Do driver materials influence subjectively perceived sound textures if measurements are identical?

[confucius_zero]

Some headphones use berylium, some speaker cones are paper instead of wtv composites, some tweeters are AMT-ribbon instead of metal or silk dome etc... Subjectivitists enjoy discussing such nuances and subtleties that are part of the analogic build design decisions of speakers, headphones etc... but I'm wondering

If all things equal measuring on the machine, do driver materials influence subjectively perceived sound textures if measurements are identical?

 

[levimax]

Some headphones use berylium, some speaker cones are paper instead of wtv composites, some tweeters are AMT-ribbon instead of metal or silk dome etc... Subjectivitists enjoy discussing such nuances and subtleties that are part of the analogic build design decisions of speakers, headphones etc... but I'm wondering

If all things equal measuring on the machine, do driver materials influence subjectively perceived sound textures if measurements are identical?

People want to believe in magic so of course Beryllium or whatever expensive material must be more magic than a plain old silk dome. The reality is that in theory if the driver and the speaker and the crossover design were all optimized for Beryllium it has the potential for higher performance but that is seldom the case in the real world. If they measure the same they are going to sound the same.

​

 

[MaxwellsEq]

There are lots of threads on ASR discussing materials and their impact on sound. The reality is that a great speaker is a combination of drivers, cabinets, crossovers and (most importantly) decisions made about compromises.

It's very difficult to maximise the the benefits from a specific material as a result. If it was possible to make two drivers with the same characteristics, from different materials, then I don't think they would sound wildly different.

Consider the BBC LS3/5A. It has a primary goal of being consistent unit to unit over many years. This was a big challenge since manufacturing tolerances are always on the move, making matching drivers more and more expensive. Matching wasn't just about frequency response, but also impedance, stiffness etc.

 

[antcollinet]

Measurements are never idetical accross different transducers. Some of the differences may come from the materials used. The real question comes down to how audible the differences are.

Framed as in the OP - if you were to find two different transducers with identical measurements - then I'd suggest the materails used would make no additional "audible-but-unmeasureable" difference.

 

[Cote Dazur]

Subjectivists enjoy discussing such nuances and subtleties

If only they could back those "nuance and subtleties" with unsighted bias, then it would be interesting to pursue.

Unfortunately, speakers measurements only tell part of the story at this point in time, as opposed to measuring electronic gear, leaving room for speculations on what is really at play. Adding that the interaction with the room also plays a big role, if someone convinced themselves that "Beryllium" is giving them better performance, it will be hard to objectively argue with them.

 

[ctrl]

If all things equal measuring on the machine, do driver materials influence subjectively perceived sound textures if measurements are identical?

As has been said, this topic has been discussed ad nauseam in all audio forums around the world.
Common sense and acceptance of physical realities provides already the answer, but it would be even better to have some kind of proof.

There is a German-language master's thesis on this subject that investigated the subjective perceptibility of sound differences with different types and materials of tweeters:
Perceptibility of tonal differences of tweeters with different principles of operation - "Wahrnehmbarkeit klanglicher Unterschiede von Hochtonlautsprechern unterschiedlicher Wirkprinzipien".

The following tweeters were tested together with identical woofer and a crossover frequency of 2kHz:

In order to exclude typical colorations of the sound due to room interaction as well as with the loudspeaker baffle, all drivers were tested in a quasi infinite baffle in a reflexion-free environment.

The on-axis frequency responses of the combinations of tweeter with woofer were phase (via FIR) and frequency response equalized (the results are not perfect but within +-0.75dB).


A dummy head was then used to create the HRTF via binaural impulse responses (BIR) for all six speakers (six different tweeters, always the same woofer) and the speaker pairing with the largest difference in HRTF was selected for an ABX listening test (in simplified terms, the exact details are in the study).
The selected tweeter pairing consisted of a magnetostatic tweeter and a fabric ring radiator.

To allow a dynamic simulation of the loudspeaker pairs in a later comparison, the BIRs were recorded at a radius of typical head movements.
With the help of the binaural, dynamic simulations of the loudspeaker pairs, ten test persons were asked to find out, if there is still a difference between the tweeters which can be determined by the subjective sound image. For this purpose, music and artificial signals were folded/convolved with the BIR's and played back through headphones.

