Is there any way to objectively measure headphone resolution?

[solderdude]

Since they could have a resonant character, smoothing a response by compensating dips might cause additional THD+N issues due to new resonances happening, doing it the passive way by bringing down could indeed regulate masking - yet it could also contribute to worse resolution, if the big peak cause distortion masking - it seems that we'd require a complex model to get down to even the basics. Don't trust me on that, though. It is just merely interesting, considering i.e. what was brought up by your K612 measurements w/ being coarse despite being EQ'd down. I believe that's what I encountered on HD-681s EQ'd to Harman (for pure experiments) and it completely put me off that it's so harsh, what I've never heard before w/o EQ. Could you try boosting the dips w/ K612 (or anything else you have on hand with non-smooth treble) programatically after having the filter for the peaks applied? Targeting a smooth treble character on measurement. Just to see whether getting the treble objectively "smoother" (either with boosting or downboosting) gets us anywhere further into the better SQ and if the impact is negligible or not.

IMO it is best to leave sharp (narrow) dips alone. Wider dips can be 'filled-in' though. Also those >6kHz measured with HATS as one may be EQing rig artifacts and f'ing things up instead of improving things.
Peaks on the other hand can be EQ'ed down. It won't held against resonances but the biggest culprit is the amplitude error. That can be addressed succesfully.
The same here is valid. Above 5-7 kHz you may be looking at peaks that appear to be lower or higher. Even the most expensive HATS are afflicted but maybe a tad less or different from others. It varies from headphone to headphone, HATS to HATS, Ear to Ear.
So when one wants to EQ look for common 'peaks' from various measurements and EQ only the 'average'. Chances are this is most succesful.

 

[bobbooo]

The ER4SR can be 0.85% still sound clean from no masking effects since it FR is so smooth

For an IEM, not particularly. In fact, the $10 Sony MH750 (to the ER4SR's $280) is around the same in that respect if not smoother, with better high-frequency extension (one of the prerequisites for good 'resolution'). Etymotics are over-hyped, and hugely over-priced. I suspect the main reasons people find them 'detailed' are due to the inadequate bass which by contrast accentuates the mids/treble, and the high isolation which reduces masking of fine detail in music from outside noise, but at those exorbitant prices you might as well go for custom-molded IEMs to get even better isolation, and far better comfort.

 

[Feelas]

IMO it is best to leave sharp (narrow) dips alone. Wider dips can be 'filled-in' though. Also those >6kHz measured with HATS as one may be EQing rig artifacts and f'ing things up instead of improving things.
Peaks on the other hand can be EQ'ed down. It won't held against resonances but the biggest culprit is the amplitude error. That can be addressed succesfully.
The same here is valid. Above 5-7 kHz you may be looking at peaks that appear to be lower or higher. Even the most expensive HATS are afflicted but maybe a tad less or different from others. It varies from headphone to headphone, HATS to HATS, Ear to Ear.
So when one wants to EQ look for common 'peaks' from various measurements and EQ only the 'average'. Chances are this is most succesful.

I get it, so better not overcompensate, since it's more than possible that the HATS itself effs things up. Got it.

I wonder whether the averaged pinna causes the drama (and that could be done with using nonstandard pinnae) or if the HATS material is just wrong. Isn't it possible that we're overestimating the impact and variance?

 

[Dreyfus]

Good observation, but what do you mean exactly by "saturation and sharpness"?

The issue with strong peaks of whatever Q is that they may be so prominant that they overshoot the mean loudness of their band. As a consequence you may have to turn down the volume which makes the peak more bearable but further attenuates the neighbouring frequencies which finally leads to a loss of detail. To make it short, it's just a matter of missing balance in the higher frequency response.

It would need to be proven that the difference between couplers (or even heads, in a large-scale test) is measurably high as to be regarded as important in measuring treble response.

The deviations induced by pinnas and ear canals of different shapes and sizes have been sufficiently investigated in the past fifty years.

