Electrostatic speakers?

[MattHooper]

@Floyd Toole

Question:

Two of the subjects you’ve often talked and written about have to do with:

1. Good sound quality. This has been based mostly upon studying people’s preferences - correlating measurements to those preferences - in order to determine what most people rate as “ good sound quality” (especially with respect to loudspeaker design).

2. Accuracy. Especially its relevance to “The circle of confusion.” Ideally we’d have neutral, accurate loudspeakers used both in music production and in the consumers home, so there is continuity between what the artists hear in making the music and what we hear at home.

Those are obviously two distinctly different subjects. Accuracy per se does not entail
“ sounds good” in the first sense above: if you play a poor sounding recording through an accurate system, you’re going to hear poor sound. And sometimes “fixing” poor recordings entails making adjustments where you are playing that signal back less accurately (e.g. boosting bass to make up for a thin recording). So in that sense accuracy and sounds good remain orthogonal.

Fortunately, though, as you discovered for us, accuracy and good sound is related in the sense that accurate systems feature low colouration, and low colouration itself is associated with preferred sound quality.

So it turns out we get a sort of two-for-one result: not only are neutral/accurate/low coloration systems “better sounding” in terms of our preferences, but they also serve the second purpose of accuracy per se: accurately transmitting the sound contained in the recording and in principle maintaining accuracy between production and reproduction of the music. Which addresses the circle of confusion problem.

Would that be a fair summation?

If so, my question is:

What if your studies on people’s preferences, especially for loudspeaker designs, had yielded different results?

Let’s imagine the world in which it turned out blind studies showed a reliable preference for speakers that had a “smiley” frequency response.

Under those conditions, do you think you would’ve found yourself advocating for such “ coloured” loudspeakers to be used both in music, production, and music playback at home?

Just a kind of thought experiment I guess

 

[Floyd Toole]

So it turns out we get a sort of two-for-one result: not only are neutral/accurate/low coloration systems “better sounding” in terms of our preferences, but they also serve the second purpose of accuracy per se: accurately transmitting the sound contained in the recording and in principle maintaining accuracy between production and reproduction of the music. Which addresses the circle of confusion problem.

Would that be a fair summation?

Yes

Let’s imagine the world in which it turned out blind studies showed a reliable preference for speakers that had a “smiley” frequency response.

Under those conditions, do you think you would’ve found yourself advocating for such “ coloured” loudspeakers to be used both in music, production, and music playback at home?

Just a kind of thought experiment I guess

A perverse thought experiment I can only see that kind of result occurring if I had not done multiple-loudspeaker comparisons, which, as I have said was a most fortuitous choice for my first experiment. In "normal" subjective evaluations the program is not a controlled variable; it is combined with the loudspeaker sound. If for some incredible reason the professionals all had smiley faced monitor loudspeakers, their recordings would obviously sound better reproduced through similar smiley faced loudspeakers. But, humans being human, this was not the case. In those early days the pros were as confused about what was "good" as consumers were, and mixes were evaluated in cars, over boom boxes, various very colored little monitors, many with white woofers (an inverse smiley face), whatever was thought to be what the customer would be listening to. I would like to think that we have passed out of that phase.

 

[goryu]

If we disregard the middle ground, it seems there are two opposing camps at the extremes: those who wish to preserve the integrity of the original signal through the playback chain. Adherents to such a philosophy can be said to desire gear that is as accurate as possible. Measurement purists would thus choose their gear based on performance metrics and for those who have followed the results of Dr. Toole's research, would be assured that the odds were overwhelmingly in their favor that they would also prefer the sound subjectively and describe the sound as "good" or other such subjective descriptor. Those on the other side claim to pay little to no attention attention to measurements of objective performance and instead, "trust their ears". There seems to be little concern with the actual accuracy of the gear, rather, they "know what's good when they hear it". One group is all about the science, the other, all emotion, and yet these seemingly polar opposite approaches each have vast numbers of die hard adherents. Fascinating.

 

[Floyd Toole]

If we disregard the middle ground, it seems there are two opposing camps at the extremes: those who wish to preserve the integrity of the original signal through the playback chain. Adherents to such a philosophy can be said to desire gear that is as accurate as possible. Measurement purists would thus choose their gear based on performance metrics and for those who have followed the results of Dr. Toole's research, would be assured that the odds were overwhelmingly in their favor that they would also prefer the sound subjectively and describe the sound as "good" or other such subjective descriptor. Those on the other side claim to pay little to no attention attention to measure, objective performance and instead, "trust their ears". There seems to be little concern with the actual accuracy of the gear, rather, they "know what's good when they hear it". One group is all about the science, the other, all emotion, and yet these seemingly polar opposite approaches each have vast numbers of die hard adherents. Fascinating.

