In IEMs - Is there a scientific proof that BA Timbre is a real thing? What about "Speed"?

[Fraxo]

This whole topic is getting annoying. I don't need people to repeat the same facts about different technologies of drivers, Planar\BA\DD etc.. But can someone actually prove that there are differences beyond FR? An actual audible proof or something that impossible to refute? Debates about this are so stupid at times...
People are going at this all the damn time, I need some science guys, I'm a sound engineer, ready to dive in as deep as necessary so bring it on if you have it and I'll nerd along the ride!

 

[JanesJr1]

This whole topic is getting annoying. I don't need people to repeat the same facts about different technologies of drivers, Planar\BA\DD etc.. But can someone actually prove that there are differences beyond FR? An actual audible proof or something that impossible to refute? Debates about this are so stupid at times...
People are going at this all the damn time, I need some science guys, I'm a sound engineer, ready to dive in as deep as necessary so bring it on if you have it and I'll nerd along the ride!

I don't have the answer, but I've gone down this rabbit hole a couple of times, and remain confused. Is there something real and measurable related to "detail", "speed", "note decay", and "settling time" with respect to BA's. I'm tentatively convinced that "speed" may be entirely captured by FR + power; but there's more subjective fuzz here, that I'd like to understand.

 

[majingotan]

The forum that must not be named touched on that topic with 3rd harmonics, burst cycle plots as well as "attack and decay" plots aka 2D plot of a CSD graph. Subjectively, a really poor CSD plot can be incorrectly perceived as "fast" due to various harmonics that lingers during playing lots of notes, leading to a false sense of perceived detail. Also, it can also be perceived subjectively correctly as "slow" where notes pretty much smear during complex songs. IMHO, doesn't matter what type of driver is in there but implementation rather. The following is an opinion only as objective statements cannot be accurately translated to subjective perception due to my lack of expertise: BA Timbre can be a result of odd harmonic peaks (like IMD but starting on 3KHz and beyond (music), not necessarily 19 and 20 KHz fundamentals) while music is playing, and can easily be eliminated with proper BA implementation since I've subjectively heard IEMs with that "plastic BA Timbre" while other BA IEMs have that speaker-like (lacking that artificial tone) timbre.

 

[cheapmessiah]

The way "speed" is used on this hobby by people who are not really versed on the thechnicalities of mechanics, fluid dynamics or electronics pisses me off a little bit, what does that term even mean for them? How can a driver have more speed than another one and be reproducing the same frequency at the same amplitude (volume) within the limitations of their THDs? I can understand inertia (and rigidity to some extent) and how a driver that has very little of it might be better at transducing the micro details of an analog electronic wave into the nuances that compose a complex sound, but speed per se should not be significatively different because we would be getting significative inacuraccies between the signal frequency and the sound frequency.

Differences in timbre I do believe in, how a transducer is suceptible to be affected by harmonic load can change how it complies mechanicaly to the electronic signal thats fed into them. Some of the harmonic overtones can resonate more than others, and some others can be dampened much faster than others and variations of the resonances between drivers can create such differences in timbre that are analogous to a violin and a trumpet playing the same note and yet produce sound thats distinctevely different.

When I hear about the BA timbre it reminds me of how 2 identic bass guitars where one has a wooden neck and the other has a carbon fiber neck produce such different timbre, specially noticeable on the high frequency harmonic load thats not dampened as fast as in the wooden neck one, producing an "abnormally" long lasting ringing because the CF neck has the resonant frequencies at much higher values than the wood one.

And I guess that where i want to get to with this is that when an object vibrates, much like the voice coil of a speaker (or a BA or planar magnetic driver diafragm) when its moving, the movement it is describing is a combination of whats exciting it and how it is reacting to such excitation, and 2 different drivers can react slightly differently while still being quite true to the signal thats being fed into them, and hence have different timbres.

If I made any mistakes in what i said feel free to correct me as I too would love to know more and understand more about accoustics and electronics on this hobby.

 

[Earfonia]

I'm not sure if this is what we should be observing:

When I tried KZ ZEX, an Electrostatic + Dynamic Hybrid, besides the relatively warn tonality, it sounds very dry, very fast decay that my ears are not used to. I was wondering what am I hearing. It is warm, but dry. Usually the dry sounding IEM is not warm but analytical. So KZ ZEX is quite unique.