Hammershøi and Møller, Sound transmission to and within the human ear canal (1996):



1st graph: free field response at the ear entrance point = transfer function of the pinna
2nd graph: free field response at the drum reference point = transfer function of pinna + ear canal
3rd graph: response at the drum with the source at the ear entrance = transfer function of the ear canal

We can say that such resonances for sure DO matter when we are talking about accurate sound reproduction.
However, that does not mean that an average approximation like the curve gathered by Harman does not serve a purpose. In fact, headphones voiced towards an average in-room preference are still way better than an arbitrary curve that has been colored by an individual taste and ideology. "Accuracy" is still another thing, though.

If we're really speaking about completely unaccurate response in >4-5kHz range then I'd say that it's absolutely pointless to bother discussing in that case, since then we'd be completely in the dark if using actual measurements. Yet, I believe on GRAS we're measuring the summarical (so we can omit predicting the actual resonances happening, since we're measuring them in anyways) response and it's discussable whether this should even get mentioned.
...
It seems that Harman found out that until 8-10kHz the difference is consistent enough to take it into account and after that there's no point EQ-ing.

The older 60318-4 ear simulators are rated with a smooth, consistent response until 10 kHz. Everything above is quite uncertain due to the high-Q tube resonance which is specified at 13.5 kHz (+/- 1.5 kHz). The resonance was measured in a closed system with the mic right at the entrance of the coupler. Now, when we attach an open ear canal to that (which is the standard for the newer measurement fixtures) the simulator turns into an open tube resonator with much wider bandwith and one or two distinct spikes at certain frequencies. The position of those is prescribed by the shape and length of the tube.



The resultant problem with that is that the measured response can get really messy when the self-resonances of the ear simulator are overlapping the resonances of the headphone driver and cavity as much as the interferences induced by the silicone pinna. Without further insight, it is very hard to isolate and judge the actual performance of the headphone because you only receive the total sum of all the components interfering with each other.

Some people may say that this totally fine since those fixtures are derived from average models and thus do actually reflect the average listening experience in total. And indeed, as long as you don't look too close and apply a certain amount of smoothing (the Harman response is heavily smoothed!) such measurements can actually be quite usable. However, keep in mind that YOUR ears are not the ones that were measured in the chain! Your ear canal length, diameter and bending might and actually will differ from that of a reference called by GRAS or B&K, so will the size and shape of the pinna.
I can only point to the research done by Hammershoi & Moller, David Griesinger and Paul Barton that shows how complex and diverse both the human anatomy as much as the auditory perception actually are.

To come back to the later point, which is the Harman research, I don't know if they already had the newer high-res pinna and coupler (ear simulator, to be exact) when they did their latest investigations. Essentially, the pinna got an anthropometric ear canal with the first and second bend (averaged from a lot of 3D scans). Moreover, they damped the self-resonance of the 60318-4 coupler which now offers much smoother responses with less inteference issues.

Still, whatever those components spit out as a final plot, there is still the final question about the anatomy of the individual listener in the "non-average" practice. The improvements described above might decrease tolerances and improve the repeatablity of the measurements. And that's a good thing without any doubt. But they tell only little about the exact sound reproduction at the listener's ear drum which would be essential when we wanted to analyze the level of detail a headphone could produce. After all, it is a very individual thing. That's why generalized measurement systems hit their limits sooner or later.

you could also do something like what Oluv's Gadgets does - get really good recordings of the output of different headphones and then A/B and try and pick the "higher detail-retrieval" phones...

As much as I appreciate Oluv's efforts in HiFi, his procedure to track and simulate the coloration of a headphone with a miniDSP EARS is totally flawed since ...
A) we have no data that confirm that the impedance of the whole system is anywhere near that of a real human ear, let alone an average
b) the rig is technically not even an ear simulator, just an electret microphone (with more or less unkown calibration) stuck into a rather hard silicon pinna
c) you are recording a transducer with a transducer, then playing back the whole thing with just another transducer and expect the outcome to be anywhere near the original. Every component in the row adds its own artifacts and distortion.

