Lack of high-end speaker reviews

[MarkS]

If you have the book, can you check how many people participated in this study? Thanks. [Note: I have been lurking for a long time, but this is my first post.]

From the book:

Forty listeners participated in a test of their abilities to maintain objectivity in the face of visible information about products. All were Harman employees, so brand loyalty would be a bias in the sighted tests. They were about equally divided between experienced listeners, those who had previously participated in controlled listening tests, and inexperienced, those who had not.

 

[MarkS]

I would like to comment on Olive & Toole's listening data, and what can be reliably concluded from it, and what cannot.

Here again is the key figure from Toole's book. I have labeled the four different speakers with red letters to make it easier to refer to them:

Note the error bars. Per the figure's caption, the error bars show one side of a 95% confidence interval.

What does that mean? Look at the blind result for Speaker A on the far left. The height of the bar is at 6.75, and the error bar at 7.0 (approximately: I am eyeballing the chart to get the numbers, which are not listed in the book). This means that, if we were able to do the experiment with millions of listeners, the probability is 95% that the blind rating of Speaker A would be in the range from 6.5 to 7.0. The height of the bar, 6.75, is in the middle of this range.

In comparing two different bars, to have 95% probability that the higher bar would remain higher with millions of similar listeners, the difference in their heights should be greater than 1.2 times the one-sided error (taking it to be same for both bars, which is true here to within eyeball estimates, and is 0.25; times 1.2 is 0.3).

Looking at the blind results, the data do not show a statistically significant difference (at the 95% confidence level) between A and B. The same is true for C and D. However, A and B are preferred over C and D (at the 95% CL).

Looking at the sighted results, the data do not show a statistically significant difference (at the 95% CL) between A and B, same as the blind case. In contrast to the blind case, D is now preferred over C (at the 95% CL), and, as in the blind case, A and B are preferred to D (at the 95% CL).

So: the only significant change from blind to sighted in the downgrading of C.

A, B, and D were all large and expensive floorstanders that were "stylishly crafted of highly polished wood", whereas C was a "small sub/sat system" that was "made of plastic". And these tests were done in the early 90s, when sub/sat systems for stereo were a relatively new thing. Today every audiophile knows that excellent results from such a system are possible; 30 years ago, this was not the case.

So "sighted bias" only significantly changed the results for a comparatively cheap looking, inadequate seeming (in those days) system. "Sighted bias" did not significantly change the results for similar looking speakers.

This is the result of the Olive-Toole research.

 

[Blumlein 88]

I would like to comment on Olive & Toole's listening data, and what can be reliably concluded from it, and what cannot.

Here again is the key figure from Toole's book. I have labeled the four different speakers with red letters to make it easier to refer to them:

View attachment 342944

Note the error bars. Per the figure's caption, the error bars show one side of a 95% confidence interval.

What does that mean? Look at the blind result for Speaker A on the far left. The height of the bar is at 6.75, and the error bar at 7.0 (approximately: I am eyeballing the chart to get the numbers, which are not listed in the book). This means that, if we were able to do the experiment with millions of listeners, the probability is 95% that the blind rating of Speaker A would be in the range from 6.5 to 7.0. The height of the bar, 6.75, is in the middle of this range.

In comparing two different bars, to have 95% probability that the higher bar would remain higher with millions of similar listeners, the difference in their heights should be greater than 1.2 times the one-sided error (taking it to be same for both bars, which is true here to within eyeball estimates, and is 0.25; times 1.2 is 0.3).

Looking at the blind results, the data do not show a statistically significant difference (at the 95% confidence level) between A and B. The same is true for C and D. However, A and B are preferred over C and D (at the 95% CL).

Looking at the sighted results, the data do not show a statistically significant difference (at the 95% CL) between A and B, same as the blind case. In contrast to the blind case, D is now preferred over C (at the 95% CL), and, as in the blind case, A and B are preferred to D (at the 95% CL).

So: the only significant change from blind to sighted in the downgrading of C.

A, B, and D were all large and expensive floorstanders that were "stylishly crafted of highly polished wood", whereas C was a "small sub/sat system" that was "made of plastic". And these tests were done in the early 90s, when sub/sat systems for stereo were a relatively new thing. Today every audiophile knows that excellent results from such a system are possible; 30 years ago, this was not the case.

