Hi all, on the recent Noise Floor podcast titled ‘The MOST TECHNICAL Headphones?’, there was some back and forth in the YT comments about driver story and whether drivers impart any sort of inherent quality to the sound, or whether any driver could be tuned to sound like any other. The comments also touched on drivers as a part of headphones making up an acoustic system that contributes to FR at the ear drum, and also the idea that EQ could be used to make one headphone sound like another. I posted a summary of what I thought were the fundamental points that I’m reposting here:
I think we could reasonably say:
Sound quality, including its goodness/badness and the user’s perception of technicalities, is ultimately due to FR at the eardrum and the interaction of this with psycho-acoustics.
The acoustic system of the headphone (including the driver) propagating sound at the eardrum is the cause of FR at the eardrum, and results in the ‘technicalities’ experienced as a psycho-acoustic effect.
If 2 headphones with different drivers were tuned to exactly the same FR, then putting aside other factors (including e.g. factors that may contribute to HpTF, such as acoustic impedance, and bias factors), those drivers could reasonably be expected to sound the same and result in the same experience of sound quality.
Changing the FR of a driver can change the presence of different qualities of the sound (e.g. ‘soundstage’/spaciousness effect, dynamics, detail).
It follows from points 3 and 4 that there is no inherent quality (or ‘special sauce’) to a driver based on type; there is nothing that makes any one driver inherently better than any other driver.
Despite point 5, it is possible for drivers/headphones to have particular advantages over other drivers/headphones in particular aspects in certain scenarios (e.g. some headphones are less prone to distortion when EQ’d, some headphones are more sensitive and therefore easier to power with EQ, i.e. with negative gain, some might be more likely to be easier to tune).
Also despite point 5, there are trends with driver types in terms of their acoustic parameters and as part of the acoustic system presented by typical headphones in which they are used (e.g. in my experience, open-back hifiman planars have that plucked quality that you talked about a lot in your earlier live streams).
Further to point 7, those trends are significant enough that it may be possible to identify un-EQ’d headphones with particular drivers (e.g. planars) in blind tests.
Related to point 4, there is the potential for EQ to be used to tune one set of headphones so that it sounds very similar to, or better than, another set of headphones, and also to achieve changes in the different sound qualities (e.g. as mentioned in point 4).
Despite point 9, there are limitations to EQ (and to the ways in which some drivers/headphones respond to EQ), and in practice we can’t easily replicate FR at the eardrum for different headphones on a single head, let alone achieving this from person to person.
Some acoustic properties of headphones may be desirable in terms of producing greater perceptions of good sound quality and being amenable to tuning, with and without EQ, e.g. acoustic impedance / better performance in open environments (and perhaps some such properties could be considered a necessary condition for good sound quality, or for a particular aspect of sound quality), perhaps higher sensitivity.
A high price is not a sufficient condition for good sound quality in headphones, but it (or at least a certain minimum price) may be a necessary condition.
Regarding the above points, do you agree? Am I missing anything obvious or pertinent? Is there anything that should be added?
With this as a jumping off point, I’m also interested in any thoughts on:
What do you consider are the necessary and sufficient conditions for good sound quality, or for particular aspects of good sound quality?
Regarding points 3, 4 and 9 above, would it be possible to EQ a well-known headphone to sound like another well-known headphone (e.g. maybe make a HD800S sound like a 600/650/6XX)? Could you make a lower cost headphone (e.g. Clear/Arya) sound like a flagship (Utopia/Susvara)? Or would there be physical limitations of the acoustic system of the headphones, or other limitations (e.g. with EQ), that precludes this?
Are there known unknowns in terms of the effects of particular FR features at the eardrum? For example, do we know if fined grained differences in FR have much of an effect on sound quality? Do we know why amps might sound different despite not having measurably different FR, e.g. Resolve consistently reported the Topping A90 as sounding mushy and the RebelAmp as sounding lively, punchy and engaging?
Thanks for reading, appreciate there’s a lot in this but thought I’d do a bit of a brain dump.
Howdy, MJ… A very thoughtful post that will hopefully stimulate a little more serious discussion around these topics.
I think there is some merit in trying to match the FR of different headphones with EQ. But probebly not for some of the same reasons or purposes you’re alluding to above.
The fundamental question you seem to be asking is one which has been contemplated in audiophile circles probably as long as they’ve existed… Can you turn a sow’s ear into a silk purse, just by making a few tweaks to FR?