Regards
Dreyfus

PS: Sorry for the wall of text. But I think you really need the context to understand why such systems have certain practical limits and should not be taken as incontestable standards.

 

[solderdude]

"IMO it is best to 'start' with a headphone that needs the least amount of EQ and take it from there"

In your experience, which headphones need the least EQ?

K812 (needs treble peak removed)
K371 (has a small peak which can be removed)
HE6 (old one)
LCD2 (needs a lot of EQ but reacts well)
DT880 (needs treble peak removed)
DT1990 (needs treble peak removed)
SRH1840 (needs some EQ, shouldn't be played loud)
Ananda (needs treble peak removed)
Sundara (new pads)
Focal Clear
Mr Speakers AEON closed, Ether.
Fidelio X2HR (bass and treble EQ needed)
HD600 (lacks bass extension)
HD650 (needs some warmth removed)
HD560S (needs upper mids lowered)
HD58X, HD660S
Verum1
Oppo PM3

 

[Sgt. Ear Ache]

As much as I appreciate Oluv's efforts in HiFi, his procedure to track and simulate the coloration of a headphone with a miniDSP EARS is totally flawed since ...
A) we have no data that confirm that the impedance of the whole system is anywhere near that of a real human ear, let alone an average
b) the rig is technically not even an ear simulator, just an electret microphone (with more or less unkown calibration) stuck into a rather hard silicon pinna
c) you are recording a transducer with a transducer, then playing back the whole thing with just another transducer and expect the outcome to be anywhere near the original. Every component in the row adds its own artifacts and distortion.

Regards
Dreyfus

PS: Sorry for the wall of text. But I think you really need the context to understand why such systems have certain practical limits and should not be taken as incontestable standards.

I get all that. I used Oluv as an example only but of course, it would be an imperfect option. But it would be a relative comparison of 2 things, not an authoritative evaluation of any one thing. Lots of us have pretty darn good headphone setups, and comparing one recording to another and choosing which one seemed the most "detailed and revealing" if there is any meaningful difference might be doable. Any artifacts and distortion (again given that many of us have some pretty clean headphone rigs) would be added to both recordings equally...

(I'm also not really on board with the whole "every ear is different and we need to account for that" idea. Afaic, a lifetime of tuning by literally every earthly sound being fed through our individual adaptive hearing systems effectively eliminates them as a factor we need to consider. We all hear exactly the same sounds through our ears our entire lives. We use those same ears when we listen to full size speakers too...)

 

[Feelas]

The issue with strong peaks of whatever Q is that they may be so prominant that they overshoot the mean loudness of their band. As a consequence you may have to turn down the volume which makes the peak more bearable but further attenuates the neighbouring frequencies which finally leads to a loss of detail. To make it short, it's just a matter of missing balance in the higher frequency response. [...]
I can only point to the research done by Hammershoi & Moller, David Griesinger and Paul Barton who give us an insight into how complex and diverse both the human anatomy as much as the auditory perception actually are. [...]

Right, that's actually a simpler idea than I thought, and MUCH more problematic than you'd think first.

As for the couplers, at least for me I think that it quite cleans up the stuff (even by getting it messier), especially in terms of highlighting how... BIG the differences are. What were the measurements taken with?
All in all, great! If you have some time on your hands (if not, that's fine too), could you point to which exact ones are especially worth exploring?

As much as I appreciate Oluv's efforts in HiFi, his procedure to track and simulate the coloration of a headphone with a miniDSP EARS is totally flawed since ...

As Ache mentions - would doing a correlation of subjective experiences with measurements hurt? It would be completely intransferable, yet it'd serve a role of rounding up where to look with correct gear. Yet, seeing all that's written above... We're risking obscuring the real information by some unknown factors of miniDSP.

It seems that providing a personal eq kit would be the only solution to get the stuff correct.

