I cannot trust the Harman speaker preference score

[thewas]

If you mean by yet as the time the Harman/Olive test were done and my speakers were designed, you are wrong. I designed Silver 5L during 1991. Harman/Olive patent was granted in 2005. I expect their tests were done in the noughties, i.e a decade later.

That is true for the final research of Olive that lead to the here discussed equation and score, though systematic testing to understand the correlation of measurements and listeners preference was started much earlier by Toole, for example in 1986 Toole published the results of a two-year study where forty-two listeners evaluated thirty-seven different loudspeakers.

Also in 1990, Klippel reported a perceptual-based loudspeaker model for predicting various sound quality dimensions and overall sound quality. The model was based on a massive study involving seven different experiments designed to examine the influence of factors on loudspeaker quality such as listener experience, room acoustics, speaker directivity, program material and method of scaling (semantic differential versus MDS). A total of forty-five different loudspeakers (both real and simulated), three different rooms, thirteen programs and forty different listeners were compared. The rooms included an anechoic chamber, an IEC listening room and a small studio. Factorial analysis revealed seven unique dimensions such as clearness, treble stressing (sharpness), general and low bass emphasis, feeling of space, clearness in bass and brightness.

Source and more: https://patentimages.storage.googleapis.com/95/03/0a/a4a8dbd7d8042c/US20050195982A1.pdf

 

[Axo1989]

Preference means...wider directivity listening in mono?

I shouldn't really like this post, but it was pretty funny.

 

[Axo1989]

That is true for the final research of Olive that lead to the here discussed equation and score, though systematic testing to understand the correlation of measurements and listeners preference was started much earlier by Toole, for example in 1986 Toole published the results of a two-year study where forty-two listeners evaluated thirty-seven different loudspeakers.

Also in 1990, Klippel reported a perceptual-based loudspeaker model for predicting various sound quality dimensions and overall sound quality. The model was based on a massive study involving seven different experiments designed to examine the influence of factors on loudspeaker quality such as listener experience, room acoustics, speaker directivity, program material and method of scaling (semantic differential versus MDS). A total of forty-five different loudspeakers (both real and simulated), three different rooms, thirteen programs and forty different listeners were compared. The rooms included an anechoic chamber, an IEC listening room and a small studio. Factorial analysis revealed seven unique dimensions such as clearness, treble stressing (sharpness), general and low bass emphasis, feeling of space, clearness in bass and brightness.

Source and more: https://patentimages.storage.googleapis.com/95/03/0a/a4a8dbd7d8042c/US20050195982A1.pdf

You can see this as prior art for the Harman patent (as cited, and in the content).

The missing elements (which play out in several of @sarumbear's examples) include baseline distortion, dynamics and associated compression and increase in distortion, time-series spectrographic analysis to identify influences on timbre, and so on.

For people who argue that the score is a good start but needs some work, then there are those (and likely other) things to look at. They can continue working on the scoring method.

For people (who don't need to design speakers, compare large numbers of speakers, test speakers etc and) who are are happier looking at the data rather than a combined index, then there are those things to look at. They can continue to ignore the scores.

 

[Holmz]

That's more or less my take on it. I think I'd trust it enough to at least help me narrow down potential choices. Then of course the final test is to listen to them if at all possible. If you can't, try to buy from somewhere that offers a trial period. A lot more companies seem to be adopting that approach nowadays.

Maybe I lack confidence.

I listened to the speakers and thought, “Damn, these things sound great.”
But thought I was likely having a sighted bias, so I awaited the measured review.
But that review came back as high Olive Preference Score.


Just last week I listened to a $10k and then $40k set of speakers.
I mentioned that the soundstage seemed pulled to the right.
The two sales people swore it was just dandy, and another person there said, maybe just a bit.

I keep putting pinky in my left ear, digging in hope… and sometimes it sounded centered and sometimes wandering to the right.
I thought, I either have an ear wax issue, I have had a stroke, or the hearing declined in a non-gradual process since last week.

The next day I was informed that there was an issue with the preamp. (All high $ ARC gear… like 100k$ system I suspect.)


So I am somewhat adverse to making decisions solely on listening.
The room, source material, etc. all makes it too easy for the seller to control the situation in their favour.
And it may not say anything about how they will sound at home.
The best it might show is how two different sets might compare in their sales space.

And I never 100% “trust my ears.”

 

[Pogre]

Maybe I lack confidence.

I listened to the speakers and thought, “Damn, these things sound great.”
But thought I was likely having a sighted bias, so I awaited the measured review.
But that review came back as high Olive Preference Score.


