Master Thread: Are measurements Everything or Nothing?

[voodooless]

have the same similar preferences in speakers

TIFIFY

 

[robh36062]

"How can you classify the sound of a system, device, loudspeaker or headphones if you have never heard what it must sound like in reality?"

Doesn't matter we are not listening to tlme machines they are reproducers of sound not the original source. Are they exact? No! Do you like what you hear? Yes done deal.

Rob

 

[Newman]

TIFIFY

MYDNFIFH

 

[ahofer]

I do however still not think it's just a matter of picking the speaker with the highest preference score and then calling it a day

I suspect that would be a good strategy for those who aren’t as far into the hobby as we are, though.

 

[ahofer]

That is a romantic and seemingly logical statement, but if it were true, the work of Toole and Olive would have come to nothing, since we would all be so different.

Instead, they found that we pretty much all give the highest weighting to the same thing: absence of colouration. That includes flatness of frequency response, because anything not flat will not reproduce the rich harmonic structure of vocals and instruments intact, and that will sound coloured. Adding resonances will also sound coloured. Toole has written that the human ear is a remarkably sensitive and reliable instrument in this regard.

The real reason that we all appear to have different preferences on speakers is that we each have a unique set of non-sonic biases that we will mistakenly attribute to the sound waves when we engage in sighted listening, as is usual. Our actual preferences for sound waves are remarkably uniform.

Cheers

I think Toole and Olive’s work is indicative of a majority preference, but I think you have perhaps overstated it. We really don’t have large N trials to go on.

 

[Rõlnnbacke]

The testing of loudspeaker preferences say otherwise. Regular people, studio people, musicians, audio reviewers, audio salesmen, and engineers even people from cultures with different native languages have the same preferences in speakers.

Seems logical too, if you consider our hearing being tuned to reality.. but can be influenced by what we see. But registered sounds coming out of boxes is not 'realistic', so maybe our brains get confused in a way, more so if we see them.

 

[Romario]

I struggle to compare speaker measurements. For arguments sake, a nice FR in a below 1K bucks a pair compared with a similar FR for 10K. The latter will sound much better but if I gave one just the measurements without the model and price, can you tell where I could pin point that sound improvement? Or do we need to disassemble and do all kinds of driver measurements like in the DIY community to be exactly sure?

 

[Axo1989]

Nobody likes an empty ad hominem either. Do you really need to pop up practically every time Newman posts, with comments like this?

Not ad hominem. My assertion is that our friend's argument is shite because it draws over-simplified and dogmatic conclusions from somewhat limited data. Not because of the personal characteristics they display in the other thread.

 

[kimmosto]

The testing of loudspeaker preferences say otherwise. Regular people, studio people, musicians, audio reviewers, audio salesmen, and engineers even people from cultures with different native languages have the same preferences in speakers.

No. They may have almost the same preference in sound, but not in speakers. The reason is that final sound reproduction is combination of devices including speakers, environment (acoustics) and locations of speakers and listener. Studies are always limited and inaccurate no matter how many speakers or people involved if the room or listening position or speaker positions or any of those does not change. Studies using the same speaker locations are quite deaf for different speaker concepts in practice. Changing system is possible to arrange for conventional smallish speakers, but in-wall, on-wall, corner, dipoles etc. may not have location or background or both what they require generally or individually. For example, blind tests by local hifi magazine fail for sure if all tested speakers are not designed for the same distance to front wall though they could have exactly the same concept such as small boxed bass reflex 2-way.
All this is part of many reasons why for example CTA-2034 Directivity Response cannot be interpreted in one way. Manufactures may feel pressure to design or weight for a particular interpretation to avoid unreasonable barking on ASR though their concept could require knowledge and experience that you can't find much on the site.
Preference studies (such as Olive's preference rating) could also ignore many obvious errors based on other studies or relative significance, but ignored errors are not insignificant. Just less significant for many, but intolerable for some.

 

[robh36062]

I struggle to compare speaker measurements. For arguments sake, a nice FR in a below 1K bucks a pair compared with a similar FR for 10K. The latter will sound much better but if I gave one just the measurements without the model and price, can you tell where I could pin point that sound improvement? Or do we need to disassemble and do all kinds of driver measurements like in the DIY community to be exactly sure?