 

[Dreyfus]

I'm also not really on board with the whole "every ear is different and we need to account for that" idea. Afaic, a lifetime of tuning by literally ever earthly sound being fed through our individual adaptive hearing systems effectively eliminates them as a factor we need to consider. We all hear exactly the same sounds through our ears our entire lives. We use those same ears when we listen to full size speakers too...

There is the opportunity to adapt, indeed. Just as you learn to hear and orientate in the world when growing up as a child. But when you take that as a justification, you could also say that no one would need something like an average listening curve (Harman).

In my experience a personalized head- and ear-related response is crucial for a natural hearing reproduction. Especially when there is no visual reference. I would expect the same for an accurate evaluation of detail.

 

[Sgt. Ear Ache]

There is the opportunity to adapt, indeed. Just as you learn to hear and orientate in the world when growing up as a child. But when you take that as a justification, you could also say that no one would need something like an average listening curve.

In my experience a personalized head- and ear-related response is crucial for a natural hearing reproduction. Especially when there is no visual reference. I would expect the same for an accurate evaluation of detail.

Why wouldn't we need an average listening curve? Our hearing is adaptive...but it doesn't adapt overnight. It adapts over years. By the time we are seriously looking for audio equipment we've mostly had 20 or more years of adapting to sound. And since we all hear every worldly sound relative to every other worldly sound, and since we all have roughly similar (thought not exactly the same) ears, we largely come to a similar average preference of sound curve...

 

[Feelas]

Why wouldn't we need an average listening curve? Our hearing is adaptive...but it doesn't adapt overnight. It adapts over years. By the time we are seriously looking for audio equipment we've mostly had 20 or more years of adapting to sound. And since we all hear every worldly sound relative to every other worldly sound, and since we all have roughly similar (thought not exactly the same) ears, we largely come to a similar average preference of sound curve...

Well, I guess Dreyfus'es irony just dodged you right there Adaptiveness is the sole reason we need the curve, since the brain is compensating the ears & body. Despite that, there's some miniscule adapting ability towards specific models over time, but that all being overshadowed by years and years, just as mentioned. Don't forget that torso also affects whats is perceived.

 

[Dreyfus]

Our hearing is adaptive...but it doesn't adapt overnight.

Personally I do only need about 15 minutes until I forget about the tonality and take it as more or less balanced. That counts for about +/- 3 dB around my average listening curve I would say. Stronger colorations do need longer listening periods, indeed.

Doesn't work so well with (non-personalized) binaural recordings, however. I think that is because the HRTF coherence is just too weak, at least on my head. That's were the specific high-frequency response gets very important.


Do also not forget that headphones - without DSP - do excite only a fraction of the HRTF our hearing expects. That's why why our brain locates the sound source just inside of the head. The spectral maps and phase informations arriving at our ear drums simply do not match our natural experience (see the "gestalt" model by Günther Theile). And without measurements that show the actual delta between a specific headphone model and a natural external sound source at the ear drum of the individual listener we have not much of a clue to fix that.
If we look at it this way, the broad sounding of a headphone (how much bass vs treble emphasis etc.) could be considered to be one of the least problems we have to address. IMO it is much more important to design headphones which have better potential to adapt to our natural spatial hearing ... because that's where our hearing is definately NOT able to compensate.

As for the sole resolution, I would expect a response that has been matched to the individual loudness curve - with all the dips and peaks fixed as good as practically possible - to be the most pleasing. I wouldn't say that you could reach that with any random headphone, though. You would need a good tonal consistency (mostly indipendent of the wearing position) and a model that somewhat matches your personal ear resonance fingerprint.

What were the measurements taken with?

Which measurements are we talking about?

Hammershoi and Moller?
https://vbn.aau.dk/ws/portalfiles/portal/227875888/1996_Hammersh_i_and_M_ller_JASA.pdf

As Ache mentions - would doing a correlation of subjective experiences with measurements hurt?