I measured it, and when observing the measurement I saw the decay T30 time is like 1/3 of ER4XR (single BA) and BLON BL-05 (single dynamic).

This is the KZ ZEX frequency measurement compared to my EITC-2021 target curve:

I'm not sure if we are looking at something that might explain / measure IEM decay characteristic that might be interpreted as that 'timbre' thingy. I'm not an expert and would like to hear expert's opinions as well.

 

[JanesJr1]

The forum that must not be named touched on that topic with 3rd harmonics, burst cycle plots as well as "attack and decay" plots aka 2D plot of a CSD graph. Subjectively, a really poor CSD plot can be incorrectly perceived as "fast" due to various harmonics that lingers during playing lots of notes, leading to a false sense of perceived detail. Also, it can also be perceived subjectively correctly as "slow" where notes pretty much smear during complex songs. IMHO, doesn't matter what type of driver is in there but implementation rather. The following is an opinion only as objective statements cannot be accurately translated to subjective perception due to my lack of expertise: BA Timbre can be a result of odd harmonic peaks (like IMD but starting on 3KHz and beyond (music), not necessarily 19 and 20 KHz fundamentals) while music is playing, and can easily be eliminated with proper BA implementation since I've subjectively heard IEMs with that "plastic BA Timbre" while other BA IEMs have that speaker-like (lacking that artificial tone) timbre.

If I (over)simplify that to a non-technical summary, there is the music signal, transformed into a sonic output that more or less matches. In reproducing a note or any burst of sound over time with a "fast" or "detailed" presentation, there are minimal harmonics or lingering resonances at the end of a note that could time-smear the match between signal and sonic output and thereby mask transient details. A "slow" presentation would be the opposite.

With respect to "plastic" BA timbre, which I believe is Crinacle's adjective, I think it was you, majingotan, who previously suggested on this forum that this might only be the tendency of BA drivers to have a weak low-end FR (possibly in association with a dry or "fast" presentation of detail in the higher frequencies?) This is not quite what you say above, which implies that some BA drivers MAY have odd-harmonic distortion. I wonder what the diff is for the non-plastic IEM's...?

I think I agree, at least with the FR part of the BA timbre: I confess that I have found Etymotic BA IEM's hard to EQ, and that I do not accept Oratory's proposed EQ's for any of my Ety's as natural-sounding, even though I generally like Oratory EQ's as a starting point for other drivers. However, I now have what sounds like a pretty natural-sounding presentation from my ER4XR's with a much-steeper cut in the high frequencies from 2-4kHz and up on Eq APO. (This doesn't sound plausible; I can't explain it but it is subjectively true.) It doesn't sound significantly different from my dynamic and planar drivers, except that the low's are mostly there (after EQ) but seem subjectively less resonant than with dynamic drivers. The mid's and high's are very detailed, withOUT harshness, and the timbre seems accurate. With the low's there is a difference, but whether it is subjectively good or bad depends on the material: sometimes a more lush or resonant low-end would seem helpful; other times a "tight bass" (= more accurate?) seems to be called for and the BA driver delivers it with a smooth transition from bass to mid's.

 

[JanesJr1]

The way "speed" is used on this hobby by people who are not really versed on the thechnicalities of mechanics, fluid dynamics or electronics pisses me off a little bit, what does that term even mean for them? How can a driver have more speed than another one and be reproducing the same frequency at the same amplitude (volume) within the limitations of their THDs? I can understand inertia (and rigidity to some extent) and how a driver that has very little of it might be better at transducing the micro details of an analog electronic wave into the nuances that compose a complex sound, but speed per se should not be significatively different because we would be getting significative inacuraccies between the signal frequency and the sound frequency.

Differences in timbre I do believe in, how a transducer is suceptible to be affected by harmonic load can change how it complies mechanicaly to the electronic signal thats fed into them. Some of the harmonic overtones can resonate more than others, and some others can be dampened much faster than others and variations of the resonances between drivers can create such differences in timbre that are analogous to a violin and a trumpet playing the same note and yet produce sound thats distinctevely different.