You can get a bit more nuanced and also phrase the question this way… Is FR and distortion, measured at the DRP (ear drum reference point), all that matters in terms of sound quality?
If you believe the answer to the 2nd question is yes, then it should also be possible to turn a proverbial sow’s ear into a silk purse… I think a lot of audiophiles of the subjectivist persuasion might tell you this is not possible though… if only to justify the large sums of cash they’ve spent on their gear over the years.
I come down somewhere in the middle on these questions. There are limits to human perception. And I think if you could match FR and distortion at the DRP, or get reasonably close to this in perceptual terms, then I think you could get pretty close to the same sound quality. Maybe even close enough to fool the average listener or audiophile in blind tests.
I think there can be quite a few practical impediments though to really achieving something like that. Some of which you’ve already discussed. So I’m not quite ready to get on board with the sow’s ear analogy.
Bottom line up front: I see a lot of armchair speculation and overgeneralization, but little evidence of extended critical listening and careful comparisons.
Underspecified. Do you mean FR at a static point in time or FR in the real world, with drivers that all have distinct resonant properties and decay timelines?
That is tautological. Yep. The headphones “acoustic system” is responsible for the technicalities of the headphones.
Impossible. Different drivers have different sizes, weights, materials, compositions, etc. As above, a point-in-time FR doesn’t equate to identical real-world musical time sequences. See the ASR crowd…do they seriously try to make their Porta Pros sound like Utopias?!?
Now, this may not apply to brands that seem to sell maybe 3 families of drivers and tune them very slightly differently for different users and price points (e.g., Focal, Hifiman, Sennheiser). In those cases one may indeed come very close to duplicating the performance of another product. And if a maker consciously tries to clone another product then “different drivers” could be tuned to be functionally the same.
Yes. That’s why EQ has value.
NO NO NO NO. BS BS BS BS. NON SEQUITUR. This is ASR-like armchair speculation. Listen with your ears. First, FR in #3 may look “fine” with a static measurement tool, but “exact” matching is impossible over a time sequence. Second, dynamic drivers routinely take on the resonant character of the materials of their composition. Paper cones sound “papery” – soft, dull, rolled-off treble, plastic cones sound “plasticky” - hollow, boomy, low-mid emphasis, and metal cones sound “metallic” – bright, precise, reflective. Planar and electrostatic drivers tend to sound diffuse and dithered versus the single-point movement of a dynamic cone.
I swap you my Koss Porta Pro for your Focal Utopia. Even trade.
So, you just admitted that #5 is BS.
Again, you admit #5 is BS and replicate my comments about driver construction and composition.
It is rather easy to identify headphones with the right testing method. I, and many people, tend to fail rapid ABX testing between setups. But, listening to a 10 to 30 second clips one after another does not allow psychoacoustic habituation to occur and does not resemble how music is actually consumed. As such, I created a marathon playlist to address the only property that cannot be ignored, denied, or avoided: fatigue.
Some music sources are recorded poorly or have baked-in noise so they make my ears ring or give me a headache. The key for testing is that the impact changes with the setup (DAC, amp, headphones), and some headphones pass my marathon test while others fail in seconds, or after 30 minutes, or after an hour, or after 2.75 hours. Each set of drivers has little tiny unique quirks and responds differently to some notes and transitions – these key points can be learned whereby any changes with a new set of headphones stands out.
This method cannot be blind, as pain and ringing ears are inescapable. The time point and type of pain becomes a fingerprint for the driver. If you use this playlist or a similar one, test with a clean DAC and a solid state amp. Some differences will be obscured by colored setups, tube amps, etc.
That’s uncontroversial. EQ can shift the coarse properties of headphones in any given direction. EQ can’t make a woody avocado from Peru taste like a sublime creamy avocado from Mexico (true story: check the labels). EQ can’t make Budweiser taste like Budvar. My offer of giving you my Porta Pro in exchange for a Utopia stands.
And so, what was the point of this armchair exercise? Coarse tweaks work well. Yes. Nuanced tweaks and “free upgrades” to flagship products fail. Yes.
As in my prior response to you, this generally means tighter tolerances and better quality control. There is no way to make a cheap mass-produced driver be as consistent as one built with care from the ground up, and built with higher quality materials. The basic HD 600 is an excellent product. I am routinely shocked at how good it sounds when I go back after a break. Still, it doesn’t have the range, blackness, foreground/background distinction of my Clear. My Clear doesn’t have the range or nuances of my HD 800 S. While I can’t handle the Utopia’s treble spikes, it greatly outclasses other headphones in 3D space and localization. No other Focal product can touch it or be EQed to come close.
