I really like this idea, and more broadly, the idea to have headphone manufacturers on the podcast to discuss things like design philosophy, inspiration, etc.
3 Likes
My audio best of the year would be live music. I like source even better than source gear. There was a great Toto concert I saw in the Spring.
Not sure if this was already covered but a discussion on different driver types would be cool. Not necessarily how they work, but more on how they sound.
One of the biggest reasons I struggle to believe “FR at the eardrum is everything” or something to that effect is how different planars and dynamics sound to me, even when FR for both headphones are harman-like. In fact, planars have such a distinct sound, dynamics as well, that I have basically defined a “planar sound” and “dynamic sound” in my mind, and I would like to know if you guys feel the same way, and how the science can maybe explain it. Same with estats sounding “ethereal” to many listeners, is that just a lack of bass or is there more to that. If I EQ bass down, would I get an estat like sound?
Even more so e-stats.
Two headphones measuring “harman-like” is not a sufficient condition for them to sound the same, especially when most of the “driver timbre” stuff comes from the response over 5 kHz—and this is exactly where we’d expect to see the responses of headphones diverge as they are placed on a different head.
This is important because matching them on one rig doesn’t mean they’ll match on another rig… or on your own head.
Here I’ve EQed two headphones measured on a clone 43AG with KB006x rig—similar to what Harman used—to Harman’s 2018 target.
Yes, they look similar, and they may even sound somewhat similar, but it is a near-certainty that they wouldn’t sound identical, or even “impossible to distinguish.” That’s because once you bring those same EQed headphones to a different head, you see the differences between them don’t track, and there’s a good deal more divergence:
Showing the differences between the two by compensating one to the other:
KB006x Clone Rig
5128
h
This is owed to what we call “non-HRTF related HpTF variation” or simply, the effect the measurement head itself has on the behavior of the headphone. Resolve and I did a video that really gets deep into this and I highly recommend checking it out for a even better explanation. The short answer is that the heads themselves make the headphone behave differently when it is placed on them, so we can’t take a single measurement of a headphone at face value as “the way this headphone performs,” nor can we assume that the trends shown on one head will track to another head.
But this is why “Harman-like headphones don’t sound similar” is not a sufficient argument against FR at the eardrum characterizing the totality of the sonic event for human listeners. Harman-like (or Harman EQed) headphones as measured on one rig are likely not measuring nearly as similarly when placed on another head, or your head, and the differences tend to arise exactly where we’d also get things that contribute to “planar timbre” in the ≥ 5 kHz region.
2 Likes
There’s a saying (I don’t remember where I got this from):
Speakers play your room,
Headphones play your ears
As it turns out, your ears also play the headphones.
2 Likes
Yes, that explains the difference between a specific dynamic and specific planar, but not necessarily the trends of driver sounds, such as “planar timbre” or “ethereal estats”. I’m wondering if there are specific FR traits that can help explain such trends. As an example, many headphones widely considered to have a large soundstage have a 2k dip. I’ve yet to encounter a similar correlation to explain, for lack of a better description, “driver sound” (in my experience, LCD-4, LCD-5, susvara, he1000stealth all share a planar sound, despite measuring quite differently).
1 Like
planar timbre
I’d say its most likely large magnitude resonances above 10 kHz, all of the headphones you mentioned have this (LCD-5 to the least degree, but it still has some).
estat ethereality
Lack of bass and forward 1 kHz contribute to a weightless presentation that is more typical of estats than planars, additionally estats are often very acoustically open which may contribute to this.
Certainly not all estats lack bass. My ancient STAX SR-5N lack base and have the forward 1Khz you mention. The NectarSound HIVE does not. It has very solid bass. I understand that STAX SR-007, in particular the later versions have good bass. And for the money the flagship STAX should have it also.
Aside from uncertainty of FR at ear drum, there’s also these factors to consider:
- We don’t have a rigorous way to define subjective experiences, from widely accepted ones such as timbre to more abstract ones such as ethereality.
Let’s take timbre for example. If you have experience with live instruments, you probably have a pretty good idea of how they should sound like. However, not all pianos, guitars, violins, etc. sound the same. Not all recording microphones sound the same, and not all sound engineers master the same way. The tricky thing is that, in the grand scheme of things, FR at ear drum (which let’s just agree is all that matters for now), only plays a portion in the entire chain of production. Dynamic compression (in the proper audio engineering meaning), mixing, non-linear added effects, and even the distance between the microphone and the instrument all contribute to the final timbre. Vocal is probably what most people base on when discerning timbre. Here’s the thing, if you look at what kind of EQ sound engineers put on vocals and microphones, suddenly all those 2 or 3 dB differences betweens FR become rounding errors. What’s the realistic timbre then? Is it just what sounds best?