When I hear about the BA timbre it reminds me of how 2 identic bass guitars where one has a wooden neck and the other has a carbon fiber neck produce such different timbre, specially noticeable on the high frequency harmonic load thats not dampened as fast as in the wooden neck one, producing an "abnormally" long lasting ringing because the CF neck has the resonant frequencies at much higher values than the wood one.

And I guess that where i want to get to with this is that when an object vibrates, much like the voice coil of a speaker (or a BA or planar magnetic driver diafragm) when its moving, the movement it is describing is a combination of whats exciting it and how it is reacting to such excitation, and 2 different drivers can react slightly differently while still being quite true to the signal thats being fed into them, and hence have different timbres.

If I made any mistakes in what i said feel free to correct me as I too would love to know more and understand more about accoustics and electronics on this hobby.

Thank you for a plain-English reply! I think I get the part about "speed" being kind of specious as a concept, since an adequately-powered transducer operating at a given frequency is by definition operating with the speed necessary to reproduce the signal (a tautology). Beyond that, I infer you believe there may different resonances at different frequencies for BA vs. dynamic drivers, resulting in a different timbre. Since the resonances or harmonics occur not as the intentional resonance of the physical musical instrument but rather the error of the transducer in conveying the musical signal, it is distortion. Not far afield from majingotan's tentative speculation about odd-harmonic distortion with some BA drivers (previous posting) is it?

Seems measurable, no?

 

[JanesJr1]

I'm not sure if this is what we should be observing:

View attachment 189434

View attachment 189433

View attachment 189432


When I tried KZ ZEX, an Electrostatic + Dynamic Hybrid, besides the relatively warn tonality, it sounds very dry, very fast decay that my ears are not used to. I was wondering what am I hearing. It is warm, but dry. Usually the dry sounding IEM is not warm but analytical. So KZ ZEX is quite unique.

I measured it, and when observing the measurement I saw the decay T30 time is like 1/3 of ER4XR (single BA) and BLON BL-05 (single dynamic).

This is the KZ ZEX frequency measurement compared to my EITC-2021 target curve:

View attachment 189435

I'm not sure if we are looking at something that might explain / measure IEM decay characteristic that might be interpreted as that 'timbre' thingy. I'm not an expert and would like to hear expert's opinions as well.

Thank you for making this graphically plain. My question is: if a fast-decay driver is subjectively "dry", wouldn't it also be more accurate? Isn't any post-impulse resonance a form of distortion?

 

[JanesJr1]

The forum that must not be named touched on that topic with 3rd harmonics, burst cycle plots as well as "attack and decay" plots aka 2D plot of a CSD graph. Subjectively, a really poor CSD plot can be incorrectly perceived as "fast" due to various harmonics that lingers during playing lots of notes, leading to a false sense of perceived detail. Also, it can also be perceived subjectively correctly as "slow" where notes pretty much smear during complex songs. IMHO, doesn't matter what type of driver is in there but implementation rather. The following is an opinion only as objective statements cannot be accurately translated to subjective perception due to my lack of expertise: BA Timbre can be a result of odd harmonic peaks (like IMD but starting on 3KHz and beyond (music), not necessarily 19 and 20 KHz fundamentals) while music is playing, and can easily be eliminated with proper BA implementation since I've subjectively heard IEMs with that "plastic BA Timbre" while other BA IEMs have that speaker-like (lacking that artificial tone) timbre.

PS, I kind of am missing the joke on the "Forum that must not be named ..." and would like to ask what it is, but I can't do that apparently ...

 

[charleski]

PS, I kind of am missing the joke on the "Forum that must not be named ..." and would like to ask what it is, but I can't do that apparently ...

I'm guessing this was an allusion to SBAF.

 

[JanesJr1]

The way "speed" is used on this hobby by people who are not really versed on the thechnicalities of mechanics, fluid dynamics or electronics pisses me off a little bit, what does that term even mean for them? How can a driver have more speed than another one and be reproducing the same frequency at the same amplitude (volume) within the limitations of their THDs? I can understand inertia (and rigidity to some extent) and how a driver that has very little of it might be better at transducing the micro details of an analog electronic wave into the nuances that compose a complex sound, but speed per se should not be significatively different because we would be getting significative inacuraccies between the signal frequency and the sound frequency.