You have to make sure that the correlation keeps consistent. That can be hard to achieve without a somewhat realistic anatomy.

Regards

 

[Blake Klondike]

Here's the best answer I've seen to this question, from reddit user Oratory1990, a professional acoustic engineer who measures headphones for a living:

So basically, resolution/detail, like the vast majority of a headphone's perceived characteristics that influence judged sound quality, can be determined by frequency response, as is supported by the scientific literature e.g. Dr Sean Olive's extensive research.

Can anyone recommend which models Amir has tested have the best results in this category? Very interested to check this out!

 

[chi2]

@ Dreyfus The wide inter-individual variation in ear canal and pinna physique leading to widely varying transfer functions is the main hindrance in objective headphone measurement. Interestingly, people with pronounced differences in transfer functions and therefore objectively differing sound signals at the eardrum level perceive what they hear, e.g. a live performance of a musical instrument, to be completely natural. It is a mix of brain functioning and lack of alternate experiences that sets the standard for the qualitative evaluation of what is heard.

Given the fact that those ear canal and pinna transfer functions have too much uncontrolled variation between individuals, one possible conclusion is to measure headphones, not earphones, not on the eardrum level with measurement rigs that only by chance fit a given individual transfer function, but instead on a damped flat plane. Such a simple measurement rig may yield more objective information, as this way all individually varying characteristics of the pinna and ear canal are left unconsidered, that is, they still apply when actually listening to the headphone.

Many headphone geeks, me included , have built such simple rigs using calibrated electret microphones, placed in wooden boxes filled with sand, and covered with cellular rubber. The possibilities of comparing the sonic characteristics of different headphones on such a rig are imo more consistent than with the minidsp EARS that I also own but never got warm with, despite the application of different compensations, among others one based on relative differences to a state of the art measurement rig provided by Marvin on SBAF.

A different and far more sophisticated approach - not with the aim of generalized assessment of the quality of a headphone but of consideration of individual transfer functions - is to do individualized measurements with microphones placed as deep as possible in the ear canal. This way, the individual transfer function can be assessed and a correction function can be calculated and applied to the sound signal.

The Smyth Realizer A8 and A16 are able to do that and the result in reproducing sound perception in a given room over headphones is astounding. When a headphone has competent technical abilities including the lack of severe resonances, the Realizer is able to neutralize many of the existing frequency-related shortcomings of the headphone, hereby reducing the sonic differences between them. Yet, the resulting musical reproduction over different headphones, while being more similar, is still far from being identical.

Some weeks ago, I had done an individual measurement/calibration with the A16 in a sound studio for a 9-1-4 Dolby Atmos system. At the same event, I had four different headphones calibrated to my ears. This way I was able to compare the reproduction of the music in the room with the A16's simulation over the different headphones. The headphones (Sennheiser HD800, Focal Utopia, Abyss 1266 Phi TC, Hedd Heddphone) were sonically closer together than without compensation. Still, there were substantial differences with the HD800 giving the most precise and convincing result - possibly due to its excellent driver properties and spatial capabilities. The remaining differences between the headphones may also come from the fact that the microphones, while being placed deeply in the ear canal, still were far from the eardrum level. Therefore, any possible interactions between resonances of the different headphones and ear canal resonances were left unconsidered.

 

[Feelas]

Given the fact that those ear canal and pinna transfer functions have too much uncontrolled variation between individuals, one possible conclusion is to measure headphones, not earphones, not on the eardrum level with measurement rigs that only by chance fit a given individual transfer function, but instead on a damped flat plane. Such a simple measurement rig may yield more objective information, as this way all individually varying characteristics of the pinna and ear canal are left unconsidered, that is, they still apply when actually listening to the headphone.

I don't feel that'd necessarily be more objective, but surely abstracted from the "noise" that the variance causes. It's not necessarily more subjective to measure on some fixture with it's own characteristics. Yet, if we can't assert that head-based phenomena don't happen in a pair (and not either add to or substract from the experience), could lead us nowhere near the real causes. I think it'd be much better to just use some non-standard (averaged to typical variation) rigs and present a family of curves, since we can't get away from the variation?