Metallic timbre vs Plastic timbre vs Organic timbre. Aside from the HD 600 series, which is the only one remotely closed to titled DF-HRTF, the way the headphones communities assigned timbre to headphones is quite fascinating to me. Metallic for Focal, which incidentally has metal drivers; Plastic for planars, which happen to have a thin moving polymer diaphragm; Organic for ZMF, which somehow has bio-cellulose drivers. That’s just a bit too convenient, isn’t it? All I’m sayng is, there’s a pretty good chance psychoacoustics affects the very real auditory signals. On the other hand, from my experience in characteristic vibrational response of materials, I understand the general trends of timbre from different materials, but that definition is probably far from what most audiophiles would describe timbre quality in headphones. Sidenote, most instruments are made of metal. I’d hope they sound metallic.
What about ethereality? My interpretation is that the most ethereal sound I can think of is from a pipe organ. Looking at it’s timbre from FFT, we can see the wide bandwidth of each harmonic and the steep downward slope of overtones. Contrasting with a violin, which features sharp overtones and a shallow slope extending much further to the upper treble. This quality is what I’d describe as clarity and imaging precision. How does it relate to headphones FR? I doubt there’s a strong literal correlation from a single note, since headphones are minimum-phase and have very low distortion. However, if you consider the entire timbre over time of a chord progression, the overtones getting louder or quieter because of FR can have similar auditory effect to instrument timbre. That’s my version of ethereality. I’m sure it’s different from yours, but yours is probably different from another listener’s as well.
- Confirmation biases: we know what the material of the drivers before we even listen to the headphones; we talk to eachother and converge to the same definition, even though our experiences might be vastly different.
The only way to find out is to conduct a blind test with different driver materials and a number of truly random FRs with the same chassis and weight. There’s no way to ensure identical FR, so doing many random ones that don’t trend towards anything. For example, leakage behaves differently for different drivers, so just ensuring that the bass level is evenly random between drivers. Back to reality, no one’s gonna do this. There’s literally zero or even negative financial incentive.
I probably have counter-biases where I couldn’t care less about the driver type maybe to a fault. I thought the 3 BA UE Reference Remastered was a punchy boy. It reminded of LCD-X, which rendered kick drums in a really satisfying way. The most impressive slam I experienced was between the Utopia, then HE 1 and Cosmo. I think HD 650 with bass-slope EQ is also quite fantastic in that area. But quite frankly, the difference between them is less than the difference between 2 tracks in the same album. So, make of that what you will. I often use phrases like ‘most dynamic-like planar’ or ‘most planar-like IEM’ to the point where they lost their meanings altogether. I’m so desensitised to driver story that the only reason I care is how heavy does it make the headphones or how large does it make the IEMs.
- Physical psychoacoustics: the coupling between the pads and head contributes a lot to my experience personally.
The Focal Clear microfibre pads with its high clamp force gives me quite a visceral sensation. However, it’s ear gain features don’t give me quite the distinct impact that I’d describe particularly hard hitting. Nevertheless, that sensation is what I’d never get from the mechanical coupling of the HD 800 S, even if I EQ everything to the same FR at ear drum (hypothetically of course since I can’t EQ that granularly, ever.) So, it seems to me that the physical design of headphones makes a larger impact to subjective experience than the driver type itself.
Which pipe organ? They differ quite a bit. And having heard Frank Lloyd Wright’s post horns - they ain’t ethereal. What about a glass harmonica? Or the celeste or carillon?
What about the Greek Hydraulis? My favorite, but more circusy than ethereal.
I don’t think the difference matters. My point was that ethereality gave me a vibe of grandeur, so I figured an instrument the size of a room was quite fitting. An organ has so many parameters that you can create all sorts of timbre in a single instrument, which makes the difference between different organs kind of a moot point. They all have the same FFT features which are wider bandwidths and steep slope compared to a typical string instrument because of the sheer size and diffusion. As for the one that I heard, it’s in Esplanade, Singapore.
So this is the organ you heard
That’s a very modern instrument and in a theater organ tradition. If you get to listen to some others and some old ones you might find differences.
Given your parameters a steam calliope would also sound ethereal. I can’t define ethereal, but like the judge said about porn, I know it when I see (hear) it. It’s what I associate with Neptune from Holst’s Planets.
I’ll throw some topics into the ring. Thanks for the show - it’s been both entertaining and informative!
What gear have you changed your mind on over time and why, e.g. something that you’ve fallen out of love with, something that you’ve come to appreciate more with time?
How has your perspective shifted on what things matter to you in audio as you’ve learned and experienced more in the hobby?
What do you think were the big trends in personal audio in 2024, both in the products that were released, and in the discourse in the community, and how have you felt about them?