There’s all sorts of luxury branding and image-driven price mark-ups with audio, watches, jewelry, Lexus vs. Toyota cars, and more. I downright hate some flagship headphones (e.g., everything with closed cup), and don’t much like others (e.g., Susvara). The minimum price depends on your interest and ability to hear nuances. For many people this could be the HD 600 or HD 650, but the HD 5 series probably won’t be good enough. Still, I’m quite happy for upgrading beyond this tier. It’s a hobby, and the more expensive headphones perform substantially better with my marathon tests.
See above. Sometimes yes, sometimes NO NO NO NO NO. Overall I think you are slipping into the ASR “equal measurements mean its just as good” error. No, listen with your ears. I own two amps that fall near the ends of ASR’s amp measurement ranking list (the last time I checked). Thse include the “terrible” Bottlehead Crack and the “excellent” THX AAA 789. I still use the fuzzy Crack but boxed up the 789 (and would sell it if there was much of a market for them). The 789 can sound like listening to sandpaper on broken glass while sliding across ice. But its measurement numbers are great.
Personal testing. Nothing else matters. Your hearing and that pleasant zing you feel from “cool looking” equipment is all that matters. About 90% of the technicalities can be reached with the HD 600. If you are value oriented, stop there.
You can downgrade sound easily, but you cannot upgrade sound quality a great degree with EQ. That’s wishful thinking. About 1/3rd of the tracks on my playlist are heavily distorted and degraded – this is to show how different headphones perform with less-than-ideal sources. Some tracks can be indistinguishable with cheap vs. expensive headphones, or they result in very different whines, whistles, missing bass, or artifacts.
This is perception science and it’s a well researched domain. I think you are again getting at sustained FR over time, not static FR charts. The A90 vs. RebelAmp comparison follows from materials and construction – the RebelAmp is hugely overbuilt (I own one) and can deliver plenty of power/current without breaking a sweat. A weak amp, say from a tablet or phone, typically struggles to deliver current to demanding headphones and sounds terrible. See the many discussions of Sennheiser “scaling” and “synergy.”
Just as an initial point, I want to make clear that I’m coming at this from a place of interest in learning more, with an open mind and without rusted on views about these matters. I don’t have a whole lot of experience with acoustics generally, other than having listened to music my whole life, getting into headphones a few years ago, and watching the Headphones Show and now the Noise Floor. I have listened to a limited but sufficient number of headphones to have a basic understanding of what physical and acoustic differences there may be (for full context here they are in order of purchase: AKG K371, HD560S, Sundara 2020, 6XX, OG Ananda, Elegia, OG Clear, Dunu Titan S, Arya Stealth, Meze Alba, HD800S, Radiance). But my understanding as expressed in my post above was a synthesis of points made by me as well as Resolve, Listener and another user (Lowrider007007), where I was attempting to capture the gist of what was said. The discussion there provides useful context which you may wish to read.
Here though, a lot of the language in your post frankly reads as an attack, and I am not sure what I have said to offend you. It was disappointing to receive this as the second response to my post, particularly when it says ‘Founding Member’ next to your name. I joined this forum maybe a month ago following my interest in the show and in getting better sound from my headphones and music. If your goal was to make the new guy feel alienated and unwelcome, well done, you have succeeded.
In my response I will try to focus on your substantive points and attempt to ignore your various accusations (armchair speculation, lack of evidence, overgeneralisation, underspecification, tautology, BS, wrongness, non-sequitur, admittance of BS (perhaps hypocrisy?), and pointlessness - can you understand how this comes across?).
Now, going to your substantive points.
I’m not sure what you mean by this. I mean sound quality (including technicalities) perceived by a user when listening to headphones. So I suppose FR in the real world. Regarding FR at a static point in time thing, my understanding is that as headphones are minimum phase devices, a measurement of headphones’ FR at a static point in time should be same as in the time domain, so in theory any distinct resonant properties and decay timelines should be reflected in the FR plot. Is that not right? I confess I have only recently come to understand what it means for headphones to be minimum phase, and there seems to be some contradictory information about the significance of this.
The context for this statement was in a discussion about driver story, or in other words drivers having inherent properties that are reflected in sound quality. So the ‘acoustic system of the headphones’ is shorthand for recognising that all parts of the headphones are potentially capable of contributing to FR at the eardrum, not limited to the driver but including the housing, baffle, earpads, etc.