So "sighted bias" only significantly changed the results for a comparatively cheap looking, inadequate seeming (in those days) system. "Sighted bias" did not significantly change the results for similar looking speakers.

This is the result of the Olive-Toole research.

However, A and B apparently had the largest sighted influence. That they also had the best sound rather clouds the issue. Speakers with best sound win blind, and that was A and B. Speakers with the best appearance win sighted and that was also A and B. You cannot reasonably conclude from this that sighted listening worked almost as well and blind listening.

 

[MarkS]

Earlier in this thread you wrote:

Most of the data from Toole and Olive's testing say ... people rank very highly from sighted listening things they rank very poorly unsighted.

I assert, in the strongest possible terms, that the data do not support this claim.

 

[Axo1989]

However, A and B apparently had the largest sighted influence. That they also had the best sound rather clouds the issue. Speakers with best sound win blind, and that was A and B. Speakers with the best appearance win sighted and that was also A and B. You cannot reasonably conclude from this that sighted listening worked almost as well and blind listening.

This is a good counterpoint, we would likely need more tests to disaggregate the influences of aesthetics and sonics. Were there any photos of that selection?

 

[Sokel]

This is a good counterpoint, we would likely need more tests to disaggregate the influences of aesthetics and sonics. Were there any photos of that selection?

What's evident about this test is not about the first three speakers but about the last (D) .
It wins the third one by a big margin sighted but looses lousily unsighted.(American pride maybe? )

 

[Axo1989]

I would like to comment on Olive & Toole's listening data, and what can be reliably concluded from it, and what cannot.

Here again is the key figure from Toole's book. I have labeled the four different speakers with red letters to make it easier to refer to them:

View attachment 342944

Note the error bars. Per the figure's caption, the error bars show one side of a 95% confidence interval.

What does that mean? Look at the blind result for Speaker A on the far left. The height of the bar is at 6.75, and the error bar at 7.0 (approximately: I am eyeballing the chart to get the numbers, which are not listed in the book). This means that, if we were able to do the experiment with millions of listeners, the probability is 95% that the blind rating of Speaker A would be in the range from 6.5 to 7.0. The height of the bar, 6.75, is in the middle of this range.

In comparing two different bars, to have 95% probability that the higher bar would remain higher with millions of similar listeners, the difference in their heights should be greater than 1.2 times the one-sided error (taking it to be same for both bars, which is true here to within eyeball estimates, and is 0.25; times 1.2 is 0.3).

Looking at the blind results, the data do not show a statistically significant difference (at the 95% confidence level) between A and B. The same is true for C and D. However, A and B are preferred over C and D (at the 95% CL).

Looking at the sighted results, the data do not show a statistically significant difference (at the 95% CL) between A and B, same as the blind case. In contrast to the blind case, D is now preferred over C (at the 95% CL), and, as in the blind case, A and B are preferred to D (at the 95% CL).

So: the only significant change from blind to sighted in the downgrading of C.

A, B, and D were all large and expensive floorstanders that were "stylishly crafted of highly polished wood", whereas C was a "small sub/sat system" that was "made of plastic". And these tests were done in the early 90s, when sub/sat systems for stereo were a relatively new thing. Today every audiophile knows that excellent results from such a system are possible; 30 years ago, this was not the case.

So "sighted bias" only significantly changed the results for a comparatively cheap looking, inadequate seeming (in those days) system. "Sighted bias" did not significantly change the results for similar looking speakers.

This is the result of the Olive-Toole research.

There is a tendency to oversimplify these and similar tests, and create dogmas—or maybe just rules of thumb—from them.

The other example often used is the stereo vs mono test summary graph, where the QUAD speakers scored low in the SQ summary data (but not across all the usually ignored genre category results) in mono but similar to others in stereo. But we'd want to see more dipoles tested to see if that result is more a function of speaker type rather than stereo dramatically obscuring poor sonics.

 

[MattHooper]

However, A and B apparently had the largest sighted influence. That they also had the best sound rather clouds the issue. Speakers with best sound win blind, and that was A and B. Speakers with the best appearance win sighted and that was also A and B. You cannot reasonably conclude from this that sighted listening worked almost as well and blind listening.

