Harman preference curve for headphones - am I the only one that doesn't like this curve?

[Robbo99999]

He's a bit of a mouth-breather, isn't he?

I'm probably the same.

Yeah, it's a bit worrying I have to listen to my headphones with my mouth open or it won't sound right!

 

[Feelas]

Yeah, it's a bit worrying I have to listen to my headphones with my mouth open or it won't sound right!

You just have to get something so amazing that'll make your jaw drop

 

[Robbo99999]

You just have to get something so amazing that'll make your jaw drop

Ha, yes! But on a serious note I've found my last headphone, AKG K702 (EQ'd to Harman), which also happens to be the first proper headphone I bought. My other headphones don't touch it for soundstage, and only the HD600 matches it for detail & nuance. I think I'd have to go HD800s to better my experience, and I'm not paying that kind of money for a headphone.

 

[Robin L]

But... pop music is produced in a way that'll have most presence in the mid/upper bass region, simply because most people don't own anything that goes under 40-50Hz or that is all unhearable (e.g. in cars)...

Lots and lots of people are wearing headgear to work. My previous job was for the IRS in Fresno. It was seasonal work. We processed forms and typed letters, 8 hour shifts with frequent overtime. The favored headgear was Beats. This was 2010-2019. I got a major-league education in pop music of the 2000's, asking people what they were listening to, researching headphones at sites like Inner Fidelity and others, getting portable gear I could use at work. I asked most people what they listen to. Naturally it was all over the map. I was drawn to modern pop music after I got into bass heavy headphones. While it's true that a lot of dance/pop stuff, the sort of stuff found in "Now" collections, focuses on "thump, more than a few pop productions go down real deep. And I found a lot of that music was enjoyable solely for the bass lines.
AKG came out with a shot of addressing this market with the K167 "Tiësto" headphones. They were intended for modern dance/techno fans while also being touted as monitors. They managed to get really low without fogging up the lower mids. However, they broke really easy. They were being cleared out for $50 a pair via B + H until they ran out of stock. I've owned multiple pairs, gave some away, rebought. The pair I got most recently is falling apart. After I got the Drop 6XX I couldn't help but notice that there was a presence bump with the Tiësto cans, much like Sony's MDDR V-6/7605 'phones.
I guess Tiësto's in the rear view mirror now, and AKG does what it can to sell as many headphones as they can. I'd say that the K 371 is a continuation of what AKG was doing when they came out with the Tiësto headphones and that the Harman curve takes into consideration that segment of the market.

 

[watchnerd]

$49.95 on Amazon, still in stock:

https://www.amazon.com/AKG-Pro-Audi...t=&hvlocphy=9033296&hvtargid=pla-625557334272

 

[Robin L]

$49.95 on Amazon, still in stock:

https://www.amazon.com/AKG-Pro-Audi...t=&hvlocphy=9033296&hvtargid=pla-625557334272

If I ever want another pair [and I won't].

 

[DEALUX]

Only if the IEMs in question don't fare well in regards to boosted bass. If they do, they won't spray whole midrange with THD, alas it'll all be clean & tight. Common misconception.

I think even the $7 Sony MH755 has less distortion in the bass than most full size headphones. That's not a factor. I have the Thieaudio Monarch which closely follows the Harman IE target but the issue is that the sub sometimes overpowers the upper bass, not midrange. So you do not get the smooth texture that you get on planars or EQ-ed dynamic drivers (HD800 for instance) but resonant sub bass that overpowers the upper bass frequencies.

 

[preload]

Every time I've EQ's a pair of headphones using Oratory's PEQ filters, the end result has been horrible. Unlike loudspeakers, there's apparently quite a bit more to listener preferences than matching a particular FR curve. As others have mentioned, Harman's own research shows substantial variability in listener preferences in bass/treble quantities, as a function of their age, gender, and country of residence.

 

[watchnerd]

Every time I've EQ's a pair of headphones using Oratory's PEQ filters, the end result has been horrible. Unlike loudspeakers, there's apparently quite a bit more to listener preferences than matching a particular FR curve. As others have mentioned, Harman's own research shows substantial variability in listener preferences in bass/treble quantities, as a function of their age, gender, and country of residence.