I agree that in practice this is probably impossible. But the point addresses a hypothetical: if 2 headphones with different drivers produced the same FR at the eardrum, they would sound the same. I think that proposition is correct. Maybe this is what you meant when you said armchair speculation. FWIW, I am speculating from an office chair.
I could say my experience overlaps with this to a limited extent. For instance, I believe the 6XX has paper drivers, and it is definitely the darkest out of all my headphones. My Focal Clears do sound reasonably bright and precise (though my planars also sound precise), though I’m not sure what you mean by reflective. I’ve heard plenty of reviewers say they sound metallic, but personally I don’t hear it, or at least that’s not a word I would use to describe what I do hear. My planars haven’t necessarily sounded more diffuse or dithered than my dynamic driver headphones, though I would say my Ananda probably sounds softest (or roundest?), and the Clear is the most punchy.
However, I don’t think it is the case that all drivers in all headphones of each of those materials or kinds takes on those qualities. For instance, I understand the Meze 109 uses paper drivers but is quite bright. If the driver / headphones are tuned differently, they will sound different. But perhaps it is also the case that drivers of particular materials (e.g. paper) or kinds (e.g. planar vs moving coil) are more likely to have particular sound qualities in their default tuning once put in a headphone housing, but they can be tuned to have a different response.
I have a Clear - one of my queries on the EQ page was whether it would be possible to EQ this to sound like the Utopia. But it was a question about what is possible, not an assertion that it actually is possible. I am interested in conducting the experiment to see what the result is. I am not proceeding from an assumption that I can get flagship experience at a cheaper price.
The proposition in point 5 is that driver types do not inherently correlate to different sound quality / qualities. I had thought that was clear from the context but perhaps I could have stated it expressly. The qualifier in point 6 is ‘in particular aspects in certain scenarios’, and the examples given are about practical matters that don’t relate to FR. So the 2 statements are not contradictory (and I am certainly not ‘admitting point 5 is BS’ - though it is clear enough that you think it is).
I got a bit lost in what you were saying in this response. Are you saying headphones with drivers using a particular kind of materials (e.g. paper) or kind of driver (e.g. planar or moving coil) will reliably give you pain and tinnitus as a result of your marathon playlist? I am not sure what your point was, if you would like to clarify. The point I was making in points 7 and 8 above were about the default tunings or perceived sound qualities of certain kinds of headphones when listened to without EQ (e.g. planars).
In 2 out of 2 responses I’ve received to my original post, non-audio analogies have been used. The essence of my question is: can you use EQ to make one headphone sound like another? Intuitively my impression is that it might be possible to achieve a similar coarse grain response to another headphone, but in every case there would likely be small variations in FR that still make the headphones distinguishable and not the same sounding. But I am not sure if that is right, and I am interested in testing this (or indeed if it has already been tested).
The point of this exercise is for me to learn more and to confirm the correctness of my understanding of the above points. Thanks for your contribution.
As I mentioned above, I am not pushing the above points as my own unassailable position. In actuality, my subjective experience has been that I have perceived differences in sources that have affected sound quality (namely, differences between my Fiio K5 Pro, Chord Mojo 2, Singxer SA-1 and Xduoo TA-26S), with the poorly measuring TA-26S sounding fantastic for my high impedance dynamics and the more expensive, well measuring and balanced Singxer sounding softer and more boring to my ear in stock form than my older, single-ended only K5 Pro. (I’ve since done the jumper mod on the Singxer and it fixed the softness, so I like it better now).
My experience does not accord with this. I mentioned this in Listener’s EQ thread - my HD800S was a recent purchase, and without EQ I thought it sounded frankly pretty bad. Cold, boring, unengaging but also fatiguing. Because of those things, I couldn’t focus on the detail that everyone praises it for. But I applied the EQ that Listener posted for it (the very first one in the thread), and to me it is a completely different headphone. It is rich, punchy, lively, dynamic, engaging, very much a joy to use now. I would say that is an upgrade to the sound quality with EQ. But I wonder if your intention was more about EQ’ing cheaper headphones to sound like more expensive ones.
Are you saying that the A90 cannot deliver sufficient power / current, and this is why it sounds bad? I mean, there could be something in that, I’m not sure.
You may not understand the audio culture or backstory, but are (perhaps innocently) repeating many cliches and well-trod discussions. I mentioned ASR specifically because that website led many people to make poor buying decisions, and it caused business harm to vendors who were not interested in measurements and an arbitrary “objective” method (e.g., see the demand crash of Audio GD and the release of the Schiit Heresy amp following ASR’s reviews). In following a rather militant founder, there were many wannabe dragon slayers several years back who said X is just as good as Y when there’s no evidence that they listened the more expensive Y, or they trusted a leader who said he evaluated high-end amps and DACs with a basic HD 650 playing at “earlobe shaking” volume.