Truly fascinating, but not surprising. As I think I made clear in my book and the slide show, for a variety of reasons most purely subjective evaluations are not trustworthy. Opinions are inconsistent, and so are the chosen products. That "method" has supported many kinds of manufacturers for decades. Everything found a buyer.

Trusting what is heard at any point in time is a natural inclination - human nature. The problem is that the coloration in the loudspeaker that made it appealing at that place and time is permanent. It will be superimposed on every recording, every movie soundtrack that is played through it. We definitely can adapt to some horrendous colorations, but it is a pity to have to expend any mental capacity to do so because we know that if those people made their choices under different circumstances, they would select timbral neutrality. In the hundreds of listeners we have used there were all kinds of people with all kinds of biases who, if their hearing was normal, in the end agreed on what was good.

 

[MattHooper]

Yes

A perverse thought experiment I can only see that kind of result occurring if I had not done multiple-loudspeaker comparisons, which, as I have said was a most fortuitous choice for my first experiment. In "normal" subjective evaluations the program is not a controlled variable; it is combined with the loudspeaker sound. If for some incredible reason the professionals all had smiley faced monitor loudspeakers, their recordings would obviously sound better reproduced through similar smiley faced loudspeakers. But, humans being human, this was not the case. In those early days the pros were as confused about what was "good" as consumers were, and mixes were evaluated in cars, over boom boxes, various very colored little monitors, many with white woofers (an inverse smiley face), whatever was thought to be what the customer would be listening to. I would like to think that we have passed out of that phase.

Indeed!

It seems to me, and I say this with admiration, that much of your work could be said to be aimed at investigating (and then disputing ) the old maxim:

“There’s no accounting for taste.”

 

[STC]

Mono is used in the multiple-loudspeaker comparison tests because the most common problem in loudspeakers, resonances, is most easily heard in that format.

Dr. Toole,

You mentioned that your experiments were conducted in mono because resonances are more readily identifiable in mono than in stereo. That observation is central to the question I have been raising. In stereo, accuracy is important, but it is not the whole story — stereo is also an art form that creates a fake illusion of a real acoustic scene neither perfectly or accurately.

I haven’t yet read the 4th edition, but your earlier books were fundamental to my understanding of these issues, and I appreciate the time you have taken to explain the 2 kHz dip and the stereo soundstage.

I’m familiar with that dip from work I did developing a related plugin. A general listener might simply hear the phantom centre as slightly less bright than a hard-panned mono signal. But many people also notice that the same recordings often sound smoother or mellower on headphones or IEMs than on speakers — in that case the missing pinna cues often seem more perceptually significant than the crosstalk dip itself.

You described stereo production as involving artistic judgement, and I agree. From my own experience, even experienced listeners who strongly preferred one remaster over another (often citing bass and vocals) were unable to accurately re-EQ a modified version of their preferred track back to the original when only a few frequency regions had been altered. This left me questioning the consistency of such subjective judgements. I have not seen published blind data showing how consistent recording and mastering engineers are in these artistic decisions.

There is a certain irony here: we place great emphasis on loudspeaker accuracy in order to preserve an inherently subjective art form, while expecting that accuracy to somehow deliver the creator’s intent intact into our minds.

For what it’s worth, if I were buying a new pair of speakers today I would still choose a faithfully neutral design (most likely an ESL). Not because it will automatically be my final preference, but because it gives me the most reliable starting point to shape the sound to my liking.

Thank you again for the engagement. I remain one of your over-enthusiastic students — your work (and Ralph Glasgal) has long been a guide for me.

ST

 

[STC]

This is a legitimate function of mastering--to account for interaural perception events at the eardrums--not from baking in "exceptions" into the loudspeakers. All mastering (and mixing) guys that I've seen used loudspeakers (stereo or multichannel) to mix and master the various recorded channels. That's not something that the loudspeakers need to deal with.

I have to say that this is yet another example of the confusion that exists in sound reproduction circles because there are no standards by which the "professionals" must adhere to.

Chris

If mastering engineers are already accounting for the two-speaker summation and interaural effects while working on neutral loudspeakers, then they are necessarily applying EQ and processing so that the combined sound at their ears is what they want. In other words, they are colouring the recording itself so that it will sound correct when later played back through similar neutral speakers.