Conversely, that would seem to mitigate against being able to draw confident conclusions in the opposite direction, wouldn't it?

If both A and B were rated higher than others in both sighted and blind listening, then while you could conclude sighted listening had *some* influence in raising their ratings, you couldn't conclude that the actual sound wasn't a factor in sighted listening. Since the blind testing also verifies the sound was indeed better, then that has to be considered a possible factor causing the consistent higher rating as well. Such a test it seems did not untangle that variable.

In other words, it still seems some caution is reasonable in how broad one's conclusions can be from that test.

 

[MattHooper]

I think it mostly boils down to one's desire to have their opinions led either by the "baggage" of various biases and preconceptions, or to have them led by an honest and (mostly) unbiased/uncolored evaluation of...whatever it is; in this case, of course, speakers.

Sure, that could be an attitude one adopts, but with respect, it doesn't answer the issue I raised. If you aren't going to actually perceive the sound correctly in your sighted conditions, then it seems to make it simply an academic exercise in choosing a speaker - something that "measured well" but which you'll never really appreciate sonically.

But of course, that's not actually how people think here: they presume that measurements will indicate how a speaker will sound to them, not just in blind conditions, but in the sighted conditions they'll employ the speaker. And it leaves open the question, if the blind listening data and speaker measurements will predict what one hears in sighted conditions, what would explain that, except that sighted conditions are sufficient to, at least to a relevant degree, allow some accuracy in what we hear?

Anyway, this is reminding me of a speaker audition experience I had ages ago. Back in the early 90s, my dad and I were wandering through a local town and decided to stop into the Bose store. Almost immediately, an associate grabbed our attention and asked if we wanted to attend a brief speaker demo that was about to begin. We agreed and were led into an isolated room that contained several rows of chairs as well as a dozen or so other customers. Shortly after we entered, everyone took a seat, and we were greeted with some fairly large speakers hanging from the front wall. The associate proceeded to spin up a handful of music tracks, which were played quite loudly. I can't attest to the speakers' tonality (or lack thereof) at this point, but I remember thinking that the speakers sounded suitably large and lively and were enjoyable to listen to. After a couple of minutes, with music still playing, the associate stepped up to the front of the room and began removing the speakers from the wall. It was at this point that we all realized that these weren't speakers at all, but simply empty enclosures that had been camouflaging the real speakers, which turned out to be little bitty Bose Cubes. Everyone predictably ooh'd and ahh'd and shortly after that the demo was over.

Eventually my dad and I left the store and resumed our journey around town. A few hours later, I returned to the store. I wanted to listen to the demo again, this time armed with the knowledge of what was really going on. The same demo tracks were played again at presumably the same volume, and, suddenly, I could "hear" the smallness of the true speakers. They sounded more strained and distorted now. Did anything really change, other than my own biases and preconceptions? Nope. I might have happily taken these speakers home during the first audition, not so much the second.

Excellent anecdote! I've had plenty like it. It's one reason why I always close my eyes when auditioning speakers, or also at audio shows. I've been in front of some very tall loudspeakers but when I closed my eyes they sounded tiny! Like they could have been a boombox!

 

[Blumlein 88]

Earlier in this thread you wrote:

I assert, in the strongest possible terms, that the data do not support this claim.

From that one test it might. The speaker with the worst appearance did the worst in sighted. It had better sound quality than the good in appearance US speaker. In did better in blind testing vs the US speaker. My point is this one test doesn't clearly support either one. It is not the only test they've done.

 

[Blumlein 88]

Conversely, that would seem to mitigate against being able to draw confident conclusions in the opposite direction, wouldn't it?

If both A and B were rated higher than others in both sighted and blind listening, then while you could conclude sighted listening had *some* influence in raising their ratings, you couldn't conclude that the actual sound wasn't a factor in sighted listening. Since the blind testing also verifies the sound was indeed better, then that has to be considered a possible factor causing the consistent higher rating as well. Such a test it seems did not untangle that variable.

In other words, it still seems some caution is reasonable in how broad one's conclusions can be from that test.

You couldn't conclude sound was no factor, but that is about all. Of course this is not the only test they did. This one test would not untangle that variable, but many other tests were done. What is reasonable is to look at the total results with many speakers.