All subjects stated that they had not found any auditory clue by which they could have identified one of the two tweeters. All hits were in the range of 50%, which suggests more or less arbitrary guessing by the subjects due to undetectable differences. The result of the evaluation of the hit rates over all test subjects is thus quasi confirmed.


Thus, when using tweeters with different design principles and diaphragm materials, no difference could be detected in stereo listening in the free field after phase and frequency response equalization.

So no one could distinguish the "super fast" membrane of the magnetostat tweeter from the fabric tweeter
No one was bothered by the different decay and distortion behavior of the tweeters - which highlights the low importance of these parameters (as long as there are no gross errors).


The chapter on the current state of research (which was 2010) summarizes interesting older work:

Pflaum [Pfl95] equalized different pairs of speakers axially linear in magnitude and phase and recorded music material played through them once recorded monaurally and once with an dummy head, i.e. binaurally.

In a listening test, subjects were asked to decide on the basis of the recordings whether they listened to the material from the same or from different speaker.

The evaluation of the listening test showed that the dummy head recordings made it possible to distinguish the loudspeakers significantly, whereas the monaurally recorded signals no longer allowed a significant distinction between the loudspeakers.

Pflaum concluded that linearly equalized loudspeakers, with the different effects of spatial radiation removed, could no longer be reliably distinguished.

Müller hypothesizes here that the discriminability of loudspeakers among each other is thus due to different magnitude frequency responses at the tympanic membrane, which result from diffraction and reflexion at the artificial head depending on the acoustic wavefields of the loudspeakers.

The last section also provides the explanation why many think they hear a difference with different tweeter membrane materials:
When using different materials, the radiation of the loudspeaker changes and this can lead to different auditory perceptions, but not the material itself, the same applies to the different operating principles for tweeters.

The deviations in radiation that can occur with completely identical tweeters but with different diaphragm materials were once shown here, for example.

 

[MaxwellsEq]

As has been said, this topic has been discussed ad nauseam in all audio forums around the world.
Common sense and acceptance of physical realities provides already the answer, but it would be even better to have some kind of proof.

There is a German-language master's thesis on this subject that investigated the subjective perceptibility of sound differences with different types and materials of tweeters:
Perceptibility of tonal differences of tweeters with different principles of operation - "Wahrnehmbarkeit klanglicher Unterschiede von Hochtonlautsprechern unterschiedlicher Wirkprinzipien".

The following tweeters were tested together with identical woofer and a crossover frequency of 2kHz:
View attachment 262198
In order to exclude typical colorations of the sound due to room interaction as well as with the loudspeaker baffle, all drivers were tested in a quasi infinite baffle in a reflexion-free environment.

The on-axis frequency responses of the combinations of tweeter with woofer were phase (via FIR) and frequency response equalized (the results are not perfect but within +-1dB).
View attachment 262203

A dummy head was then used to create the HRTF via binaural impulse responses (BIR) for all six speakers (six different tweeters, always the same woofer) and the speaker pairing with the largest difference in HRTF was selected for an ABX listening test (in simplified terms, the exact details are in the study).
The selected tweeter pairing consisted of a magnetostatic tweeter and a fabric ring radiator.

To allow a dynamic simulation of the loudspeaker pairs in a later comparison, the BIRs were recorded at a radius of typical head movements.
With the help of the binaural, dynamic simulations of the loudspeaker pairs, ten test persons were asked to find out, if there is still a difference between the tweeters, which can be determined by the subjective sound image - for this purpose, music and artificial signals were folded with the BIR.

All subjects stated that they had not found any auditory clue by which they could have identified one of the two tweeters. All hits were in the range of 50%, which suggests more or less arbitrary guessing by the subjects due to undetectable differences. The result of the evaluation of the hit rates over all test subjects is thus quasi confirmed.
View attachment 262264

Thus, when using tweeters with different design principles and diaphragm materials, no difference could be detected in stereo listening in the free field after phase and frequency response equalization.

So no one could distinguish the "super fast" membrane of the magnetostat tweeter from the fabric tweeter
No one was bothered by the different decay and distortion behavior of the tweeters - which highlights the low importance of these parameters (as long as there are no gross errors).


The chapter on the current state of research (which was 2010) summarizes interesting older work:






The last section also provides the explanation why many think they hear a difference with different tweeter membrane materials:
When using different materials, the radiation of the loudspeaker changes and this can lead to different auditory perceptions, but not the material itself, the same applies to the different operating principles for tweeters.