I don’t think @Resolve would make some of the arguments you listed, but I don’t know enough about @listen_r or Lowrider007007 to say one way or another.
I’ve got to call it as I see it and stand by those assessments. Consider the possibility that they are factual despite being critical. BUT, you wrote a quite radical position with bullet #5–so radical that I almost wrote you off and stopped answering. While your general point is fine (EQ has an impact), the goals of slaying dragons in higher price tiers and your apparent reliance on EQ measurements is misguided.
Compare until you pick up on “slow” versus “fast” drivers. Slow ones will be smooth and gentle but risk being fuzzy or muddy, while fast ones have clear transitions with details and risk being edgy or harsh.
It’s not a remotely possible hypothetical. It is not worth your time or further discussion. This speculation can be easily disproven with a couple back-to-back demos. EQ to your hearts content and then actually listen for an hour or two. See above regarding perceived attacks.
The 6XX has plastic drivers. It also has boosted mid bass (forced bass versus the thinner 600) that makes many sources sound artificially thick while it takes away from the clarity and details of vocals and the upper mids. While quite popular, the 650/6XX is quite colored and makes a lot of music sound overly similar.
Here is the exercise: tap your finger on a cardboard box (paper drivers), a plastic box (plastic drivers) and a metal box (metal drivers). Listen to the sounds. Metal reflects sounds more precisely than either plastic or paper. Headphone drivers cannot avoid their physical construction materials.
Planars emit sound from all over the driver. Dynamic drivers have a cone that moves as a whole. Softest/roundest is indeed the diffuse character of the driver. I find dynamic drivers to be well unified across time, whereby all the sound comes out and arrives together → this is punchy.
No need to speculate here, you say “perhaps it is also the case” – it is the case that dynamic drivers typically take on the resonant tone and timbre of their construction material. This may be harder to hear with some products, or digital processing may obscure it.
What are we talking about? It is NOT possible to deeply transform performance. Full stop. The drivers are made from different materials, have different stiffnesses, different thicknesses, etc. You can make the tone similar, but the Clear’s clipped low and high range plus hazy grain will persist relative to the razor-clear deeper and more precise Utopia.
My prior comments stand.
One can learn to hear subtle differences by listening to one set of headphones for a long period. My 3.5 hour list is just the start and bare minimum. You can pick up on much more if you listen to the same set every day for a month, then switch to another set. Training happens. As such, your hypotheticals are not grounded in reality. Fatigue is a reliable indicator that something different is happening with Setup A vs. Setup B.
The answer is a firm NO NO NO you cannot make one set match another if you have close to average/normal hearing. You can make the tone profile similar with EQ, but not match the speed, lower and upper range, harmonic resonances, and other qualities of the driver.
The context was about making X sound like Y. When you buy X you can do anything you want to improve it, so your HD 800 S story is not relevant. I too EQ the 800 S to get the best from it.
No specific comments, as I don’t own one. Class A vs. AB vs. D amps sound different to me.
Could likely be sighted biases or poor level matching. I want to stress that I did not do a blind ABX there. Could also be something to do with the distortion profile as Cameron talks about. My position on amps/DACs is that when I did do the blind testing, even in scenarios where I thought I may have heard a difference, at no point did I ever care. So I just leave that to the people who do care. The exception here would be to do with certain tube amps that I think are fun, that’s when the differences are far more noticeable.
No I would make the majority of those arguments, as would Listener. I hear you with respect to the ASR-like notions about EQ, and I would definitely agree that there are those there who regularly lead people astray because they have what amounts to a poor understanding of both audio metrology and acoustic systems but that kind of thinking really is not what the original poster here is indicating.
You also said some things about test tones, and I’m not sure I’m entirely clear on what you mean by “with a static measurement tool”. You can measure the FR with a sweep, with an FFT (real time with pink noise), or even music. It’s all the same, with the exception of certain active headphones that have dynamic or loudness based equalization where they change based on the crest factor of the music.
ALL of which are subjective qualities that… if they have a genuine acoustic cause (as in not just placebo), can be found in FR at the eardrum. Like as the original poster indicated, it’s not that the drivers here don’t matter, they contribute meaningfully to the FR at the eardrum. But we shouldn’t consider an extra dimension of ‘technicalities’ that one driver imparts over another if the FR at the eardrum is the same. The problem of course is that in practice, no two headphones will have the same FR at the eardrum.