 

[MarkS]

It had better sound quality than the good in appearance "American" speaker.

The difference was well within the statistical error on the rankings.

To their great credit, Olive & Toole published these error estimates. We should not draw conclusions that fail to take them into account.

 

[Blumlein 88]

Maybe worth reading this short little article by Kal Rubinson about his experience and thoughts having visited Harman and done the test for himself.

Blind Listening at Harman International | Stereophile.com

Revel began demonstrating prototypes of the Performa F228Be, reviewed elsewhere in this issue, at the Consumer Electronics Show and other audio events in 2015, though they had yet to settle on the model name.

www.stereophile.com

His concluding paragraph:

I came home with ears and mind refreshed, and certain that any thinking audiophile would benefit from the experience. That said, from now on, good-looking Spinorama graphs or similar data (footnote 3) will be major criteria in my considering what I'll review—and what I'll buy.—Kalman Rubinson

 

[Galliardist]

Earlier in this thread you wrote:

I assert, in the strongest possible terms, that the data do not support this claim.

But the data (not just the one graph, and not just speaker testing) does clearly support the assertion that sighted and blind tests give different results for the same sound waves. This at least indicates that a property of the listeners is involved in the differences shown in the two types of test.

I'm not sure that the causes of that difference has ever been properly investigated. Nor is there a great deal in general literature on hearing from which we can deduce even a good working hypothesis. "Biases" - I hear the cry, but then what causes which of the long list of biases regularly produced here to affect what result?

It appears that to answer all of the assertions of both sides, we need a longitudinal qualitative study backed up by good quantitative measurements as a framework for understanding the qualitative data. I'm not aware that anybody would ever fund such a study, or that anyone who did do the study would ever publish their results.

So we have to stick with what we've got for now and, I guess we can keep arguing for ever about this.

 

[MattHooper]

As I say, if you push the skepticism too far, you reach incoherence (and you start to losing explanatory power)."

I don't know if you do lose explanatory power. You lose any practical way to choose speakers. I don't think you'll get far telling people they should not listen to speakers instead buy only by the specs or you are foolish. At least for the time being we are not in a place where skepticism is too much, rather that it is not enough in general. I would say while there is some accuracy for sighted listening it is more limited than most think.

What I mean about losing explanatory power through hyper skepticism:

To go back to an example I've used: If a scientist spends most of his time focused on researching optical illusions, that could influence him to take a dim view on the reliability of our sight. We develop biases. But to push the point: Imagine if that scientist took the skepticism so far that he declared, on the basis of optical illusions, that this applies to all conditions - that is our sight is wholly unreliable and inaccurate. Everyone would recognize he has leveraged the data to spread far too broadly. That's going to face huge explanatory challenges with that type of hypothesis: how will he explain people passing eye exams? Reading? Driving cars? The endless examples of people clearly navigating the world successfully using sight? If he doesn't recognize the relevance of those examples, he's not a very responsible scientist. Any comprehensive view of our vision will have to acknowledge and account for the successes as well as the failures, and explore the gray areas in between, where we will have to make lower confidence judgments, yet still justified in a practical sense.

The same goes for our perception in regard to hearing. Skeptical claims about our perception in sighted conditions are not, as some have seemed to imagine here, hermetically sealed as to the implications for audio - they rely on fundamental (and true!) issues of bias in hearing perception. You can have bias effects, of course outside of just judging loudspeakers. By the same token, pushing the skepticism of sighted listening too far, and you start running in to the type of explanatory problems I've raised before. At some point you have to start explaining all the apparent successes of our hearing in non-laboratory/uncontrolled conditions. Just as people successfully navigate the world and communicate via sight, they do so with hearing perception. Whether it's every day life, or music/record production, or whatever. (As I've said before, my work in post production sound actually becomes both irrational, and it's success inscrutable, if sighted skepticism is pushed to far). We seem to get along pretty darned well appealing to hearing differences outside of the "lab," and at some point saying "It's all bias effects" is going to have pretty poor explanatory power vs the assumption people are often "really hearing what they think they hear, and really communicating truths about what they hear." There's tons of noise in our perceptual and cognitive system, but we've evolved to ultimately rise above lots of the noise to manage some significant level of reliability.