The deviations in radiation that can occur with completely identical tweeters with different diaphragm materials were once shown here, for example.

Fascinating. I'd not seen that.

 

[Multicore]

The question was asked in two different phrasings.

Do driver materials influence subjectively perceived sound textures if measurements are identical?

If all things equal measuring on the machine, do driver materials influence subjectively perceived sound textures if measurements are identical?

Either way, I think the answer is that it very well might. If the subject thinks that materials make a difference then she may be able to hear the expected difference.

If we additionally assume a properly blinded test then I think it depends. Are all the potential factors that can influence subjective perception accounted for in the measurements? I really don't know. I remember a video with Dan Clark presenting his work and answering questions. Someone asked about stereo image or sound-stage or something like that and Clark's answer (something about angled orientation of the driver plane, iirc) suggested to me that such matters aren't fully understood yet. So are the headphone measurements typically used today sufficient to predict all the perceptual regularities? Idk.

 

[ctrl]

Are all the potential factors that can influence subjective perception accounted for in the measurements?

The eardrum and the microphone record sound pressure fluctuations, so from that point of view all information is contained in it.

As far as the spatial (partly also tonal***) information is concerned, which our brain receives due to the individual shape of the outer ear, positions of the two ears,... i.e. the HRTF, not all information is included in the BIR (binaural impulse responses) - see below.

*** see e.g. timbre change from monophonic to stereo (see post#2), depends on the individual HRTF

So are the headphone measurements typically used today sufficient to predict all the perceptual regularities? Idk.

BIRs (binaural impulse responses) were recorded for each tweeter/speaker using a dummy head (for different head positions in the range +-80° for the left and right speaker).
This then allows the BIR made in the anechoic chamber to be convolved in real time, taking into account the head position, with any audio signals (music, pink noise,...) and transmitted to the listener via headphones:

The headtracker was attached directly to the headphone to determine the subject's head position. It continuously sent magnetic position data to a receiver, which in turn transmitted information to the convolution engine for correct selection of a BIR pair matching the subject's head direction from a "left" and a "right" speaker, where they were then convolved in parallel.

The criticism is that the HRTF of the dummy head and thus the signal fed to the subject/listener via headphones, of course, does not correspond exactly to the signal generated with the individual HRTF of the subject/listener.

This means that the spatial information the listener can derive from the binaural signal of the dummy head source is not as good as if he were sitting directly in front of the speakers (with his individual HRTF "at work") in the anechoic chamber.

For our question, however, this is even advantageous, since we are not asking whether the tweeters differ in their spatial radiation (which of course they do and as shown above is very likely the answer), but whether the tweeter material (and the construction concept) are decisive for the sound. So if some spatial information is lost in this type of experimental setup, so much the better.

 

[pablolie]

Some headphones use berylium, some speaker cones are paper instead of wtv composites, some tweeters are AMT-ribbon instead of metal or silk dome etc... Subjectivitists enjoy discussing such nuances and subtleties that are part of the analogic build design decisions of speakers, headphones etc... but I'm wondering

If all things equal measuring on the machine, do driver materials influence subjectively perceived sound textures if measurements are identical?

If they measure the same in every way, at first I'd doubt you'd get a "different perception" other than visual and preconceptions.

But the other question is - do all materials sound the same over time and retain their original specs? It's often said paper cones age and decay over time. OTOH I have a pair of speakers with paper cones I sometimes cycle in for fun, and they still sound awesome after close to 30 years.

 

[fredstuhl]

Also, i haven‘t noticed a positive correlation of expensive/ exotic cone materials with loudspeaker performance in Amir‘s tests. In fact, I wouldn‘t be surprised if there was actually a reverse correlation between fancy cone materials (and marketing that is centered on the use of such) and objective loudspeaker performance, as measurement-oriented speaker design companies tend not to bother with magic ingredients.

 

[Cars-N-Cans]

Some headphones use berylium, some speaker cones are paper instead of wtv composites, some tweeters are AMT-ribbon instead of metal or silk dome etc... Subjectivitists enjoy discussing such nuances and subtleties that are part of the analogic build design decisions of speakers, headphones etc... but I'm wondering

If all things equal measuring on the machine, do driver materials influence subjectively perceived sound textures if measurements are identical?

