakers of many sorts. The fellows theory: all speakers are uneven, and sound different. The more of them you use together the more it averages out to a really good even result. Didn't sound that way to me. It was LOUD. Sounded like a horrid indistinct mess. Even in mono.

 

[jkeny]

Yes, agreed, it might be easier to hear issues when using a mono speaker (& avoiding room boundary reflections) but it is not the way we normally listen to our playback systems.

But is this not the exact opposite to the argument I've seen you use to question Amir's ABX test results - you suspected he was using an unnaturally high volume & listening to the background noise differences i.e not normal listening conditions?

 

[esldude]

Yes, agreed, it might be easier to hear issues when using a mono speaker (& avoiding room boundary reflections) but it is not the way we normally listen to our playback systems.

But is this not the exact opposite to the argument I've seen you use to question Amir's ABX test results - you suspected he was using an unnaturally high volume & listening to the background noise differences i.e not normal listening conditions?

It isn't the opposite. They were different purposes. I can take two music files and boost them 50 db. Hear some differences clearly though it is a low level difference. Drop them back to their natural level and those differences are down near -100 db. Do they sound the same or not? In normal use, yes they sound the same. So using a boosted example to say they sound different makes little sense.

On the other hand, were I investigating some low level phenomna or altering a design to make it more linear, sure that 50 db boost makes sense.

In the same way if using mono speakers gets more discernment than stereo for design purposes it makes plenty of sense. If you get two designs that test very closely in mono with one winning out, and if that difference is obscured in stereo, then maybe saying they are different in their normal use isn't a sensible thing to say.

In the contentious debate about whether short term testing is more discriminating or not, much evidence says short term quickly switched comparisons are most discerning. The idea being if you can get things indistinguishable in short immediate comparisons you can be sure it is really audibly the same. Obviously this isn't how we normally listen. Normal listening might well be so much less sensitive there is no real difference in some things detected in rapid switching. Carrying out the test in a rapidly switched manner provides a safety margin of sorts. (please note I don't wish to debate in this thread short vs long term listening. If you don't agree, just take my description as a suppostion to follow the logic of it)

 

[jkeny]

It isn't the opposite. They were different purposes. I can take two music files and boost them 50 db. Hear some differences clearly though it is a low level difference. Drop them back to their natural level and those differences are down near -100 db. Do they sound the same or not? In normal use, yes they sound the same. So using a boosted example to say they sound different makes little sense.

On the other hand, were I investigating some low level phenomna or altering a design to make it more linear, sure that 50 db boost makes sense.

In the same way if using mono speakers gets more discernment than stereo for design purposes it makes plenty of sense. If you get two designs that test very closely in mono with one winning out, and if that difference is obscured in stereo, then maybe saying they are different in their normal use isn't a sensible thing to say.

Well that's one aspect of what I'm saying - their tests are far enough divorced from normal usage (they avoid stereo & room boundary reflections) that the applicability to our "normal listening" is tenuous, at best.

The phrase that comes to mind is one that objectivists often cite "every measureable difference is not necessarily audible" in this case what's heard in mono may not be of great importance in stereo listening or there may be aspects of stereo listening that are perceptually more important than what's heard in mono.

As regards the rest of your analysis:
You suggested Amir was gaming the system by listening at an unnaturally high volume (btw, I don't accept this was the case) & hence you dismissed his positive ABX results as him gaming the system. As I said I very much doubt he was doing this but let's say he was - is this not a valid way of differentiating the differences? It shows that there are actual audible differences but your contention that they are they are only apparent in an abnormal listening situation (high volume).

On the other hand you agree with Harmon's use of one speaker (an abnormal usage) to better differentiate between speakers even though this may not be discernible in normal listening. They further compound this by avoiding room boundary reflections (or minimise them) - another unnatural usage.

BTW, I will agree that listening loud Vs listening in mono is different in one important aspect - in listening loud we are not removing any part of the signal as we are by doing so in mono (& without room boundary reflections)

In the contentious debate about whether short term testing is more discriminating or not, much evidence says short term quickly switched comparisons are most discerning. The idea being if you can get things indistinguishable in short immediate comparisons you can be sure it is really audibly the same.

Yes, if you can remain within the time window for auditory echoic memory & you can eliminate all other psychological factors that influence the test. As this second condition is impossible to meet, the only answer is to use statistical analysis of the results which requires a large sample set for confidence. This usually requires many repetitions of the same quick A/B trial which introduces another factor - that of fatigue/loss of focus. Internal controls are necessary to test for this but seldom used. Even if all these criteria are met then the quick A/B listening will only reveal elements in the soundfield which are amenable to quick A/B comparison. These elements do not encompass all the differences between two soundfield. As a result most DBTs are exercises in belief.