As far as other acoustic properties that contribute to ‘special sauce’… we can’t say for sure there’s NOTHING else or that it’s all just FR at the eardrum (even if it’s likely). Axel Grell for example is testing the effects of acoustic impedance for precisely this reason. My understanding is that he feels there’s a unique effect that we can perceive with high acoustic Z vs low acoustic Z even if FR at the eardrum is controlled for, which is very interesting. But that’s less to do with the driver than it is to do with the system as a whole.
@MichaelJames with regards to your post, I think you alluded to this but the big issue here is that without proper tools, you’ll never be able to assess the goodness and badness of a given FR at your eardrum. So EQ of this nature is extremely sophisticated and annoyingly difficult to actually investigate. And beyond that, who is to say that people don’t have a preference for a given coloration. So even if you get their HRTF and can measure at their eardrum, you don’t know what’s actually contributing to the psychoacoustic qualities a person enjoys.
A general comment: You describe the frequency response at the eardrum as the primary or overwhelmingly important parameter affecting sound quality of headphones.
I disagree. The ear/brain system is exquisitely sensitive especially to the time domain phase and transient response of the reproduced audio signal - this affects perceived directionality, 3-dimensional depth, and the degree of so-called holographic 3D perceptions of soundstage. This high sensitivity probably evolved because it was necessary for accurate perception of distance and direction of threats in the environment. Time domain smearing and other such factors subjectively cause a considerable blurring, smearing and sometimes overbrightness of instrumental sonic images.
With the headphone system as a whole some of this time domain response is not due to the headphone itself but also to the amplifier and the preceding DAC outputs having things like transient IM distortion due to excessive feedback. The key point is that many of these mechanisms of time domain distortion result in what could be termed time smearing of transients and ringing, which the ear/brain system is very sensitive to.
However, much of the time domain response is indeed caused by the headphones, where there can be mechanical resonances at certain frequencies, causing ringing at the resonance frequencies with the sharpness and amplitude of the ringing determining the “Q”, which factor is also one of the important parameters in each section of the parametric EQ systems often used with headphones.
As can be seen, a basic understanding of perceived sound quality of a headphone is a complicated matter where time domain response is at least as important as frequency response. The time domain behavior of the system is critical because the ear/brain system is so very sensitive to it.
No, time domain information isn’t relevant in headphones as a separate thing from FR. Unless something is very wrong/broken, headphones are minimum phase systems, which means time domain is proportional to the FR. You can show if there are reasons to care about time domain by plotting excess group delay. This gets tricky at ultra high frequencies, but essentially that would be the ‘smoking gun’ as to whether or not you need to care about CSD or waterfall plots, and in general if you see issues there it may be a disqualifier for the product because it would be behaving… poorly.
Then what about the adverse sonic effects of skin effect in thicker conductors? This can occur in the headphone cables, and in the interconnects used in a distributed system with separate DAC and amplifier. I know from experience that that can cause a considerable blurring and smearing of the sonic “image”, and is due to the frequency-dependent differential phase shifts of the signal currents propagating near or at the surface of the conductor and time delayed reduced amplitude components propagating at various depths inside the conductor. This phenomenon is proportional to conductor diameter, makes it important to use as thin conductor wires as possible, either Litz type construction or very thin ribbons that minimize max depth of penetration, and is easily demonstratable by comparative evaluation of different cable designs.
I don’t think this phenomenon could be calculated and predicted from the frequency response.
Yes, though I want to be clear that HpTF is literally “FR as measured on x head”. I feel like what you meant is “contribute to non-HRTF related HpTF variation” but I just wanted to make that clear
Yes.
Yes, at least when it comes to sound quality itself. Even if some drivers are better suited for certain tasks—drivers with low free-air resonance and low acoustic impedance are generally better for things like wireless devices, closed back headphones, or EQ—this does not mean they will necessarily sound better in practice.
Yes.
Yes, and these features are typically found in the FR of the headphone because the FR of the headphone is a result of the driver’s motion.
Yes, though perhaps some results might surprise people. Listeners generally aren’t as discerning as they think they are, and that goes for me as well
Yes.
Limitations to EQ, not so much. We can fairly easily transform the entirety of a device’s FR with EQ (barring any issues with frequency extension, non-minimum phase behavior or other nonlinear behaviors like distortion, but these would be headphone limitations not EQ limitations).