On our end the frequency response, distortion characteristics, and radiation pattern for speakers, or headphone driver and cup layout/type of headphone will be what we ultimately hear in regards to imaging. I would say the material used is more a design choice than anything else and something we should not really be too concerned with as end-users, only the capabilities of the device.

 

[valerianf]

Let Us be objective, I have to make citation of the post of @ctrl:
"
The different physical parameters (such as stiffness, material sound velocity, specific weight,...) lead to different radiation characteristics of the tweeters, which in turn explains the tonal differences."
It is in the thread :

BlieSMa T25A-6, T25B-6, T25D-6, T25S-6 - a comparison of dome materials

https://hificompass.com/en/reviews/bliesma-t25a-6-t25b-6-t25d-6-t25s-6 I thought this was a really neat writeup about some high-end tweeters. For me, the most interesting part was how silk performed. T25A-6 = Aluminium + magnesium T25B-6 = beryllium T25D-6 = diamond T25S-6 = silk Thanks...

www.audiosciencereview.com

It is well known that the speaker material is changing the break up behavior of the speaker.
This can enforce or distord some frequencies.
It is a phenomenon present in all speakers, not only the tweeter.

 

[antcollinet]

Let Us be objective, I have to make citation of the post of @ctrl:
"
The different physical parameters (such as stiffness, material sound velocity, specific weight,...) lead to different radiation characteristics of the tweeters, which in turn explains the tonal differences."
It is in the thread :

BlieSMa T25A-6, T25B-6, T25D-6, T25S-6 - a comparison of dome materials

https://hificompass.com/en/reviews/bliesma-t25a-6-t25b-6-t25d-6-t25s-6 I thought this was a really neat writeup about some high-end tweeters. For me, the most interesting part was how silk performed. T25A-6 = Aluminium + magnesium T25B-6 = beryllium T25D-6 = diamond T25S-6 = silk Thanks...

www.audiosciencereview.com

It is well known that the speaker material is changing the break up behavior of the speaker.
This can enforce or distord some frequencies.
It is a phenomenon present in all speakers, not only the tweeter.

right.

But - those differences will show up in the measurements.

 

[MaxwellsEq]

Let Us be objective, I have to make citation of the post of @ctrl:
"
The different physical parameters (such as stiffness, material sound velocity, specific weight,...) lead to different radiation characteristics of the tweeters, which in turn explains the tonal differences."
It is in the thread :

BlieSMa T25A-6, T25B-6, T25D-6, T25S-6 - a comparison of dome materials

https://hificompass.com/en/reviews/bliesma-t25a-6-t25b-6-t25d-6-t25s-6 I thought this was a really neat writeup about some high-end tweeters. For me, the most interesting part was how silk performed. T25A-6 = Aluminium + magnesium T25B-6 = beryllium T25D-6 = diamond T25S-6 = silk Thanks...

www.audiosciencereview.com

It is well known that the speaker material is changing the break up behavior of the speaker.
This can enforce or distord some frequencies.
It is a phenomenon present in all speakers, not only the tweeter.

The original post asked "If all things equal measuring on the machine, do driver materials influence subjectively perceived sound textures if measurements are identical?"

The things you point out are well known by everyone - BUT a headphone or speaker designer if aiming for an accurate frequency response and minimum distortion will build cabinets/crossovers/headphones to compensate for peaks, troughs and dispersion. The question from the OP was effectively - if you do all the corrections so that two very different drivers are measuring the same, will we hear a difference. The objective consensus so far is, possibly not.

 

[valerianf]

"if you do all the corrections so that two very different drivers are measuring the same"
Let us take the crossover of a 3 way speaker.
BMR audio need 90 element crossover to be able to get a suitable FR curve.
Not many speaker designers are taking this precaution.
Better choice is to find a material for the driver that bring a flat FR curve with break up peaks as small as possible.
This is a real world choice: ideal cross-over does not exist, and no crossover remove fully a tweeter break-up.

 

[MaxwellsEq]

"if you do all the corrections so that two very different drivers are measuring the same"
Let us take the crossover of a 3 way speaker.
BMR audio need 90 element crossover to be able to get a suitable FR curve.
Not many speaker designers are taking this precaution.
Better choice is to find a material for the driver that bring a flat FR curve with break up peaks as small as possible.
This is a real world choice: ideal cross-over does not exist, and no crossover remove fully a tweeter break-up.