As for the Realiser experiments - can you somehow describe the perceived variation between pairs? Even in vague terms.

Off-topic a bit - does the Realizer really get that close to speaker soundstage reproduction as can be read in places, from your perception? I was wondering if solely matching the headphones to personal HRIR ends up with a proper soundstage reproduction, and am dying to discuss this, since I didn't have a chance to do it myself.

 

[Dreyfus]

... one possible conclusion is to measure headphones, not earphones, not on the eardrum level with measurement rigs that only by chance fit a given individual transfer function, but instead on a damped flat plane. Such a simple measurement rig may yield more objective information, as this way all individually varying characteristics of the pinna and ear canal are left unconsidered, that is, they still apply when actually listening to the headphone.

I get your point.

The problem I see with the flat plate, however, is that it renders a completely different working environment which drastically changes the impedance of the resonanting volume. Hence, you will get a treble response which introduces other constructive and desctructive errors in correlation with the boundary layer. Without the exact (or at least rough interpretation of the) anatomy of the listener in place, it is really hard to interpolate what is really going on there since you are completly skipping the PRTF interaction which can change substantially between various headphone designs. The pinna activation is not a linear function and depends on the design of the headphone ... and the exact shape of the ear, of course. That's why the interpolation of the treble to the expected real life performance can be really tricky on a flat plate fixture.

On the other hand, bass and mids will be mostly fine and to my experience actually quite consistent. You just have to keep in mind that a plate is much better at providing a proper seal. So in practice, most people will probably not preceive the same bass level unless the design is really good at sealing around the human ear (see Beyer as an example, their pad design keeps the bass quite consistent on different head shapes).

At the end of the day it is indeed an alternative to measuring with ear simulators and such. But not neccessarily a better one. It strongly depends on the demands the fixture shall fulfill and - once again - how the user handles the data. As for a precise high frequency measurement, I still see no alternative to measuring with the listeners ear anatomy. Unfortunately, this is hard to achieve technically and not really usefull as a common reference in the community. That's why we consider ourselves mostly satisfied with the average responses spit out by ear and cheek simulators. We just have to keep in mind that it is a smoothed average! Some people appear to ignore that quite consistently.

Still, there were substantial differences with the HD800 giving the most precise and convincing result - possibly due to its excellent driver properties and spatial capabilities. The remaining differences between the headphones may also come from the fact that the microphones, while being placed deeply in the ear canal, still were far from the eardrum level. Therefore, any possible interactions between resonances of the different headphones and ear canal resonances were left unconsidered.

The HD800 is often used as a reference for surround simulations because of its large ear cup design which appears to offer a great pinna activation and tonal consistency when measuring with in-ear mics.

I've done a few experiments with Jaakko Pasanen's Impulcifer project and measured my HRTF in front of my studio monitors. I used a DIY in-ear mic which fits into the entrance of the ear canal. For playback I choose my HD 700 which was calibrated to flat in the same procedure. There are some errors due to the lack of ear canal resonance which leaves room for improvements, indeed. Still, the result is quite impressive already. I haven't tried the Realizer universe, yet. But so far I can say that personalized surround simulations are miles ahead of generic HRIRs and BRIRs.

By the way, Hammershoi and Moller did also comparisons between blocked and open ear canal measurements versus the results right at the ear drum.

In short, measurements at the ear canal entrance seem to capture all the directional dependencies introduced by the pinna. The canals add some slight variations on the position and level of the high-Q peaks. But its hard to say if that is already an evidence for a practically relevant directional dependency of the canal. As for the open vs closed canal entrance, it is probably better to measure with the blocked canal. The blocked measurements are more consistent and thus allow better repeatability. Both are still quite different from the response at the ear drum, though. So you will loose some data either way. According to David Griesinger, it is the timbre matching our personal HRTF that is being lost.