This sits in tension with the approach of certain labels (such as Linn and Water Lily Acoustics) that deliberately aim for minimal or no EQ, valuing a more pure and unprocessed signal path. If the recording must be adjusted to make neutral loudspeakers sound “right” in stereo, then the ideal of a transparent recording chain becomes harder to maintain.

That seems to me another illustration of the lack of consistent standards you mentioned.

ST

 

[STC]

Edited- this response meant to address Chris’s post. Wrongly quoted Dr Toole earlier.
———————————
“

I remember the Polk "SDA monitors" from the late 1970s-early 1980s that gave the listener a sense of 180 degrees of soundstage illusion--as long as you were sitting really close to the centerline between the loudspeakers, the space between the listeners and the loudspeakers were completely free of early reflections, and a pristine condition existed for side wall and front wall early reflections. Bob Carver did the same thing in his bigger preamp by introducing his "Sonic Holography", but not at the loudspeakers, rather via cross-channel mixing electronically upstream of the amplifier(s).

What happened to the SDA monitor and the Carver Sonic Holography products? Why did they stop making them?

Chris

Regarding the Carver Sonic Holography units: they were an early analogue attempt at crosstalk cancellation. There are still few people still use them. Modern DSP-based crosstalk cancellation is significantly more effective and flexible. There are several good plugins available now — you can find them by searching for current crosstalk cancellation tools, or check the one I developed (linked in my profile).

As for Polk SDA, they still exist in as soundbars. That is understandable as that the best way to do crosstalk cancellation. With conventional stereo speakers, the most effective crosstalk cancellation and spatial effect tends to occur when the speakers are placed at a narrower angle — around 20 degrees works best in my experience. It can still function at 60 degrees, but the effect is noticeably weaker. Closely placed speakers setups generally give the strongest result.

ST

 

[MattHooper]

Regarding the Carver Sonic Holography units: they were an early analogue attempt at crosstalk cancellation.

My dad was something of an audiophile and I grew up listening to KEF 105.2 flagship speakers powered by Carver amplification, including the famous C4000 preamp with the Sonic Holography feature. (we also had two rear speakers that could be engaged for a more immersive sound).

I remember playing quite a bit with the sonic holography settings. They certainly produce the most impressively spacious and dimensional sound. But I often preferred playback with Sonic holography turned off.
Without that trickery engaged, the sound seemed a little bit more solid and natural and neutral to my ears… I guess I might’ve said back then “less phasey.”

 

[dlaloum]

Interesting. I am the person who started the loudspeaker "Harman curve" thing but I don't understand your comment There is another Harman curve for headphones that Sean Olive originated and that one has some legitimacy).

First, all room curves, including the so-called Harman curve are results not targets. The original curve was what was measured in a room for a number of highly rated forward firing cone/dome/horn loudspeakers. That is all. It is explained in Section 12.3.5 in the 4th edition (originating in Section 12.2 in the 3rd edition). People jumped to the conclusion that equalizing any loudspeaker to match that room curve would result in "highly rated" sound - that is simply wrong! Room curves are dominated by far off-axis (reflected) sounds and the on-axis, direct, sound is a more important factor in the sound quality we perceive.

There is nothing uniquely good or bad about ESLs - like all designs they come in many "flavors". The dipole radiation characteristic is a distinguishing characteristic and it delivers a very different sound field from the common forward-firing designs. As you correctly point out, bass output is often a limitation, but good ones have enough bass to be very gratifying.

The amount of bass one "prefers" has to depend on what recording one is listening to. I find myself occasionally adjusting the bass level to compensate for recordings with too much or too little. Many listeners may have experienced bass deprivation in their youth and enjoy a little boost if it is available

Bass is disproportionately important to one's overalll impression of sound quality (about 30%), and the listening room dominates the quality of bass we hear so generalizations are risky.

Excuse the confusion, and that includes my own confusion - I am well aware of the curve being a result rather than a target, however, even though I have frequently reminded others of this, I fell into that trap myself!

In the room EQ result, I find myself preferring a response curve that has less bass prominence....

My own preference for panel speakers in the past, may have had an impact on my preferences and/or the differing way that the types of speakers I prefer (multi polar rather than uni polar) results in them activating the room differently, resulting in a typical room response with less of a low frequency hump.

Using Room EQ software (Dirac ART in my case) - also results in unconsciously working towards a target - as that is what one sets in the software - rather than focusing on it as a result....