Just last week I listened to a $10k and then $40k set of speakers.
I mentioned that the soundstage seemed pulled to the right.
The two sales people swore it was just dandy, and another person there said, maybe just a bit.

I keep putting pinky in my left ear, digging in hope… and sometimes it sounded centered and sometimes wandering to the right.
I thought, I either have an ear wax issue, I have had a stroke, or the hearing declined in a non-gradual process since last week.

The next day I was informed that there was an issue with the preamp. (All high $ ARC gear… like 100k$ system I suspect.)


So I am somewhat adverse to making decisions solely on listening.
The room, source material, etc. all makes it too easy for the seller to control the situation in their favour.
And it may not say anything about how they will sound at home.
The best it might show is how two different sets might compare in their sales space.

And I never 100% “trust my ears.”

Well, head to head listening comparisons at hifi shops can be dubious depending on the room, setup and sighted bias for sure, yes.

I don't have a singular method for choosing speakers. It's a matter of collecting data and listening. The Harman score is just another data point worthy of taking into consideration, imo. It's not the final arbiter, nor is a single listening test under dubious conditions. The home test is the ultimate test at the end of the day.

 

[lherrm]

The home test is the ultimate test at the end of the day.

Sighted or unsighted ? ;-)
(derivated question : if one prefers a speaker sighted and the other unsighted, should he buy what he prefers with ears and eyes or ears only ? My naïve answer would be : with the same configuration as he will regularly enjoy music with - probably with both)

 

[Pogre]

Sighted or unsighted ? ;-)
(derivated question : if one prefers a speaker sighted and the other unsighted, should he buy what he prefers with ears and eyes or ears only ?)

Unsighted for ultimate accuracy of course, but I've personally never gone through the lengths to set one up like a dbt.

I trust the science behind the method and believe there's definitely a corellation. I'm not advocating otherwise. I'm just saying that I don't rely on just one metric or one uncontrolled listening session. When I shop speakers I take my time and collect data first, then I seek out potential candidates and start listening.

 

[Pogre]

I could see myself in a situation where I get it narrowed down to 2 or 3 candidates and can't make up my mind listening to them. In a case like that, I can't decide just listening, then I could see using something like the Harman score to help me make up my mind.

 

[steve59]

I've always complained that single speaker testing in an unfamiliar room would be useless for selecting a stereo pair for a different room, but then I saw a pic of the room and the speakers aren't in the exact same place in the room, just the same distance out in the room with each platform moving front to back, wouldn't this offer another variable since listeners would be a different distant off axis?

 

[dominikz]

I've always complained that single speaker testing in an unfamiliar room would be useless for selecting a stereo pair for a different room, but then I saw a pic of the room and the speakers aren't in the exact same place in the room, just the same distance out in the room with each platform moving front to back, wouldn't this offer another variable since listeners would be a different distant off axis?

If you think about the Harman loudspeaker blind test facility, their loudspeaker shuffler apparently moves all tested loudspeakers to the same, center position:

In order to yield the most accurate A/B test results, JBL manufactured the only known Speaker Shuffler, a hydraulic-driven, computer-operated device capable of swapping the speakers into a precise center position within four seconds

(source)

Also nicely illustrated in this video (source). As you can see, while individual loudspeakers indeed only move front-back on the small platforms, the larger platform beneath all of them moves left-right to center the speaker currently playing. Hope this helps!

 

[IPunchCholla]

Thesaurus.com - The world's favorite online thesaurus!

Thesaurus.com is the world’s largest and most trusted online thesaurus for 25+ years. Join millions of people and grow your mastery of the English language.

www.lexico.com

You do see that meaningless and useless are synonyms in the link you posted, right?

 

[sarumbear]

You do see that meaningless and useless are synonyms in the link you posted, right?

I do but I also know that words should be used in context. My use case was as in "a meaningless statement" not as in "she felt her life was meaningless". Without nuance communications fail, as it has been demonstrated...

 

[IPunchCholla]

I do but I also know that words should be used in context. My use case was as in "a meaningless statement" not as in "she felt her life was meaningless". Without nuance communications fail, as it has been demonstrated...

And you do understand that those two cases are not mutually exclusive, nor are thesauruses proscriptive? “The useless statement” and “the statement is useless” are not understood to be meaningfully different statements by a reasonable person, therefore the relationships in the second example apply to the first, to the degree that any synonym would. Substituting meaningless doesn’t seem to change anything.

Your focus on this almost seems to be a red herring.