It's not just FR. You have sensitivity, power handling, and directivity not getting into distortion levels, what the decay characteristics are like a CSD measurement or step response. All of this can give you valuable information independent of just FR alone.

It's a possibility the system with the rougher FR may sound better because differences noted above. Selecting speakers is personal, sure you can lump a large portion of people into the "Toole Group" but there will also be outliers who don't go with the modern design philosophies.

Just find something you like.

Rob

 

[Romario]

It's not just FR. You have sensitivity, power handling, and directivity not getting into distortion levels, what the decay characteristics are like a CSD measurement or step response. All of this can give you valuable information independent of just FR alone.

It's a possibility the system with the rougher FR may sound better because differences noted above. Selecting speakers is personal, sure you can lump a large portion of people into the "Toole Group" but there will also be outliers who don't go with the modern design philosophies.

Just find something you like.

Rob

Yes exactly. But can one look at all those measurements and pinpoint the better speaker is what I'm trying to grab my head around. I'm not in for a new pair, just discussing to learn more.

 

[sigbergaudio]

Yes exactly. But can one look at all those measurements and pinpoint the better speaker is what I'm trying to grab my head around. I'm not in for a new pair, just discussing to learn more.

Among similar measuring speakers, it can be difficult. If a speaker has clear problems, then that can be seen in the measurements. And one can assume it will likely not sound as accurate as a different speaker that measures much better.

But we can also see pretty cheap and small speakers measure well in many respects, but still not sound the same as a larger, more expensive speaker - so it's pretty complicated.

 

[Galliardist]

Not ad hominem. Our friend's argument is shite because it draws over-simplified and dogmatic conclusions from somewhat limited data. Not because of the personal characteristics they display in the other thread.

Well, that's what you should have said, then. Can you show in support of that, the data sets concerned are "somewhat limited"? Can you point to a larger dataset that shows something different? What are your conclusions based on the available data?

Lay out your own argument in detail. I'm here to learn.

 

[MattHooper]

Among similar measuring speakers, it can be difficult. If a speaker has clear problems, then that can be seen in the measurements. And one can assume it will likely not sound as accurate as a different speaker that measures much better.

But we can also see pretty cheap and small speakers measure well in many respects, but still not sound the same as a larger, more expensive speaker - so it's pretty complicated.

As I’ve said before, there is a way in which speaker measurements can be made easier to comprehend, which is by an adopting a target curve as your general criteria.
So if you just take the general type of response cited on this forum as indicative of a well designed loudspeaker with good sound, you’d be looking at essentially what some call the Harman Curve. It doesn’t take long to recognize the pattern there of the essentially flat on axis response with a gently sloping evenly spaced set of off access responses.

So you can say “ look for a measurement like this this sounds good.” And then one can also have their own experience of having greatly win down their selections, listening to loud speakers with essentially that curve, and understanding with that sounds like.

But where things get far more difficult is when you ask “ OK but what do all the other loudspeaker designs sound like?”

In that case, there are so many different speaker designs that measure so differently, you are being asked to look at a large variety of all sorts of different measurement aspects and somehow knowing just from the measurements exactly how any loudspeaker at all would sound. That takes a different level of understanding and direct experience correlating measurements to their Sonic consequences.

And then, what do you do if your interests extend beyond the relatively few available loudspeakers that have not spinorama-level measurements? Then it seems to be a “ try to hear them yourself” situation.

So on that level of knowing what all loudspeaker sounds like just from the measurements, not to put you on the spot, but I’m curious about what you would infer from this set of measurements.

If you were put into words what these measurements tell you in terms of the subjective effect, what would that speaker sound like?

I’d open up that question to anyone else here who would take a stab at it.