Regards
Dreyfus

 

[chi2]

The problem I see with the flat plate, however, is that it renders a completely different working environment which drastically changes the impedance of the resonanting volume. Hence, you will get a treble response which introduces other constructive and desctructive errors in correlation with the boundary layer. Without the exact (or at least rough interpretation of the) anatomy of the listener in place, it is really hard to interpolate what is really going on there since you are completly skipping the PRTF interaction which can change substantially between various headphone designs. The pinna activation is not a linear function and depends on the design of the headphone ... and the exact shape of the ear, of course. That's why the interpolation of the treble to the expected real life performance can be really tricky on a flat plate fixture.

Valid points regarding the aim of objective and generalizable measurements, thanks for the insight.

The HD800 is often used as a reference for surround simulations because of its large ear cup design which appears to offer a great pinna activation and tonal consistency when measuring with in-ear mics.

As for the Realiser experiments - can you somehow describe the perceived variation between pairs? Even in vague terms.

The HD800 to my knowledge is still one of the phones with the largest stage. The A16 manages to neutralize its tonal weaknesses (too light in the low and lowest freqiencies, peaking mid highs). The resulting quality in all frequency ranges is impressive. I expected the planar Abyss Phi TC to excel in the bass but I actually preferred the sound of the compensated HD800. The Utopia was very clean and precise sounding, yet the smaller stage had a negative impact on the authenticity of the simulated room. Same with the Heddphone which, on top of that, had remaining weaknesses in the frequency range below 100Hz. The ribbon driver seems to have inherent limitations, that can't be compensated by means of EQ.

I've done a few experiments with Jaakko Pasanen's Impulcifer project and measured my HRTF in front of my studio monitors. I used a DIY in-ear mic which fits into the entrance of the ear canal. For playback I choose my HD 700 which was calibrated to flat in the same procedure. There are some errors due to the lack of ear canal resonance which leaves room for improvements, indeed. Still, the result is quite impressive already. I haven't tried the Realizer universe, yet. But so far I can say that personalized surround simulations are miles ahead of generic HRIRs and BRIRs.

Impressive to see such a project in the public domain. I wonder how long ot takes till this kind of dsp is easily available without expert knowledge and usable in portable setups.

 

[100rounddrum]

High resolution = low distortion and balanced, extended FR response. Though the resolution can be masked if there’s an obvious boost or dip somewhere in the FR range, despite having low distortion.

Oh, and Harman curve isn’t anywhere clse to being perceptually neutral. Lower-mids are recessed, upper-mids too forward and bass has too much boost.

 

[Feelas]

Oh, and Harman curve isn’t anywhere clse to being perceptually neutral. Lower-mids are recessed, upper-mids too forward and bass has too much boost.

I guess same can be said about mostly any target curve if you account for what Dreyfus compiled for us. It's closer than both FF and DF targets, I think. I'd just not agree about too big bass boost, since getting the character in-line with room curves is actually a nice treat. The headphone crowd is just too accustomed to thin sound.

IMO I haven't yet found albums which translate badly onto K371s, and that means (I guess) that I can be hardly taken as a non-average person

 

[daftcombo]

The concept of "resolution" has been imported from the world of image & video and makes little sense in audio.

 

[dominikz]

Well, if we can go from impulse response to FR and back-forth, then (as far as I remember, tho I'm no major in the topic) both FR and IR contain complete system's response to the stimulus.

The FR/IR will predict the system's complete response to a stimulus only if the system in question is linear and time-invariant ('LTI' system). If there are non-linearities (which exist in most real systems to some extent), those are unfortunately not described by the IR/FR. Examples of non-linear distortions are e.g. those which we attempt to quantize with THD and IMD figures. These can be described by a system's transfer function / curve (here's a nice visual example of that).

Of course, one of the aims of high fidelity audio engineering is to create audio system as close to LTI as possible, and then also minimize linear distortions