Obviously this isn't how we normally listen. Normal listening might well be so much less sensitive there is no real difference in some things detected in rapid switching. Carrying out the test in a rapidly switched manner provides a safety margin of sorts. (please note I don't wish to debate in this thread short vs long term listening. If you don't agree, just take my description as a suppostion to follow the logic of it)

Well, I have followed your logic & answered with mine. BTW, this goes to the heart of the O/P - psychoacoustics, so I don't know why you wish not to discuss it?

 

[Whatmore]

choose the better testing methodology. This is a "blind" quiz, I won't tell you which is which

1. Makes little or no attempt to remove known biases. Almost always results in false positives.
2. Succeeds at removing most known biases. May, very occasionally produce false negatives which are most likely random rather than systematic.

 

[jkeny]

choose the better testing methodology. This is a "blind" quiz, I won't tell you which is which

1. Makes little or no attempt to remove known biases. Almost always results in false positives.

Do you have evidence of this in the audio field that will stand up to scrutiny i.e not the Harmon test above - that long term sighted listening almost always returns false positives?

2. Succeeds at removing most known biases. May, very occasionally produce false negatives which are most likely random rather than systematic.

If "most known biases" are removed by the simple job of removing knowledge, why would the known ITU, MUSHRA standards go to great lengths to ensure the removal of known biases? One of the recommendations in these standards is to use internal controls, specifically to identify false negatives (& false positives).

Do you have evidence of the percentage of tests that return false negatives & what level of false negatives have been found in each test? Or are you just stating your blind belief, which as I said, is the usual accompaniment to blind tests?

 

[esldude]

Near the entirety of your post is being ignored though it was read and considered....................

Well, I have followed your logic & answered with my. BTW, this goes to the heart of the O/P - psychoacoustics, so I don't know why you wish not to discuss it?

Your supposition is flawed at its premise. Firstly by saying objectivists don't wish to discuss psychoacoustics. You don't seem to realize what that encompasses. It includes very basic things about what is audible by the human ear as well as going onto auditory scene analysis. You wish to skip the more basic psychoacoustics to jump to an interpretation of auditory scene analysis that I find not only strange, but one that would appear to contradict the more elementary elements of psychoacoustics.

PS-I will note I did not dismiss Amir's results. Amir's own comments make me think he was listening at an elevated volume. The idea a difference is apparrent at elevated volumes and not at lower volumes is not controversial. The rest of your conclusions don't follow from what I wrote.

 

[jkeny]

Your supposition is flawed at its premise. Firstly by saying objectivists don't wish to discuss psychoacoustics. You don't seem to realize what that encompasses. It includes very basic things about what is audible by the human ear as well as going onto auditory scene analysis. You wish to skip the more basic psychoacoustics to jump to an interpretation of auditory scene analysis that I find not only strange, but one that would appear to contradict the more elementary elements of psychoacoustics.

Hmmm, I think you are confused between acoustics & psychoacoustics. I seem to have to repeat myself. I already said that JNDs need to be revisited as they have been based on simplistic test signals - signals that ignore what we now know about psychoacoustics.

PS-I will note I did not dismiss Amir's results. Amir's own comments make me think he was listening at an elevated volume. The idea a difference is apparrent at elevated volumes and not at lower volumes is not controversial. The rest of your conclusions don't follow from what I wrote.

OK, lets clear this up as I believe what you said on another forum to be dismissive of Amir's results:

After I resampled with new software, the remainder of the difference in the two files is way, way down. It isn't nothing, but is around -90 db mostly due to dithering. There is just close to nothing there. However, Amir commented at some point he was listening to something like .4 or .5 seconds in one segment. And accidentally skipped to another area which was dangerously loud. That sounds to me like he picked a segment decaying to silence, and jacked the level up to max. That can let you hear a bit of noise difference in the noise floor. You wouldn't hear that without the unnaturally high volume. So I am not saying Amir is lying. I guess I would say he is gaming the system to prove it can be gamed and isn't a reliable determination of audibility effects. That fits with a wide swath of his behavior and opinions on the matter.