(I ask because I have quite a bit of experience, listening this loudspeaker):

(From stereophile)

 

[Blumlein 88]

No. They may have almost the same preference in sound, but not in speakers. The reason is that final sound reproduction is combination of devices including speakers, environment (acoustics) and locations of speakers and listener. Studies are always limited and inaccurate no matter how many speakers or people involved if the room or listening position or speaker positions or any of those does not change. Studies using the same speaker locations are quite deaf for different speaker concepts in practice. Changing system is possible to arrange for conventional smallish speakers, but in-wall, on-wall, corner, dipoles etc. may not have location or background or both what they require generally or individually. For example, blind tests by local hifi magazine fail for sure if all tested speakers are not designed for the same distance to front wall though they could have exactly the same concept such as small boxed bass reflex 2-way.
All this is part of many reasons why for example CTA-2034 Directivity Response cannot be interpreted in one way. Manufactures may feel pressure to design or weight for a particular interpretation to avoid unreasonable barking on ASR though their concept could require knowledge and experience that you can't find much on the site.
Preference studies (such as Olive's preference rating) could also ignore many obvious errors based on other studies or relative significance, but ignored errors are not insignificant. Just less significant for many, but intolerable for some.

The correlation factor for speakers without response below 100 hz in Harman testing was .96 irc. It was still very high for other speakers. One may need to adjust position or eq on the low end. The preference of most people seems consistent across groups. Otherwise you just hoping for some lucky coincidence of room and odd response.

 

[pablolie]

The testing of loudspeaker preferences say otherwise. Regular people, studio people, musicians, audio reviewers, audio salesmen, and engineers even people from cultures with different native languages have the same preferences in speakers.

Everybody converges to the same preferences in speakers? That would be unknown news to us that follow the audio scene?

 

[Blumlein 88]

Everybody converges to the same preferences in speakers? That would be unknown news to us that follow the audio scene?

That was the result of Harman testing. Trained listeners were more discerning and consistent. All groups converge on the preferences. If not exactly completely identical at least very similar.

 

[NTK]

Everybody converges to the same preferences in speakers? That would be unknown news to us that follow the audio scene?

Part 1- Do Untrained Listeners Prefer the Same Loudspeakers as Trained Listeners?

One of the more controversial topics among audio researchers is whether or not trained listeners should be used for audio product testing an...

seanolive.blogspot.com

 

[MAB]

Part 1- Do Untrained Listeners Prefer the Same Loudspeakers as Trained Listeners?

One of the more controversial topics among audio researchers is whether or not trained listeners should be used for audio product testing an...

seanolive.blogspot.com

View attachment 405786

It's a great study.
The four speakers speakers in the study do have preference predicted from measurements.

Part 3 - Relationship between Loudspeaker Measurements and Listener Preferences

Part 1 of this article presented experimental evidence from a study conducted by the author demonstrating that trained and untrained liste...

seanolive.blogspot.com

"There are clear visual correlations between listeners' loudspeaker preferences and the set of frequency graphs. Both trained and untrained listeners clearly preferred the loudspeakers with the flattest, smoothest and most extended frequency response curves, as exhibited in the measurements of loudspeakers P and I. Loudspeaker B was rated lower due to its less extended, bumpy bass, and a large hole centered at 3 kHz in its sound power curve. The measurements of Loudspeaker M indicate it has a lack of low bass, and has a non-smooth frequency response in all of its measured curves. Both the direct and reflected sounds produced by this loudspeaker will contribute serious colorations to the timbre of reproduced sounds.

It is both satisfying and reassuring to know that both trained and untrained listeners recognize and prefer accurate loudspeakers, and that the accuracy can be characterized with a set of comprehensive anechoic measurements. The next logical step is to use these technical measurements as the basis for modeling and predicting listeners' preference ratings. This will be the topic of a future post in this blog."

 

[Newman]

I think Toole and Olive’s work is indicative of a majority preference, but I think you have perhaps overstated it. We really don’t have large N trials to go on.

Well, to say 'majority preference' is to very much understate it. That gives the entirely wrong impression that, hey, maybe only 55%, or 66%, or 75%, gave speaker A that preference. If you look at the confidence intervals on their preference ratings of individual speakers, they are very tight indeed. Such confidence levels also suggest that the trial N is sufficient for the purpose of drawing the conclusion drawn.

Also, Olive has over the years conducted informal facility demonstrations that include eliciting preferences from visitors that number in the hundreds if not 1000+, and IIRC has said that, if he predicts from the measurements that you would prefer speaker A over B at the facility, then it would be truly extraordinary if you didn't. That's not the same as saying something like 'then a majority of you would prefer A over B'.

cheers