And so, if we take examples where sighted listening may be totally unreliable (I dunno, maybe when there are little to no actual sonic differences) on the other end of the spectrum will be sighted perception that is very reliable (e.g. test if I can tell the difference between my mother's voice and Trump's voice, and the reliability under both sighted and blind conditions are likely to be a match). So then we have to keep both ends in view, and realize as we trace from each side towards the centre, the least reliable and the most reliable, as they converge we'll be hitting gray or fuzzy areas in which we start to make judgments "do we NEED blind tests to have ANY justification for a conclusion in X case?" Where, given the impracticality of putting all those scenarios to the test, you adopt lower-than-scientific confidence levels, but with the proper caveats, still justified.

The question then is what, in audio, might fit in that gray area: where we aren't putting our perception to scientific rigour, it's just not practical, and yet allowing for lower confidence levels, we can justify provisional conclusions.

For myself, I would put, say, owning and comparing loudspeakers in that category. If I've owned a loudspeaker, had plenty of time playing with positions to see what it can do, I've played tons of familiar music, and then I own another speaker too and go through the same efforts, I think I'm justified enough that I"m getting some real picture of the sound, and it's not utterly obscured by bias effects. The fact that what I think I hear is also very plausible given the design of each speaker adds to the confidence. It COULD be that I'm wrong, but that caveat is built in to my conclusion. I'm being practical. I would push that to auditioning speakers in stores too. That kind of thing. Others may not.

Think about what it would mean if we could figure out the best way the speaker should perform. I can now buy DACs, amps, most anything not a transducer without worrying about anything other than features, cost, appearance and reliability. Transparent sound quality is a given. If speakers should be developed to that level, there might be people who just want something different. Like SETs still have a market and are definitively not transparent or accurate. So what would you think if speakers as dependably perfect as amps were available. You wouldn't listen not because there is no accuracy to doing so, but because the component accuracy is beyond us and there is no need. Look at how much trouble we have convincing people the electronics don't sound different.

Yeah, (assuming for sake of argument tube amps can sound different)...that's why tube amps will likely always be around. The ever more amazing specs for solid state gear don't matter, because we've effectively had sonic transparency for a long time. We've heard it. And still, some audiophiles like something extra added in, which the transparency does not provide.

As for the speakers question, it depends on the approach. Some people's approach is that they want to "not hear" the equipment. They just want it out of the way. Other people enjoy playing with sound. Being creative with it. That's their fun. Achieving a certain "sound" they have in their mind. And that's an entirely legitimate approach too (I'd say that, because it's at least partially my approach). It's what many of the guys playing with tube amps and horns and all that are doing (or at least think they are doing).
So I'm not sure a perfectly transparent speaker will obviate other options for the latter crowd.

Cheers.

 

[Blumlein 88]

What I mean about losing explanatory power through hyper skepticism:

To go back to an example I've used: If a scientist spends most of his time focused on researching optical illusions, that could influence him to take a dim view on the reliability of our sight. We develop biases. But to push the point: Imagine if that scientist took the skepticism so far that he declared, on the basis of optical illusions, that this applies to all conditions - that is our sight is wholly unreliable and inaccurate. Everyone would recognize he has leveraged the data to spread far too broadly. That's going to face huge explanatory challenges with that type of hypothesis: how will he explain people passing eye exams? Reading? Driving cars? The endless examples of people clearly navigating the world successfully using sight? If he doesn't recognize the relevance of those examples, he's not a very responsible scientist. Any comprehensive view of our vision will have to acknowledge and account for the successes as well as the failures, and explore the gray areas in between, where we will have to make lower confidence judgments, yet still justified in a practical sense.

The same goes for our perception in regard to hearing. Skeptical claims about our perception in sighted conditions are not, as some have seemed to imagine here, hermetically sealed as to the implications for audio - they rely on fundamental (and true!) issues of bias in hearing perception. You can have bias effects, of course outside of just judging loudspeakers. By the same token, pushing the skepticism of sighted listening too far, and you start running in to the type of explanatory problems I've raised before. At some point you have to start explaining all the apparent successes of our hearing in non-laboratory/uncontrolled conditions. Just as people successfully navigate the world and communicate via sight, they do so with hearing perception. Whether it's every day life, or music/record production, or whatever. (As I've said before, my work in post production sound actually becomes both irrational, and it's success inscrutable, if sighted skepticism is pushed to far). We seem to get along pretty darned well appealing to hearing differences outside of the "lab," and at some point saying "It's all bias effects" is going to have pretty poor explanatory power vs the assumption people are often "really hearing what they think they hear, and really communicating truths about what they hear." There's tons of noise in our perceptual and cognitive system, but we've evolved to ultimately rise above lots of the noise to manage some significant level of reliability.