 

[esldude]

OK, lets clear this up as I believe what you said on another forum to be dismissive of Amir's results:

Yes, I posted that on another forum. It doesn't contradict what I posted earlier. He can answer what he did if he wishes. He already has posted his point was to show how casually performed blind tests should not be used to brow beat people about what is or is not audible. That his point in the other thread here was about what constitutes proof one way or the other.

 

[Whatmore]

Do you have evidence of this in the audio field that will stand up to scrutiny i.e not the Harmon test above - that long term sighted listening almost always returns false positives? If "most known biases" are removed by the simple job of removing knowledge, why would the known ITU, MUSHRA standards go to great lengths to ensure the removal of known biases? One of the recommendations in these standards is to use internal controls, specifically to identify false negatives (& false positives).

Do you have evidence of the percentage of tests that return false negatives & what level of false negatives have been found in each test? Or are you just stating your blind belief, which as I said, is the usual accompaniment to blind tests?

John, you are the one who repeatedly dismisses blind tests on the grounds of false negatives. Perhaps you could furnish us with some evidence of this?

 

[jkeny]

John, you are the one who repeatedly dismisses blind tests on the grounds of false negatives. Perhaps you could furnish us with some evidence of this?

I'm saying that without some measure of how prone the test is to false negatives, we have no way of evaluating the validity of the results - it's about rigour of the test - we simply don't know what percentage of the results are false negatives.
If you disagree with that, then please tell me why?

 

[Whatmore]

I'm saying that without some measure of how prone the test is to false negatives, we have no way of evaluating the validity of the results - it's about rigour of the test - we simply don't know what percentage of the results are false negatives.
If you disagree with that, then please tell me why?

How valid are sighted tests? Can you provide some rigour around the incidence of false positives?
If you can't, then sighted tests are not valid.

If you disagree, please tell me why?

 

[jkeny]

How valid are sighted tests? Can you provide some rigour around the incidence of false positives?
If you can't, then sighted tests are not valid.

If you disagree, please tell me why?

But I don't insist on sighted tests to "prove" anything but many objectivists insist on DBTs.
Even when a valid, rigorous DBT is run, (& not many are), I find it only revealing of some elements in the soundfield
I find that long term listening to be more revealing of the other elements - the ones I consider more important - things like the illusion of realism of the sound. This "realism" illusion won't be so evident if the distortions, frequency/amplitude anomalies, are present - the ones DBTs are directed towards.
So, to my mind, long term listening are more revealing than quick A/B style DBTs & less prone to error

What I'm suggesting is the inclusion of internal controls in DBTs - it's not my idea, it's recommended in the published standards for conducting DBTs. I'm just highlighting why it's needed & the lack of such controls in almost all BTs I've seen.
If you or Tomelex's wish to argue that these controls are not necessary then do so.
Don't distract by talking about another test - defend the test you support or admit that it's flawed & do something about it.

Is this objectivity of objctivists in action?

 

[jkeny]

Harmon did a test, and published it, to this day I have not seen a published sighted test with any where near the controls.

Guess for now, a test without complete controls is hands down winner in this scenario. And apparently the scientific community nor all the audio manufacturers and distributors are not about to conduct a sighted test with controls as their whole selling scenario depends on how each year, year after year, decade after decade, they audibly improve their kit and its audibly better and more expensive than the last,

and hundreds of manufacturers do this in audio, and no one has tested this whole sighted live with it for a while to know how it sounds theory.......pull the other one it has bells attached.

But I am not against psychoacoustics but don't try to make two channel stereo equipment out to be the way out of the woods, it will take another approach, perhaps all your danged walls need to be turned into speakers then with the appropriate input you will have the concert transported to you, but two channel stereo, its like trying to make orange juice from apples. ahahahah

Yes, you heard me right, two channel stereo can go no further than the tricks the mix and master engineer already play. As long as you have two channels and two areas you radiate sound from, stereo is at the mercy of its two channels. They are allright, but they like horse drawn wagons to airplanes, they are pitifully inadequate and out of date for what audiophiles who claim that they want to hear the real things in their home are trying to get. It just can not do it, can we play some hickery jiggery in the mix and master process and play with ambiosonics etc, yes, but still go those two areas you radiate from that put the absolute limits on this particular technology.

Yes, tomelex, I know your message, you state it every time you post.
But tell me how you will develop a new replay system that is more realistic if you don''t know & apply what we know of (& continue to gain knowledge in) psychoacoustics.
I propose, like all good objectivists, that we start with the easy stuff - 2 channel stereo & see where we get with that before we move on to more difficult