Either way the statement that the Garmin preference score is useless or meaningless or valueless isn’t an opinion. It is a statement of fact. As such, as the claimant, it is your obligation to prove the claim. Which is where much of the objection to your post seems to be coming from, including mine. I find meaning, use, and value in the score. Given other criteria equalization (SPL, influence of sub, cost, use case…) it (to .8 of a point difference) indicates an increased probability that one speaker would be preferred more than another. Having no reason to think I differ in my preferences significantly, I can use that to help focus speaker investigation.

If your claim is that you don’t LIKE it (because it ranks speakers you admire equivalent to those you don’t), fine. But that isn’t useful to me. So I’m out.

 

[sarumbear]

Either way the statement that the Garmin preference score is useless or meaningless or valueless isn’t an opinion. It is a statement of fact. As such, as the claimant, it is your obligation to prove the claim

We have left nuances of language and entered into miss-understanding what is written. I start with saying “I cannot trust the Harman speaker preference score”. It’s what I think. Check the meaning of IMHO.

Why some people can’t accept that others may have different opinions?

 

[Andrej]

Several things have always bothered me with subjective evaluations of speakers (assuming double blind, etc. to make it repeatable).
-- Any such comparison is very much influenced by the volume at which it is conducted given our variable hearing acuity at different frequencies
-- They have only a marginal relationship with accuracy/fidelity (with electronics we compare what goes out with what went in, with speakers, realistically, all we have is what came out)
-- Listening environment influences, speaker bandwidth, ability to play loud enough (one can limit the range of speakers to those with same intended use-cases), etc.

My preferred method would be to (obviously) adjust the volume level for each speaker, after it was first adjusted (by Dirac, for example) to a desirable full bandwidth target curve for the chosen listening volume.

Anything else is too noisy, in my opinion, to be particularly useful.

EDIT: This is probably not directly relevant for this topic!

 

[Frgirard]

We have left nuances of language and entered into miss-understanding what is written. I start with saying “I cannot trust the Harman speaker preference score”. It’s what I think. Check the meaning of IMHO.

Why some people can’t accept that others may have different opinions?

You touched an idol. There are numbers and a computer, on a forum that takes to doing science, it's blasphemy.

 

[tuga]

We have left nuances of language and entered into miss-understanding what is written. I start with saying “I cannot trust the Harman speaker preference score”. It’s what I think. Check the meaning of IMHO.

Why some people can’t accept that others may have different opinions?

I looked up indoctrination - the process of teaching a person or group to accept a set of beliefs uncritically

 

[Axo1989]

And you do understand that those two cases are not mutually exclusive, nor are thesauruses proscriptive? “The useless statement” and “the statement is useless” are not understood to be meaningfully different statements by a reasonable person, therefore the relationships in the second example apply to the first, to the degree that any synonym would. Substituting meaningless doesn’t seem to change anything.

Your focus on this almost seems to be a red herring.

Either way the statement that the Garmin preference score is useless or meaningless or valueless isn’t an opinion. It is a statement of fact. As such, as the claimant, it is your obligation to prove the claim. Which is where much of the objection to your post seems to be coming from, including mine. I find meaning, use, and value in the score. Given other criteria equalization (SPL, influence of sub, cost, use case…) it (to .8 of a point difference) indicates an increased probability that one speaker would be preferred more than another. Having no reason to think I differ in my preferences significantly, I can use that to help focus speaker investigation.

If your claim is that you don’t LIKE it (because it ranks speakers you admire equivalent to those you don’t), fine. But that isn’t useful to me. So I’m out.

Usefulness, meaning and value apply at least in part to the person (they are not intrinsic or fixed) so discussion of same is inherently subjective.

For example, I don't use the Harman score as it doesn't convey meaning, therefore it has no value. To me. Someone else may find the score useful because it does convey meaning, therefore they do value it. These are statements of fact (about usefulness, meaning and value to me versus someone else) and both these statements are true (even though judgements about usefulness, meaning and value differ in each case).

 

[Frgirard]

Usefulness, meaning and value apply at least in part to the person (they are not intrinsic or fixed) so discussion of same is inherently subjective.

For example, I don't use the Harman score as it doesn't convey meaning, therefore it has no value. To me. Someone else may find the score useful because it does convey meaning, therefore they do value it. These are statements of fact (about usefulness, meaning and value to me versus someone else) and both these statements are true (even though judgements about usefulness, meaning and value differ in each case).

It's a good definition of the nihilism. Are we on a woke forum?

 

[Axo1989]

It's a good definition of the nihilism. Are we on a woke forum?

Is that a language barrier or just you being a dick? Different people assigning different meanings and values to things has nothing to do with nihilism at all. Nihilism denies values and meaning altogether (more or less).