Whichever, in my own setups, the resulting curve that I find most "real" (neutral?) when listening to acoustic instruments (my baseline) - tend to approach more of a flat F/R in the low to midrange frequencies. - This is atypical, and most people seem to report preferring substantially more bass....

(I have also been conflating the headphone curve with the speaker response curve from your work....which I probably should not do!)

 

[STC]

My dad was something of an audiophile and I grew up listening to KEF 105.2 flagship speakers powered by Carver amplification, including the famous C4000 preamp with the Sonic Holography feature. (we also had two rear speakers that could be engaged for a more immersive sound).

I remember playing quite a bit with the sonic holography settings. They certainly produce the most impressively spacious and dimensional sound. But I often preferred playback with Sonic holography turned off.
Without that trickery engaged, the sound seemed a little bit more solid and natural and neutral to my ears… I guess I might’ve said back then “less phasey.”

Carver was one of the earliest to bring practical crosstalk cancellation to consumers with Sonic Holography. By today’s standards it’s a relatively simple analog approach—something we can now do (and refine) far more precisely in a DAW or with modern DSP.

I recognize the “phasey” quality you describe; it’s a common side-effect when the cancellation isn’t perfect, and it can still appear even with digital XTC if the geometry or filters aren’t dialed in carefully. Carver’s manuals walked through the setup in detail (equal path lengths to the listening position, careful toe-in, and often closer-than-usual speaker spacing). The circuit used fixed delays and filtering that approximate typical interaural differences rather than a measured personal HRTF, so results varied with head position and individual hearing. Unless, you calibrated to your HRTF it is going be phasey.

It was a fixed analog “one-shot” cancellation using inverted cross-fed signals and didn’t fully exploit the psychoacoustic advantages of optimized (closer) speaker placement the way later research and digital systems can. If you try a well-implemented modern XTC system today you might find the spatial effect more convincing and less colored—but just like 3D television, a lot of people are perfectly happy with conventional stereo. Preference is personal.

I am impressed that you also had the rear speakers. I am sure your dad had many SQ records.

ST

 

[Newman]

Excuse the confusion, and that includes my own confusion - I am well aware of the curve being a result rather than a target, however, even though I have frequently reminded others of this, I fell into that trap myself!

In the room EQ result, I find myself preferring a response curve that has less bass prominence....

Hmm. Less than what? The evidence-based advice on tuning one's own bass output is, "smoothened output at the listening position(s), extending into the deep bass, and level adjusted to taste".

There is no guideline that suggests one "will like bass level X best".

Your level is right for you, and the advice suggests no other.

If your only point is that you are one of the lower-than-typical people in your bass level preference.....so? It means only that what may come out of a default starting point setting needs to be.... adjusted to taste. As per the advice.

But if your point is more that the advice is exceedingly pedantic, then I suspect that you misread it.

cheers

 

[thewas]

Also preferred bass level depends on many variables outside of personal preferences, as listening SPL, room decay, listening material and if the bass target is followed as a summation or single response.

 

[Newman]

If we disregard the middle ground, it seems there are two opposing camps at the extremes: those who wish to preserve the integrity of the original signal through the playback chain. Adherents to such a philosophy can be said to desire gear that is as accurate as possible. Measurement purists would thus choose their gear based on performance metrics and for those who have followed the results of Dr. Toole's research, would be assured that the odds were overwhelmingly in their favor that they would also prefer the sound subjectively and describe the sound as "good" or other such subjective descriptor.

I am sure to be repeating myself here, but it is important to recognise that the dichotomy you want to paint across the issue is mistaken: this is not about "measurement purists on one side". It never has been, and to taint it as such is suspect to your understanding.

Those on the other side claim to pay little to no attention attention to measurements of objective performance and instead, "trust their ears". There seems to be little concern with the actual accuracy of the gear, rather, they "know what's good when they hear it". One group is all about the science, the other, all emotion, and yet these seemingly polar opposite approaches each have vast numbers of die hard adherents. Fascinating.

Rubbish! It has nothing to do with "science vs emotion", and never has so been. It has to do with understanding vs misunderstanding.

At the core of the misunderstanding is the misguided (but understandably common-sense) notion that whatever one is perceiving as sonic under any circumstances is entirely in response to the sound waves reaching one's ears. This notion has been tested and utterly fails the test of truth.

Instead, it turns out that the listening conditions have to be controlled (in an experimental sense) before one's sonic experience is entirely, or even reliably, in response to the sound waves themselves. Yes, then one can trust one's ears, but only in circumstances that one almost never establishes when making audio gear comparisons.... in circumstances where 'science-y' words like double-blind, randomised, and placebo-controlled apply.