You are still missing the point.

No-one disagrees with you. These things are all objectively measurable.

The OP wondered whether, if they were corrected for (which may be impossible or financially unacceptable), would there be an audible difference? The jury is out, but there seems to be evidence that they would not be audible.

 

[HarmonicTHD]

As has been said, this topic has been discussed ad nauseam in all audio forums around the world.
Common sense and acceptance of physical realities provides already the answer, but it would be even better to have some kind of proof.

There is a German-language master's thesis on this subject that investigated the subjective perceptibility of sound differences with different types and materials of tweeters:
Perceptibility of tonal differences of tweeters with different principles of operation - "Wahrnehmbarkeit klanglicher Unterschiede von Hochtonlautsprechern unterschiedlicher Wirkprinzipien".

The following tweeters were tested together with identical woofer and a crossover frequency of 2kHz:
View attachment 262198
In order to exclude typical colorations of the sound due to room interaction as well as with the loudspeaker baffle, all drivers were tested in a quasi infinite baffle in a reflexion-free environment.

The on-axis frequency responses of the combinations of tweeter with woofer were phase (via FIR) and frequency response equalized (the results are not perfect but within +-0.75dB).
View attachment 262203

A dummy head was then used to create the HRTF via binaural impulse responses (BIR) for all six speakers (six different tweeters, always the same woofer) and the speaker pairing with the largest difference in HRTF was selected for an ABX listening test (in simplified terms, the exact details are in the study).
The selected tweeter pairing consisted of a magnetostatic tweeter and a fabric ring radiator.

To allow a dynamic simulation of the loudspeaker pairs in a later comparison, the BIRs were recorded at a radius of typical head movements.
With the help of the binaural, dynamic simulations of the loudspeaker pairs, ten test persons were asked to find out, if there is still a difference between the tweeters which can be determined by the subjective sound image. For this purpose, music and artificial signals were folded/convolved with the BIR's and played back through headphones.

All subjects stated that they had not found any auditory clue by which they could have identified one of the two tweeters. All hits were in the range of 50%, which suggests more or less arbitrary guessing by the subjects due to undetectable differences. The result of the evaluation of the hit rates over all test subjects is thus quasi confirmed.
View attachment 262264

Thus, when using tweeters with different design principles and diaphragm materials, no difference could be detected in stereo listening in the free field after phase and frequency response equalization.

So no one could distinguish the "super fast" membrane of the magnetostat tweeter from the fabric tweeter
No one was bothered by the different decay and distortion behavior of the tweeters - which highlights the low importance of these parameters (as long as there are no gross errors).


The chapter on the current state of research (which was 2010) summarizes interesting older work:






The last section also provides the explanation why many think they hear a difference with different tweeter membrane materials:
When using different materials, the radiation of the loudspeaker changes and this can lead to different auditory perceptions, but not the material itself, the same applies to the different operating principles for tweeters.

The deviations in radiation that can occur with completely identical tweeters but with different diaphragm materials were once shown here, for example.

Great post. Thx.

Unfortunately, as always, no matter what facts one brings to the table, it won't convince the ignorant.

 

[Multicore]

Thanks for the interesting reply and link.

The eardrum and the microphone record sound pressure fluctuations, so from that point of view all information is contained in it.

Agreed, from that point of view. I am trying to be cautions in what assumptions we make without stating. For one thing, in visual perception the patterns of light that fall on the retinas is rather different from what subject report seeing. Second, reading @Floyd Toole 's book gives me the impression that we are capable of some similarly clever tricks in hearing.

But perhaps it is the case that the measured sound pressure signal at the simulated eardrum captures everything that a headphone can do to a human listener. It seems likely. Now, assuming that's the case, what is OP's question driving at?

 

[Multicore]

Unfortunately, as always, no matter what facts one brings to the table, it won't convince the ignorant.

Idk if you're referring to me here. Doesn't really matter because ignorance is what I bring to ASR. It's one of the two pre-conditions to learning. The other precondition is to hold wrong beliefs. Either way learning can change the understanding, either from ignorance to something, or from something wrong to something else. So it seems more likely to me that the one with the correct understanding is the one least likely to be convinced of anything.