Differences in timbre I do believe in, how a transducer is suceptible to be affected by harmonic load can change how it complies mechanicaly to the electronic signal thats fed into them. Some of the harmonic overtones can resonate more than others, and some others can be dampened much faster than others and variations of the resonances between drivers can create such differences in timbre that are analogous to a violin and a trumpet playing the same note and yet produce sound thats distinctevely different.

When I hear about the BA timbre it reminds me of how 2 identic bass guitars where one has a wooden neck and the other has a carbon fiber neck produce such different timbre, specially noticeable on the high frequency harmonic load thats not dampened as fast as in the wooden neck one, producing an "abnormally" long lasting ringing because the CF neck has the resonant frequencies at much higher values than the wood one.

And I guess that where i want to get to with this is that when an object vibrates, much like the voice coil of a speaker (or a BA or planar magnetic driver diafragm) when its moving, the movement it is describing is a combination of whats exciting it and how it is reacting to such excitation, and 2 different drivers can react slightly differently while still being quite true to the signal thats being fed into them, and hence have different timbres.

If I made any mistakes in what i said feel free to correct me as I too would love to know more and understand more about accoustics and electronics on this hobby.

A PS on the idea of different harmonics at different frequencies for balanced armature vs dynamic drivers, I recalled Crinacle's comment that the Etymotic ER2XR (with a dynamic driver) has almost the same FR as the Etymotic ER4XR (with a balanced armature driver), except that the ER2 version has better lows. I thought it would be good to find the waterfall plot for each driver to compare, to see if there were obvious differences between the BA and DD drivers in resonance and decay. (Unless I misunderstand how to use those graphs.) Anyway, I couldn't find such plots on ASR, Crinacle or Rtings. If anyone has an idea of where to get such info, it might have something.

 

[JanesJr1]

I'm guessing this was an allusion to SBAF.

Should I feel humiliated that I am still lost in the woods? The only "SBA Forum" I can find is that of the Simulation Basketball Association...

 

[tifune]

Should I feel humiliated that I am still lost in the woods? The only "SBA Forum" I can find is that of the Simulation Basketball Association...

Super Best Audio Friends (although I don't see much friendliness there myself). A few ostensibly well informed mods but a lot of negative attitudes

 

[Earfonia]

Thank you for making this graphically plain. My question is: if a fast-decay driver is subjectively "dry", wouldn't it also be more accurate? Isn't any post-impulse resonance a form of distortion?

I'm not sure if everyone has the same understanding and definition of 'dry' sound. My definition coming from room reverberations when wet is associated with more reverberations. And audio mixing when we use to call dry signal before we apply any reverb.

I guess theoretically the drier the more accurate. What I'm saying was, I was not used to that quality having used to other IEMs with longer decay.

And most recordings were mixed using studio monitors in room that I guess won't have decay properties as short as IEM. Just my guess. What I mean is, when I apply reverb to my recording the amount is depending on the monitoring environment during mixing. So even the shorter decay monitor is probably the more accurate monitor but the recordings that we listen might not be intended to be listened using such monitor.

 

[RHO]

There are plenty IEM models using the same shell. Shouldn't be too hard to set up a blind test with single DD, hybrid and full BA models to test if the attributes linked to the different driver types are real or not.

 

[JanesJr1]

Super Best Audio Friends (although I don't see much friendliness there myself). A few ostensibly well informed mods but a lot of negative attitudes

Thanks. I confess to a web search on SBAF. Didn't show up on any of the several best-audiophile-websites lists, go figure.

 

[TheBatsEar]

This whole topic is getting annoying. I don't need people to repeat the same facts about different technologies of drivers, Planar\BA\DD etc.. But can someone actually prove that there are differences beyond FR?

You could take a distortion measurement.

A speakers membrane swings never in and out as clean as the input signal, it adds distortion, no matter the technology.
If the distortion is inaudible, the device is considered transparent (the signal is indistinguishable from the original to human ears). If the distortion is large enough to be audible, the device should be considered defective and reenter the engineering cycle.

Everything else (speed, wideness, authority) is imagination. A rational sound engineer will not use those terms because they are undefined.

An actual audible proof or something that impossible to refute? Debates about this are so stupid at times...