The two-car garage is a popular topic of discussion among car people, where the premise is that only two cars have to satisfy all of a person’s automotive wants, and needs. Now suppose you had to whittle down your personal audio collection to two transducers (you’re not getting a loophole where you get to keep all your IEMs while you have only two headphones.) Which two would you choose?
2 Likes
None of us has unlimited funds. I got into this hobby, and after a few cheapie single DD IEM’s from KZ and the 7hz Zero 2, I bought an ARTTI T10, simply to discover what this planar magnetic technology might sound like. And the T10 has blown my mind, simply far more resolving than any other single DD IEM I have heard.
I am not one who enjoys purchasing products to “sample” them, and return them. I think this is ethically wrong, and an abuse of the laws guiding retail, which permit us to return items, for legitimate reasons. So if I buy something, I prefer to have researched it thoroughly, before I buy.
Where do I go next, from something like the ARTTI T10, if I wish to hear more resolution, sense of 3D, realism, detail, and listener enthrallment?
IEM’s vs Headphones, which is a better approach to invest in, to get the best audio quality, as well as value for money.
Should I stick with planar magnetics, or some other driver combination may sound even better. I definitely cannot afford any of the mainstream electrostatics, so let’s not go there. Will not bother to look at electrostatics. But outside of electrostatics what’s the upgrade path for someone like me with an ARTTI T10?
Is there any value in looking out for impulse response plots and waterfall plots when studying headphone reviews. I feel that the audiophile community is missing out on these measurements that go a long way to sort out the men from the boys, in the speaker world.
Having owned the Zero 2 with it’s highly appraised low distortion, according to AmirM of AudioScienceReview, and been underwhelmed with its sonics, which are easily superceded by the ARTTI T10, in my opinion, what measurements will really tell me and show me, without listening to the device, a good measure of what I can expect.
I buy DAC’s and AMPs and Speakers purely based on test results I see in magazines or on the web, and these are a good indication of what to expect, but with IEM’s and headphones, there does not seem to be an authoritative single set of tests, that we can review, which corresponds, closely enough, to the experience of listening to and IEM or headphone.
Many other products have these measures, that can be relied upon to describe how good they should be, but with headphones and IEMs we still seem to be in the dark ages, with all manner of opinions, that have not drilled down to exactly what makes an IEM or Headphone sound great.
The Zero 2 measures OK, but these measurements mean nothing, in real life, or are insufficient to tell me, without listening, if it is a great IEM or NOT.
This vacuum of a proper set of tests that give me an accurate expectation, is one that the manufacturers are taking full advantage of, even as we “speak”.
When a we going to have tests, that we can use, to decide what to buy, in IEMs and Headphones, with confidence, in a similar manner to what one can do today, when buying speakers.
1 Like
You’re suggesting a great topic. It’s slightly beaten to death by the Headphones.com guys over the last few years in various videos, but if there’s anything new on the topic, it’s always a good one to re-visit. It is their job after all to inform the buying public and I’m sure they wish they could continue to improve and provide relevant results.
From your well thought out post, I’d suggest that there is simply no better way than to go to Canjam or Tokyo to test out headphones and IEMs. Another way that no one really talks about (and are way, way too dismissive about) is learning how to EQ in a way that eliminates peaks and valleys for each individual headphone/iem on your own head (using Oratory’s presets and modifying is not good enough). It makes a world of difference.
We have talked about what we think the solution could be, but as Blaine’s talk indicates… we’re not there yet.
The proposed solution is something along the lines of showing the FR as a field, including the range of non-HRTF related HpTF variation from different heads/ears, relative to a preference bounds.
The other improvement we’re working on right now is finding ways to show that it’s not an ‘anything goes’ situation within the preference bounds. While you can currently use measurements fairly easily to rule out products that’ll sound terrible, it’s much harder to know from the stuff that’s generally good, which will be most good. At this point it becomes far more about probability, and that needs to be reflected in the bounds.
Excellent thoughts. Thanks.
I typically, at this time have three methods, to EQing Headphones/IEMs, cos I do not have any measurement gear of mine for these. I’m comfortable with measuring speakers with a calibrated measurement microphone, and REW(Room Equalisation Wizard). and deriving EQ corrections for these.
For Headphones/IEMs, unfortunately I have to rely on whatever measurements have been done in databases such as squid.links. Which are also replicated in the database used by AutoEQ.app.
Method 1: I pick one of the measurements in AutoEQ.app’s database, for my Headphone/IEM, and rely on AutoEQ to generate the correction. I do not usually make any changes to AutoEQ’s default settings. I may generate to different targets in addition to the AutoEQ default target, to give me alternatives. I have also found a custom target, kindly provided by one of the users on AudioScienceReview - StaticV3, to be an exceptional alternative to AutoEQ’s default target. He provides this target as a text file, which I import into AutoEQ.