I think this is worth repeating, as there sometimes seems to be an attitude that something's wrong with the listener if the various PEQ filters that move things closer to the Harman curve aren't liked.

To quote from the public preso:

”Harman curve Lovers”: This group, which constitutes 64% of listeners, includes mostly a broad spectrum of people, although they’re generally under age 50. They prefer headphones tuned close to the Harman curve.

“More Bass Is Better”: This next group, which makes up 15% of listeners, prefers headphones with 3 to 6dB more bass than Harman curve below 300Hz, and 1dB more output above 1kHz. This group is predominantly male and younger — the listeners JBL is targeting with its headphones.

“Less Bass Is Better”: This group, 21% of listeners, prefers 2 to 3dB less bass than the Harman curve and 1dB more output above 1kHz. This group is disproportionately female and older than 50.”


So in a nutshell, 36% of the listeners, overall, don't prefer the Harman curve, which is a sizable chunk.

Weirdly, (or perhaps because of) given I play bass, I fall in the "Less Bass is Better" camp.


So the question then becomes:

If you're in one of the outlier groups, what EQ target should you normalize to?

 

[Feelas]

I think this is worth repeating, as there sometimes seems to be an attitude that something's wrong with the listener if the various PEQ filters that move things closer to the Harman curve aren't liked.

Is there? I'm feeling that there's something akin to that perception... but only on the side of people who don't believe that Harman's the end-all, but discrediting anyone doing the research instead of merely confirming that they just don't like the curve.

 

[preload]

Is there? I'm feeling that there's something akin to that perception... but only on the side of people who don't believe that Harman's the end-all, but discrediting anyone doing the research instead of merely confirming that they just don't like the curve.

The thing is, Harman's research doesn't actually demonstrate that the curve is the "end-all." I'd reckon that most people who "quote" the published Harman headphone research haven't actually read the actual papers themselves. I could be wrong, but that's the feeling I get. To be clear, the Harman curve is pretty good for IEM's, but "just okay" for OE/AE's based on a couple of their papers using an "interesting" methodology.

 

[Feelas]

The thing is, Harman's research doesn't actually demonstrate that the curve is the "end-all." I'd reckon that most people who "quote" the published Harman headphone research haven't actually read the actual papers themselves. I could be wrong, but that's the feeling I get. To be clear, the Harman curve is pretty good for IEM's, but "just okay" for OE/AE's based on a couple of their papers using an "interesting" methodology.

And that feeling is possibly true, so what we should do (at least here on ASR, maybe?) is to try to prevent the herd mentality in regards to anything.

Oratory's Reddit is... well, Reddit, so that isn't something unexpected...

The curious part: what do you mean by "interesting" methodology? I don't think there's anything stinky there, having read two of three of these.

@watchnerd seeing how far the variance in HRTFs goes, I'd be really, REALLY, careful interpolating what the rest should use, simply because the variance is much bigger than merely "3-6dB more in bass and 1dB more of treble" - this just doesn't add up!

 

[Robbo99999]

I think this is worth repeating, as there sometimes seems to be an attitude that something's wrong with the listener if the various PEQ filters that move things closer to the Harman curve aren't liked.

To quote from the public preso:

”Harman curve Lovers”: This group, which constitutes 64% of listeners, includes mostly a broad spectrum of people, although they’re generally under age 50. They prefer headphones tuned close to the Harman curve.

“More Bass Is Better”: This next group, which makes up 15% of listeners, prefers headphones with 3 to 6dB more bass than Harman curve below 300Hz, and 1dB more output above 1kHz. This group is predominantly male and younger — the listeners JBL is targeting with its headphones.

“Less Bass Is Better”: This group, 21% of listeners, prefers 2 to 3dB less bass than the Harman curve and 1dB more output above 1kHz. This group is disproportionately female and older than 50.”


So in a nutshell, 36% of the listeners, overall, don't prefer the Harman curve, which is a sizable chunk.