If such controlled listening tests were routine in the world of audiophile gear selection processes, practically everyone who you are calling "measurement purists" and "science-y" would drop all that and choose their gear with their ears, knowing that they are indeed assessing the sonics of the actual sound waves per their hearing mechanism.

But because that isn't reality, because such tests are not easy for any of the parties involved in the process, they just don't happen.

Audiophiles can be sorted into two groups on this point: those who don't understand it, and those who understand.

Those who don't understand it, are utterly convinced that their uncontrolled gear listening experiences are all about the sound waves, are deeply insightful as to the sonic attributes of the sound waves and hence the gear, and for that reason, they insist that they can "trust their ears". They are wrong, but they are convinced they are right. Sometimes they have never had their misunderstanding explained to them, so it is understandable that they are so convinced of their method. After all, like I said in paragraph two, it is common-sense, intuitive, to think that we would only perceive something as sonic if it is actually in the sound waves. But there is also a large cohort in this group who have had it explained to them, and who are plenty smart enough to understand the explanation, but who wilfully enter "denial phase" on the issue, and just....fight it. Endlessly. With every excuse in the book. (Including here on ASR. Many get shunted to the "measurements are everything or nothing" thread where they may get short shrift and an assisted exit. But quite a lot have smooth-talked their way past the first checkpoint and become part of the furniture here, where they form something of a smokescreen between the audio science available on ASR and the access to it by readers and visitors, while perhaps seeing themselves as important sceptics with a role of endless questioning, including sealioning.)

But for those of us who do understand it, an obvious next question arises: if I can't determine the audio gear that I will most enjoy the sound saves from, by simple listening tests, how indeed am I to approach the task? And audio science has rushed in to help us here, looking into the problem and finding a reasonably effective, doable solution. (Hooray for audio science!) After a lot of experimentation, and the discovery that we are a lot less unique than may first appear in terms of our preferences for sound reproduction, audio science has come up with a suite of measurements and thresholds and guidelines for measurement results, that are a more reliable guide to our personal sonic preferences for the actual sound waves during sound reproduction than any other method short of controlled personal listening tests.

It has nothing to do with being "measurement purists" or being "all science, no emotion". Zero. It is all about understanding the most effective method currently available and doable, to get the one thing that every audiophile wants: sound reproduction with the sonic attributes that they most prefer, and can use in service of their personal music (or other audio playback) appreciation goals.

cheers

 

[goryu]

Semantics. You can try to explain the dichotomy as understanding versus not understanding but the fact remains- There are those who trust science and those who refuse to believe anything but their own ears. You can try to explain the cause (understand vs not understanding) or simply note the effect (believe or disbelieve). Call it what you like. I personally tend to suspect that there are multiple reasons people choose to ignore science and it isn't simply a question of not understanding. There are plenty of people who willfully believe in things that science doesn't support. It's called "faith" and the world's religions are based upon it.

 

[RetroStereo]

Semantics. You can try to explain the dichotomy as understanding versus not understanding but the fact remains- There are those who trust science and those who refuse to believe anything but their own ears. You can try to explain the cause (understand vs not understanding) or simply note the effect (believe or disbelieve). Call it what you like. I personally tend to suspect that there are multiple reasons people choose to ignore science and it isn't simply a question of not understanding. There are plenty of people who willfully believe in things that science doesn't support. It's called "faith" and the world's religions are based upon it.

Then there are those whose ears are as golden as the price tags of the cables on permanent loan.... This might apply to a few of those faith leaders, as well.

 

[traind]

On a somewhat related note: In my business (sound design for Film and TV) we generally depend on subjective descriptions to communicate among ourselves and with clients.

Wait a minute, I thought you were a marine biologist...

 

[MattHooper]

Semantics. You can try to explain the dichotomy as understanding versus not understanding but the fact remains- There are those who trust science and those who refuse to believe anything but their own ears. You can try to explain the cause (understand vs not understanding) or simply note the effect (believe or disbelieve). Call it what you like. I personally tend to suspect that there are multiple reasons people choose to ignore science and it isn't simply a question of not understanding. There are plenty of people who willfully believe in things that science doesn't support. It's called "faith" and the world's religions are based upon it.