Limitations to measurements, however, indeed there are many limitations to measurements, and measurements would likely be necessary to put such a thing into practice in earnest.
I’ll say yes, though my evidence here is limited admittedly. Lower acoustic impedance does tend to result in positive attributions of spaciousness (in my experience, at least) but I would say there are plenty of cases wherein devices with higher acoustic impedance—IEMs, for example—are often rated better than headphones for one subjective quality or another. It’s something I’d like to test for sure.
Definitely, as many very inexpensive headphones do not always have basics like good channel balance or frequency extension handled well.
As said, channel balance and reasonable (if not perfect) frequency extension is what I’d call “necessary” but not sufficient. As far as nailing down the “sufficient” stuff, people seem to disagree pretty vehemently re: what that entails. For me its just lack of noticeable or distracting coloration across the FR, but for some people I’m sure the opposite is true and they want every area to have distinct colorations that make it interesting. Such is audio.
IMO yes, it can be done. I’ve been able to get fairly close even without using in-ear mics on my own head, but friends of mine have been able to do it near-perfectly by using blocked-canal microphones in their own ears and EQing both headphones to the same response.
Absolutely yes.
We still don’t know the difference between how most headphones measure on human heads (post-DF compensation) and the measurement fixtures we use, for starters. I’m sure you’d agree this matters a ton, as its humans listening to headphones and not measurement rigs.
We still don’t know how much narrow band resonances contribute to preference in headphones; in Harman’s work we know what people preferred relative to an HD800 or a Sennheiser HD 518 measured on a rig that either no one uses except a few enthusiasts (GRAS with KB006x pinnae), or that no one uses and that lacks a DF HRTF (Welti Pinna 45CA).
We don’t know the responses people actually were getting on their heads so we don’t know what FR is actually preferred by humans outside of their preferred response as measured with those headphones on their rigs, so there very well may have been narrow band resonances affecting what people preferred and we wouldn’t know.
I’m not an authority on this but as someone who has heard differences between amplifiers in sighted tests, bought amplifiers and DACs based on that experience, and then ended up selling them all: I’m pretty positive I would fail at distinguishing between them in blind tests. The differences, if they do exist (and I doubt they do, personally) have always been small enough to essentially never be worth the money they require spending unless you’re spending like < $500 for the whole system.
All of the effects of those factors you mention would be contained in the FR, so changing the FR would make the sound at your eardrum different in whatever way you determine. That said, headphones do have actual differences in weight, size, feel, etc. that mean even if they produced identical frequency responses at your ear, they would have sufficient physical differences that may cause a listener to perceive the sound differently.
Additionally, if they’re two different acoustic designs and you EQ them the same on a measurement fixture, this does not guarantee they will still measure the same when brought to another measurement fixture (or to your own head). Indeed the evidence seems to support them measuring differently most times, and that variation would be a factor specific to each headphone’s acoustic design.
First off, I really find the tone of this message in multiple places condescending and frankly quite rude. Please speak with a bit less hostility and assume the best of other participants in conversation going forward. We are all here to learn.
FR is, itself, a time sequence. Nothing contained in the behavior of the pattern, speed, and magnitude of the headphone’s oscillation of air is not present in the frequency response.
I think this is generalizing a fair bit, though not unfairly. The thing is, these materials/driver types also have common FR features among them that would result in the sonic characteristics you’re talking about, and you even mention a few of them, so I’m not sure how what you’re saying is incompatible with what they’ve said.
I unironically would take that trade if they were priced the same; PortaPro is a much more livable sound signature for me
No, they didn’t, they’re just seeing the nuance that some driver’s acoustic properties do inherently make them better fits for certain tasks, but not overall.
Large diaphragm planars like the Hifiman stuff have beeg drivers with low tension and this often results in a few good things for certain tasks like EQ—eg. low frequency extension, low positional variation. These things do not necessarily mean inherently better for all cases/reasons/sounds though, eg. these same devices tend to be quite bright, have comparatively high distortion in the audible band (though this likely won’t be audible with music) and have varying levels of modal behavior that can may contribute to some weirdness in the audible band as well.
It’s possible to admit that there are characteristics inherent to/typical of driver types while also acknowledging that these characteristics contribute to the FR of the device, which can be changed with EQ. These truths are not mutually exclusive, and indeed it points to MichaelJames having a good handle on how headphones actually work.