I repeat this process for as many of the different measurements for my listening device, as reside in AutoEQ’s database. leaving me with a set of candidates which I may then rotate through as profiles, when I am listening, and after a while pick one. Sometimes every few days I change to use another profile. The corrections, generally sound pretty close to each other anyway, so I do not sweat over this.
Truth be told, in my opinion EQ of headphones, is an optimisation, and nothing more. We do not listen to white or pink noise, so the impact that EQ has on actual music and audio, is NOT as significant as the frequency response curves tell us. EQ will only apply equivalently, if the music we listen to has identical level in all frequencies, which is certainly not the case.
With the ARTT T10, I could well ignore any EQ attempts, and the stock sound is still very impressive, EQ is just the icing on the cake.
Method 2 : Roll of the highs and lows, to reverse the V-Shape, and boost the mids, These roll offs are gentle, no more than 3 dB max, at the extremes.
Method 3 : Introduce a “room like” speaker enhancement with boosted lows and attenuated highs, like a room curve, really just an EQ tilt with manually constructed high and low shelf parametric filters.
In practice I combine (Method 1 OR Method 2) + Method 3, i.e ending up with two layers of EQ.
I would like to hear more about this other method of EQ which you describe, would this be based on some kind of test tones at different frequencies. I’m open to trying most things, as long as they are safe!. Looking forward to your further enlightenment on the approach.
You beat me to it. Only a few hours ago, the same thought occured to me, to attend a CANJAM and hear as many of these listening devices for myself, and bring along one or two of the IEM’s/Headphones I am familiar with, as well as a portable audio player (or stream via mobile phone), with my best DAC dongle, and hear what all the fuss is about, with regard to many other Headphones/IEM’s that I read about.
Hitherto the only way I have been able to audition Headphones/IEMs is to buy them at my own cost, and experience them for myself. Obviously, that can become a bit expensive, as I reach for better quality, and demand more from my listening devices.
I come from the professional audio side of things, where I mix music for live events, and also mix albums/master them in a studio. In that world, I am very comfortable with buying products “blind”, based on a review of credible specifications. And what I hear always corresponds to the specs (of course that could be placebo!!). My point being, in that world, for things like DACs/ADCs, Microphone preamps, microphones, speakers, the measurements give me pretty much everything I need to know. Caveat - I also listen to speakers in showrooms and at events, but do so less and less, cos over time, I have developed a familiarity with the house sound of the various main manufacturers’ speakers., then I can always use custom EQ to optimise further. But by and large the detailed info on driver and cabinet, crossovers, etc, gives me a good ballpark of what to expect, and if it’s not too expensive, I can buy with confidence, without having to 1st audition. Also the reviews of these products in magazines and independent online reviews by users, complements the measurements and tends to converge.
Headphones and IEM’s have been the exception, where listening for myself, has been compulsory.
I think though, that looking back, I had already been through an early phase with professional audio equipment, where I had a learning phase, listening to some equipment, and correlating what my ears told me, with what I had read in measurements. And thereafter, I had references, and I could compare specs of other devices with my references that I owned or knew very well, from personal experience.
Having owned some IEM’s like the CCA CRA, Zero 2, and ARTTI T10, which have been extensively reviewed online, looks like I can now use these as references to appreciate the delta between them and any other listening devices, whose reviews I study., and can more reliably extrapolate what to expect.
Where I do not yet have a reference that I own, would be hybrids, or anything with BA’s and electrostatics, but guess it’s a journey, we take it one step after another, and do our best to keep the FOMO in check. No hurry.
The two key things I would like to clear up in my learning are :
-
How good are planar magnetic Over Ears - ideally Open back, in the sub $500 category, if I had to choose one or the other, do I go for a Sennheiser HD600 (or 650 or 660s) or one of the HiFiMans (HE 400SE, Sundara, Edition XS, Ananda Nano - or some other planar OE Open Back, in this price category)
-
How much better would it be, to invest in something like a Letshouer S12 2024, or Hidizs MP145 or Timeless 2, compared with my T10, which already just wows me every single time I listen, with the level of detail and effortless portrayal of information that I have never heard on any other device(including on speakers), of music I have been listening to for decades, e.g. Tracy Chapman. I’m thinking where do the diminishing returns start, and spending more is more of a sidegrade rather than an upgrade. My priority is purely sonics, and resolution, not looks. The T10 is supposedly ugly to some, but I would not notice, cos looks are NOT a criteria for me. Transparency and Comfort and Reliability/Manufacturing Quality - such as driver/FR/level consistency, between left and right earpieces, that’s all I care about.
Yeah attending a CANJAM and listening to all these for myself is the ideal solution. But t would be great to hear what others here think.