And so, if we take examples where sighted listening may be totally unreliable (I dunno, maybe when there are little to no actual sonic differences) on the other end of the spectrum will be sighted perception that is very reliable (e.g. test if I can tell the difference between my mother's voice and Trump's voice, and the reliability under both sighted and blind conditions are likely to be a match). So then we have to keep both ends in view, and realize as we trace from each side towards the centre, the least reliable and the most reliable, as they converge we'll be hitting gray or fuzzy areas in which we start to make judgments "do we NEED blind tests to have ANY justification for a conclusion in X case?" Where, given the impracticality of putting all those scenarios to the test, you adopt lower-than-scientific confidence levels, but with the proper caveats, still justified.

The question then is what, in audio, might fit in that gray area: where we aren't putting our perception to scientific rigour, it's just not practical, and yet allowing for lower confidence levels, we can justify provisional conclusions.

For myself, I would put, say, owning and comparing loudspeakers in that category. If I've owned a loudspeaker, had plenty of time playing with positions to see what it can do, I've played tons of familiar music, and then I own another speaker too and go through the same efforts, I think I'm justified enough that I"m getting some real picture of the sound, and it's not utterly obscured by bias effects. The fact that what I think I hear is also very plausible given the design of each speaker adds to the confidence. It COULD be that I'm wrong, but that caveat is built in to my conclusion. I'm being practical. I would push that to auditioning speakers in stores too. That kind of thing. Others may not.





Yeah, (assuming for sake of argument tube amps can sound different)...that's why tube amps will likely always be around. The ever more amazing specs for solid state gear don't matter, because we've effectively had sonic transparency for a long time. We've heard it. And still, some audiophiles like something extra added in, which the transparency does not provide.

As for the speakers question, it depends on the approach. Some people's approach is that they want to "not hear" the equipment. They just want it out of the way. Other people enjoy playing with sound. Being creative with it. That's their fun. Achieving a certain "sound" they have in their mind. And that's an entirely legitimate approach too (I'd say that, because it's at least partially my approach). It's what many of the guys playing with tube amps and horns and all that are doing (or at least think they are doing).
So I'm not sure a perfectly transparent speaker will obviate other options for the latter crowd.

Cheers.

Where I worked at one time, a large industrial facility, one area had a 30x40 ft metal building with hard concrete floor. 3 walls were identical metal panels with overloads and alarms in these basically 1x2 foot sections. The alarm was the 3 khz buzzer. Stupidly no lights. When a panel alarmed in that building it was nearly impossible to hear where it was. It was so reflective and all these dozens of identical panels. From one place it sounds over there, over there it sounds over yonder, over yonder it sounds like where you first were. You could wander around having a devil of a time locating it. One trick I learned was to put a finger in one ear and you could eventually have more success finding which sidewall it was on. Even then it was tough. Would I say this indicates our hearing doesn't work for direction? No, only that it doesn't work under those conditions. It does work for hearing the alarm, but not for finding it. It works just fine as soon as we walk out of that environment.