Weirdly, (or perhaps because of) given I play bass, I fall in the "Less Bass is Better" camp.


So the question then becomes:

If you're in one of the outlier groups, what EQ target should you normalize to?

All that data you quoted is based on testing around the Harman Curve, so if you're one of the "21% of listeners, prefers 2 to 3dB less bass than the Harman curve and 1dB more output above 1kHz", then what you'd do is lower the Low Shelf Boost Filter that is at 105Hz down by 2 or 3dB, and then you'd add a High Shelf Filter at around 1100Hz with a 1dB Gain. If you were in any of the other brackets then you'd take the same approach....turning the words/description into parametric filters. Oratory1990 who is famous for his headphone measurements & EQ's to the Harman Curve, he provides guidance at the bottom of his pdf files (for each headphone) on which filters to change to tailor the sound to your own preferences. He always includes the Low Shelf Filter where you can tweak the bass level to your hearts content, but he doesn't always have the High Shelf Filter in place to tweak the treble preference, but that's relatively easy to add yourself, the 1100Hz High Shelf Filter I mentioned. So, in conclusion, I think the easiest/best approach to take if you don't like the Harman Curve settings "out of the box" then tweak the bass and/or treble like I've described.....then if you still can't balance it then consider yourself a Harman Curve hater and look for different ways to EQ your headphone (although if you're confident using REW you could make some more subtle adjustments like using Peak Filters to change the amplitude of the 3kHz Hump to see if that fits you better, before throwing out the approach of "tweaking around the Harman Curve").

 

[preload]

And that feeling is possibly true, so what we should do (at least here on ASR, maybe?) is to try to prevent the herd mentality in regards to anything.

You mean like the Revel/Rythmik/Anti-B&W herd mentality? Ooops, said too much.

The curious part: what do you mean by "interesting" methodology? I don't think there's anything stinky there, having read two of three of these.

Well, looking at Olive et al "A Statistical Model that Predicts Listeners’ Preference Ratings of Around-Ear and On-Ear Headphones" AES Convention Paper 9919 (2018), the authors obtained preference ratings for 32 headphones and used it to create a prediction formula based partially on the degree of deviation from their Harman curve. But instead of testing the actual 32 headphones, they simply gave their listeners the same model of headphones, and then EQ'd it to within +/-1 dB tolerance (not great) up to 12kHz (seriously?) in order to "simulate" each of the actual 32 different headphones. Yes, it eliminated some of the bias from the weight/feel of each headphone, but we all know that EQ'ing an HD650 isn't going to turn it into an HD800 or an Orpheus, especially with +/- 1dB tolerance up to 12kHz. So, I found that "interesting." Perhaps you have a different take on the paper.

Also, just to get a reality check on this paper, I purchased their #2 preferred headphone, and it sound like dog crap (and so did Amir in his official review, but he was a little nicer about it). I also found that "interesting."

 

[Feelas]

You mean like the Revel/Rythmik/Anti-B&W herd mentality? Ooops, said too much.

Being in neither, I bet we can agree we'd both say too much. Whoopsie!

Well, looking at Olive et al "A Statistical Model that Predicts Listeners’ Preference Ratings of Around-Ear and On-Ear Headphones" AES Convention Paper 9919 (2018), the authors obtained preference ratings for 32 headphones and used it to create a prediction formula based partially on the degree of deviation from their Harman curve. But instead of testing the actual 32 headphones, they simply gave their listeners the same model of headphones, and then EQ'd it to within +/-1 dB tolerance (not great) up to 12kHz (seriously?) in order to "simulate" each of the actual 32 different headphones. Yes, it eliminated some of the bias from the weight/feel of each headphone, but we all know that EQ'ing an HD650 isn't going to turn it into an HD800 or an Orpheus, especially with +/- 1dB tolerance up to 12kHz. So, I found that "interesting." Perhaps you have a different take on the paper.