Because it debates belief, not fact. I'm with you, it's stupid and proof that indeed, all the people, including us, are full of biases and assumptions and plenty of BS.

I'm a sound engineer, ready to dive in as deep as necessary so bring it on if you have it and I'll nerd along the ride!

I think you need to understand that distortion in the realm of the audible (there is science behind it, papers you can read) is worthy of your attention.

Measure the amount of distortion that reaches your ear, then look up the table of audible distortion, then relax because it's so small it's not a problem or re-engineer your setup.

"detail", "speed", "note decay", and "settling time"

These could be seen as different words for more or less distortion. Take settling time. It could mean the speed with which a sine wave decays if you remove the signal. Which is distortion.

but there's more subjective fuzz here, that I'd like to understand.

Why? It's subjective, no one knows what its supposed to mean.

3rd harmonics

Another word for distortion. You don't want any harmonics. Not first, not second or third. All of them distort the signal. Not a problem of course if the distortion is small enough to be inaudible.

various harmonics that lingers during playing lots of notes, leading to a false sense of perceived detail.

If anything lingers that is not signal, it is distortion. If it is audible, the product is defective.

notes pretty much smear during complex songs

Distortion of the signal.

BA Timbre can be a result of odd harmonic peaks

Harmonics are distortion. BA Timbre is distortion.

music is playing

Ahh, signal at last!

View attachment 189434

Seems this measurement tries to tell how long a distortion lasts and how far away it is from the signal. This makes sense to me. You want the time of the distortion short and the distance from the signal in dB as large as possible.
Not sure what would be audible distortion. Maybe this is already transparent?

View attachment 189435

Another measurement of distortion. In headphones you want a perfect line on top of the Harman curve. Luckily, bent FR is fixable.

A sound engineer should never have to talk about speed, color, timbre or words like that, they are undefined.
Talk about distortion (based on measurements) and transparency (which depends on our brain/ears limited resolution to detect said distortions).

 

[JanesJr1]

You could take a distortion measurement.

A speakers membrane swings never in and out as clean as the input signal, it adds distortion, no matter the technology.
If the distortion is inaudible, the device is considered transparent (the signal is indistinguishable from the original to human ears). If the distortion is large enough to be audible, the device should be considered defective and reenter the engineering cycle.

Everything else (speed, wideness, authority) is imagination. A rational sound engineer will not use those terms because they are undefined.


Because it debates belief, not fact. I'm with you, it's stupid and proof that indeed, all the people, including us, are full of biases and assumptions and plenty of BS.


I think you need to understand that distortion in the realm of the audible (there is science behind it, papers you can read) is worthy of your attention.

Measure the amount of distortion that reaches your ear, then look up the table of audible distortion, then relax because it's so small it's not a problem or re-engineer your setup.


These could be seen as different words for more or less distortion. Take settling time. It could mean the speed with which a sine wave decays if you remove the signal. Which is distortion.


Why? It's subjective, no one knows what its supposed to mean.


Another word for distortion. You don't want any harmonics. Not first, not second or third. All of them distort the signal. Not a problem of course if the distortion is small enough to be inaudible.


If anything lingers that is not signal, it is distortion. If it is audible, the product is defective.


Distortion of the signal.


Harmonics are distortion. BA Timbre is distortion.


Ahh, signal at last!


Seems this measurement tries to tell how long a distortion lasts and how far away it is from the signal. This makes sense to me. You want the time of the distortion short and the distance from the signal in dB as large as possible.
Not sure what would be audible distortion. Maybe this is already transparent?



Another measurement of distortion. In headphones you want a perfect line on top of the Harman curve. Luckily, bent FR is fixable.

A sound engineer should never have to talk about speed, color, timbre or words like that, they are undefined.
Talk about distortion (based on measurements) and transparency (which depends on our brain/ears limited resolution to detect said distortions).

I like most everything you say here. I still have some question with respect to many reviewers who (if your comments are true) seem to associate a lack of distortion with poor subjective fidelity. For example, headphones with "short decay time" are very often described as "blunted". Balanced armature IEM's with detail and a "short settling time" are "plasticy". An absence of distortion seems to be connected to poor sound way too often. I'm wondering what the reality is underneath the airy language. Is there a preference for some kind of euphonic resonances? Or is there another kind of distortion involved that is just not being measured?