Yup, it was obvious what you meant. You’d given the context for what you were getting at: The difference between those taking a more scientific approach (see for instance, the name of this forum), letting science derived data drive their understanding of how audio gear performs, versus those taking a more “ I know it because I feel it when I’m listening” approach. Those following the science instead of letting subjective emotion lead their beliefs about audio. (Which obviously does not rule out the science followers having as much emotion in listening to music playback as anyone else).

Paying attention to context to try and understand what people mean, rather than picking out trigger words to react to, can save misunderstandings and unnecessary reactions.

Often the imagined “smoke” is “coming
from inside the house.”

 

[dlaloum]

Hmm. Less than what? The evidence-based advice on tuning one's own bass output is, "smoothened output at the listening position(s), extending into the deep bass, and level adjusted to taste".

There is no guideline that suggests one "will like bass level X best".

Your level is right for you, and the advice suggests no other.

If your only point is that you are one of the lower-than-typical people in your bass level preference.....so? It means only that what may come out of a default starting point setting needs to be.... adjusted to taste. As per the advice.

But if your point is more that the advice is exceedingly pedantic, then I suspect that you misread it.

cheers

So I was referring to this as the "Harman Curve"

And primarily the "all listeners" profile.... which demonstrates that most people prefer what I would refer to as a bass boosted F/R profile.

My personal preference approximates to the "trained listeners only" F/R

From memory (without searching for the relevant references) - the preferred profile for 60%+ of the population is similar (if not the same as) to the "all listeners" or "untrained listeners" profiles on this chart. And as a rule of thumb / colloquial manner of speech I would call those F/R profiles "Bass boosted" as contrasted to the predicted steady state room f/r for flat direct sound and the "trained listener" F/R.

And yes it is SUBJECTIVE (the name of the chart is "Subjectively preferred...." - and part of my point was that it is my own subjective preference (ie: to taste) - and I was originally responding to an earlier post, and making the remark that for many of us outside the majority preferred 64%... a profile with less bass (or more!) would be subjectively preferred

 

[Floyd Toole]

So I was referring to this as the "Harman Curve"

View attachment 550881

And primarily the "all listeners" profile.... which demonstrates that most people prefer what I would refer to as a bass boosted F/R profile.

My personal preference approximates to the "trained listeners only" F/R

From memory (without searching for the relevant references) - the preferred profile for 60%+ of the population is similar (if not the same as) to the "all listeners" or "untrained listeners" profiles on this chart. And as a rule of thumb / colloquial manner of speech I would call those F/R profiles "Bass boosted" as contrasted to the predicted steady state room f/r for flat direct sound and the "trained listener" F/R.

And yes it is SUBJECTIVE (the name of the chart is "Subjectively preferred...." - and part of my point was that it is my own subjective preference (ie: to taste) - and I was originally responding to an earlier post, and making the remark that for many of us outside the majority preferred 64%... a profile with less bass (or more!) would be subjectively preferred

As I explained in post 1391, this is not the origin of the "Harman curve" that matters. It is the result of a very different experiment, as I explain in some detail in Section 12.3.6 in the 4th edition. This was a simple experiment - using too few listeners to draw important conclusions - at a time when Sean Olive and Todd Welti were investigating listener preferences in headphones.

Headphones are fundamentally different from loudspeakers in that a single curve defines performance, there is no room and 3D sound field between the transducer and the listener. It was their thought that there might be a useful relationship between a loudspeaker room curve and a headphone curve - it was the beginning of an investigative process.

The method was to equalize a "highly rated" loudspeaker to yield a flat room curve. This was the starting point, and listeners were given bass and treble controls to adjust the spectrum until they liked what they heard. The sound of the recording was combined with the sound of the loudspeaker, so one would expect a different result for different recordings. This is a fundamental issue with normal subjective loudspeaker and headphone reviews.

A forward-firing loudspeaker delivering a flat room curve would normally be perceived as being excessively bright, and probably most of us would turn the treble down. In this case the experienced listeners did it with remarkable precision, adjusting the treble control so that the direct sound was flat and the room curve tilted down. They succumbed to the attractions of slightly elevated bass, which is a very common finding, especially with movies where dramatic booms and thumps cannot be too loud it seems. Boosted bass is a forgivable sin, if not a desirable feature, depending . . .

The inexperienced listeners simply wanted more of everything, including loudness (which was not controlled). I often find this in the tone control settings of rental cars - more must be better in the naive mind. Another experiment with controlled overall loudness would be needed to get closer to a useful finding.

So, although conducted for a different reason, the experiment added evidence to the primacy of the direct sound in determining overall timbral accuracy.