I mean, they’re right though. It’s fairly easy to EQ an inexpensive headphone—say a Sennheiser HD 560S—to sound better than a Utopia or Susvara, but that’s because frankly the latter don’t sound all that perfect anyway.
Utopia might have “dynamics” but its also incredibly small and boxy sounding and has narrow band treble resonances while otherwise being littered with dips across its FR. The Susvara is incredibly bright on my head.
Either of these things, or any expensive headphone I’ve tried, could be supplanted by a cheap headphone (provided it has the requisite frequency extension) with EQ, IME.
Here we agree Though thinking about technicalities as a percentage doesn’t really make much sense to me, since there’s basically no such thing as 100% without EQ (and IME, once you’ve done that you kinda end up acknowledging that the technicalities were always something that came as a result of the FR you heard, not a result of the driver or headphone’s inherent qualities separate from FR).
I’m not gonna get too much deeper into the disagreements here, but I do want to say that I believe @MichaelJames has a great handle on the current state of understanding of headphones, and I appreciate him starting the discussion.
So if you’re genuinely hearing differences from cables in properly controlled blind tests… that means you’re dealing with devices that are extremely low impedance where the impedance curve is nonlinear, and then that’s essentially the same as EQ (because it’s changing the FR).
If you’re genuinely hearing differences from interconnects… something could be very wrong with one of them. I imagine there’s a scenario where you get a difference in overall SPL, and that may lead you to believe you’re hearing a meaningful difference, but this is where proper controls are likely to reveal there’s nothing there.
Cursed take. Mind you, I also think it might be more challenging to EQ an on-ear headphone to match the DRP FR of a Utopia given pinna crush. Though with the right tools, could be doable.
I am far too busy eating my freedom turkey to acknowledge this incredibly rude behavior, but you’ll receive a strongly worded letter in 3-7 business weeks.
I just wanted to add a couple of things to this discussion. Here’s the live stream being referenced:
And one thing I want to push back on more ardently is Blaine’s position that these ‘technicalities’, even as construed in more appropriately subjective terms, is just down to how much a listener is prompted to pay attention to certain aspects of the experience. So one reason might be… because they paid a lot for the product and therefore are more likely to listen harder, thus genuinely hearing more of the details or whatever.
There’s definitely some truth to this, insofar as your attention really is the guiding factor in determining what’s compelling, and your attention to the experience can certainly be amplified by non-acoustic aspects of these products: how much they cost, how they look and feel, the driver story being expressed and so on.
But I don’t want the takeaway here to be that there aren’t differences in perceived technicalities caused by genuine acoustic differences as well. These are subjective psychoacoustic effects, and the reason we can be confident in them being impacted by very real acoustic changes is that these effects do change when you EQ the same product.
Like we’d probably say it’s not prudent to EQ the fine-grained features in the treble past a certain point unless you know what you’re doing and are doing so based on the in-situ response on your own head. Or in other words… do so subjectively, not to the graph on a HATS, because of this.
But for those who have gone down this rabbit hole, it typically reveals a lot about those psychoacoustic effects, and how they’re impacted. It shows that it is possible to take a generally cheap and poor performing headphone and inject some ‘technicalities’ into it with the right adjustments. It’s just that doing so can be extremely difficult and time consuming.
Achieving “good sound” is one thing, but trying to emulate the exact in-situ FR of a different headphone that you find has better ‘technicalities’ is nearly impossible, and certainly impractical. At that point you’d also likely need to have that more preferred headphone on hand anyway.
So to me this isn’t really a matter of “turning a sow’s ear into a silk purse”, as much as it’s an indicator of how we should be treating FR as also describing ‘technicalities’, not just tonality. And we really should stop thinking of this as some separate magical property of driver stories. Their acoustic performance is something that can be measured, and the previous gaps in correlation between what we hear and what we measure is down to the fact that the in-situ response at our own eardrums is different in various respects from what the graph shows.
At the very least, that last part is extremely likely to be true, and if so, it seems bizarre to me that we’d continue with the mystery of “but there’s more to my experience than what the graph shows” in terms of unmeasurable, unknown quantities. Let’s look at the in-situ FR at the eardrums of real people relative to their HRTFs and then go from there.
Just curious if the livestream above is the video you were planning on posting on the subject of technicalities.
I haven’t had a chance to listen to the stream yet. But was sort of waitin on your new vid before commenting in a little more detail on some aspects of this subject.
Maybe @listen_r or someone else can explain a little more about what IME is as well, in between bites of freedom turkey. Because I’m not exactly sure what that acronym refers to in this context.