On the other hand, some of the very headphones and IEM's that others find dry, blunted or plasticy, I find transparent, dynamic and delightful.. At least if I've had a good crack at truing up the EQ. Maybe that's my underlying discomfort: I hear something different from others and wonder what's going on.

 

[Earfonia]

Seems this measurement tries to tell how long a distortion lasts and how far away it is from the signal. This makes sense to me. You want the time of the distortion short and the distance from the signal in dB as large as possible.
Not sure what would be audible distortion. Maybe this is already transparent?



Another measurement of distortion. In headphones you want a perfect line on top of the Harman curve. Luckily, bent FR is fixable.

A sound engineer should never have to talk about speed, color, timbre or words like that, they are undefined.
Talk about distortion (based on measurements) and transparency (which depends on our brain/ears limited resolution to detect said distortions).

Please enlighten me how an Impulse Response measurement can be related to distortion measurement. Thanks!

I showed 3 Impulse Response measurements, the 1st one with shortest decay sounds worst to my ears than the other 2 In-Ear Monitors with longer decay. What I can see from my own measurement is the 1st IEM sounds worst than the other 2 is because its' FR deviates further than my target curve. So FR is still the most important factor for sound quality. What I'm trying to observe here is a sound quality that I cannot describe based on FR measurement alone, and measured decay from IR measurement might say something about it.

The last picture is Frequency Response measurement, not a 'measurement of distortion'. And I don't really follow why you relate distortion measurement with Harman Target? What I know so far Harman Target is only target for frequency response, not distortion.
And there is a reason why I cannot use Harman Target for my measurement and I explained it here:

Earfonia IEM Measurement Setup & Methodology

One of the main objectives of In-Ear Monitor measurement (as well as other audio equipment measurements) is to understand the correlation of the perceived sound quality with the numbers and graphs. I’ve been doing IEM measurement for a while now, just as a hobby, to learn how to correlate what I...

www.audiosciencereview.com

Scroll down to this section:

Comparison to Harman IE Target Curve - 2019​

I explained there why Harman Target is not a realistic target for area 5kHz - 10kHz.
Resolve has similar opinion when he explains stage 8 in his video here:

Btw I think this is not the place to discuss Harman Target.

Regarding how people describe sound quality using words from other senses, like warm, bright, etc, I guess because we simply don't have a rich vocabulary of words to describe sound quality. So people borrow words from other senses to express what they hear. Nothing wrong with that. Not everyone is highly educated or well versed with all the technical terms in audio. Let's help each other to define what they are trying to express about sound quality when they use certain words that are not well defined, or common to express audio quality.

Btw 'timbre' is a well defined word for sound quality:

Timbre - Wikipedia

en.wikipedia.org

What measurements number in the dashboard leads to a "warm and intimate" sound vs "colder more technical/clinical" sound?

It is true that higher sinad lead to something more separated more defined and true to the source. However, I wonder which measurement data can define the "personality" of an audio device when it comes to warm vs cold? The frequency response?

www.audiosciencereview.com

Learn the Language of Good Sound

Better understand the difference a hi-fi system can make

www.crutchfield.com

 

[JanesJr1]

I'm not sure if everyone has the same understanding and definition of 'dry' sound. My definition coming from room reverberations when wet is associated with more reverberations. And audio mixing when we use to call dry signal before we apply any reverb.

I guess theoretically the drier the more accurate. What I'm saying was, I was not used to that quality having used to other IEMs with longer decay.

And most recordings were mixed using studio monitors in room that I guess won't have decay properties as short as IEM. Just my guess. What I mean is, when I apply reverb to my recording the amount is depending on the monitoring environment during mixing. So even the shorter decay monitor is probably the more accurate monitor but the recordings that we listen might not be intended to be listened using such monitor.

I agree from a consumer (not sound engineer) perspective. I appreciate neutral, dry, even blunted.

(The exception is that I once did comparisons between Etymotic ER4XR IEM's (balanced armature drivers) and ER2XR IEM's (dynamic drivers). Same sound per FR graphs, except the ER2's had more bass extension, and subjectively sounded more resonant in the low's. On some material, I preferred the resonant lows, even though it might have been euphonic distortion. )