Now does our sighted hearing work listening to speakers. Mostly yes. You can hear direction in the recording and many other factors. Does it work for quality evaluation sighted with reliability and accuracy? Not very well, not as well as blind. In that situation other cues can overwhelm our acuity for basic sound quality. One trick I've played on several people. I once had a room with two pairs of speakers in it and room for another if I pulled some inside it. I'd put a speaker forward, and let people come and listen who knew me well and knew my gear. The sight of the forward speaker would make them assume that is what was being heard. However, they were hearing one of the others. I did cheat a bit more, I did a basic EQ to match the playing speaker to the one being faux auditioned. There were limits to that of course. At different times one of the speakers were Soundlabs or Quads playing while a box speaker was the faux speaker. Every time that visual clue (and I guess the sonic clue from EQ) completely fooled people. To the point these people had trouble believing the truth. Most only fully accepted it when I pointed behind where the amps were showing them the speaker cables to the faux speakers ran behind the amp stand laying on the floor not connected. So this is not analogous to someone saying our eyesight is only good under blind conditions or coming to a conclusion like eyes are so unreliable despite people being able to drive a car, read, pass eye exams (a single blind test btw) etc etc. It is that our ability to judge sound quality for music over speakers in stereo (a highly un-natural situation sometimes forgotten) is easily degraded by lots of non-sonic cues. Our ability to acclimate to sound over time mars it even more. Again, I won't claim sighted listening does not work, or even is of no value. It gives everyone a warm fuzzy feeling they know how good something sounds when the data indicates their confidence is not warranted. It means we can like and enjoy speakers that aren't very good.

Which brings me back to Kal's conclusion above after experiencing the blind listening tests.
.......good-looking Spinorama graphs or similar data will be major criteria in my considering what I'll review—and what I'll buy.

I think that is a reasonable conclusion. Spinoramas aren't fully capable of giving us perfect speakers, but in general it is a good filter to start with. That way even if you are influenced by other cues you'll be working amongst pretty good speaker designs to make your choice. It also moves us toward cost effective good designs vs some very expensive gear in a wrong-headed design. For someone who wants something different, or likes it even if less than stellar there is always that market as those designs aren't going completely away. For someone that works in sound production, I don't see how having better speakers would do anything other than enhance the accuracy of what is being done in such production choices. I don't see that these facts indicate such work is meaningless, misguided or of no value. No one said your hearing is useless altogether unless it is blind. Or at least I'm not.

 

[KenA]

Draw from this what you may…

 

[GXAlan]

https://www.audiosciencereview.com/...dn’t-we-be-looking-more-at-pa-speakers.51522/

I started a different thread but why aren’t we looking at ultra high SPL PA speakers with good directivity and multitone and then attenuating with Dirac/Acourate/PEQ to correct the frequency response? Should these be the “high end” speakers to explore/review?

 

[cavedriver]

Where I worked at one time, a large industrial facility, one area had a 30x40 ft metal building with hard concrete floor. 3 walls were identical metal panels with overloads and alarms in these basically 1x2 foot sections. The alarm was the 3 khz buzzer. Stupidly no lights. When a panel alarmed in that building it was nearly impossible to hear where it was. It was so reflective and all these dozens of identical panels. From one place it sounds over there, over there it sounds over yonder, over yonder it sounds like where you first were. You could wander around having a devil of a time locating it. One trick I learned was to put a finger in one ear and you could eventually have more success finding which sidewall it was on. Even then it was tough. Would I say this indicates our hearing doesn't work for direction? No, only that it doesn't work under those conditions. It does work for hearing the alarm, but not for finding it. It works just fine as soon as we walk out of that environment.

Ah, reinforcement in highly reflective metal environments. I use to have an Army-spec Chevy full size Blazer. It had almost zero interior, just a dash and door cards and the seats. Two 6x9 Sony 2-ways were mounted just behind the front seats firing across the interior. One of the great joys of that vehicle was listening to songs like Son of a Preacher Man with all that fast reverb on Dusty's voice - way more than the studios were already adding to some albums. Of course I knew it was not at all what the song was supposed to sound like, but it WAS entertaining. Kind of like headphones if the ear cups were 4 feet tall and 2 feet from your body.

 

[cavedriver]

https://www.audiosciencereview.com/...dn’t-we-be-looking-more-at-pa-speakers.51522/

I started a different thread but why aren’t we looking at ultra high SPL PA speakers with good directivity and multitone and then attenuating with Dirac/Acourate/PEQ to correct the frequency response? Should these be the “high end” speakers to explore/review?

not sure what aspect of the conversation you're addressing, but it gives me the idea of replacing each individual speaker with 2 or 4 speakers ala the new Genelecs and then using a custom algorithm to manage the directivity based on the room's response as measured at the listening position. Pretty obvious I suppose, but still, really not all that unattainable. Has anyone heard the Genelecs to judge how successful they have been?