Yet we're all here in doubt whether it's possible to measure anything reliable on testrigs in regards to 12kHz, so in that regard I'm confident that in regards to this, it's impossible to do it the other way. I think that the whole "EQ-d isn't the same" can be explained away scientifically; and somehow Olive et. al mention, that the non-FR based factors were omitted from the study, simply because it was done for other reasons. Why would you analyze the actual headphones in question, rather their FR curves, if you're going to be working on a driver w/ perceptibly low THD and the desired FR, rather than working out which headphones will sell. This is not the same! The work was from the start based to assess which of the thirty FRs they used on K712 worked best. Simply because they can work w/ a low THD driver and chisel the FR, it'd be unnecessary to see how THD affects the buying preference. Add to that, that most people (I'm myself a terrible unskilled listener) don't necessarily perceive and/or care about factors other than FR and we're done on the commerical grounds. Same goes for 12kHz cutoff - yes, it matters, but the FRs in questions all had the same level post-12kHz, unless the EQs were flawed. We can digress for days whether the 12k+ of K712 overthrows the tonal balance of anything else than K712 so far, that the whole study is pointless - this might be the case!

From here all goes downhill, having flawed ideas like assessing the ratings based on FR-only, disregarding the THD & other factors. Shortly said, we're delving into the cult. Olive et al mention the faults outright and here I'd drop the rant on my side, simply because IMO the rest isn't on them - it's not their fault, that people who can't for themselves assess the importance of other parameters (GD, THD, etc.), buy into the narrative that FR is the only thing that matters. On that ground, it's hard to dump the faults on Harman, isn't it? The methodology is flawed, it's all laid bare - the covered FR is partial, THD is not assessed, other distortion factors are not addressed.

Also, just to get a reality check on this paper, I purchased their #2 preferred headphone, and it sound like dog crap (and so did Amir in his official review, but he was a little nicer about it). I also found that "interesting."

You mean the MDR's? Well, THD in bass is terrible. Cut the THD and leave the FR - and I believe we'd all be happy owners of 7506's.

 

[Robbo99999]

You mean like the Revel/Rythmik/Anti-B&W herd mentality? Ooops, said too much.



Well, looking at Olive et al "A Statistical Model that Predicts Listeners’ Preference Ratings of Around-Ear and On-Ear Headphones" AES Convention Paper 9919 (2018), the authors obtained preference ratings for 32 headphones and used it to create a prediction formula based partially on the degree of deviation from their Harman curve. But instead of testing the actual 32 headphones, they simply gave their listeners the same model of headphones, and then EQ'd it to within +/-1 dB tolerance (not great) up to 12kHz (seriously?) in order to "simulate" each of the actual 32 different headphones. Yes, it eliminated some of the bias from the weight/feel of each headphone, but we all know that EQ'ing an HD650 isn't going to turn it into an HD800 or an Orpheus, especially with +/- 1dB tolerance up to 12kHz. So, I found that "interesting." Perhaps you have a different take on the paper.

Also, just to get a reality check on this paper, I purchased their #2 preferred headphone, and it sound like dog crap (and so did Amir in his official review, but he was a little nicer about it). I also found that "interesting."

However, it does isolate the variable of Frequency Response up to 12kHz, which is about the functional limit for how high you can accurately measure a headphone on a fixture.....and Frequency Response is what they have control over to manipulate "for the good of the people". Having lots of different headphones in the test would muddy that variable, because impedance between ear & headphone change (hopefully I've worded that right) based on the individual and the particular headphone used....which affects the frequency response....so by using the same headphone then they're eliminating that variable and focussing on the effect of the measured frequency response - which is all they can do really. And as you say it takes away any kind of bias associated with physical feel of the headphone on your head....plus it means you don't have to keep putting on & taking off headphones, all of which will change the frequency response somewhat just by virtue of a different headphone position placement (even when taking off & replacing the same headphone) - so I imagine it would have eliminated that variable too, although I don't know the specific regime of how they cycled through their testing.

 

[watchnerd]

Is there? I'm feeling that there's something akin to that perception... but only on the side of people who don't believe that Harman's the end-all, but discrediting anyone doing the research instead of merely confirming that they just don't like the curve.

What do you mean by 'the end all'?

 

[Feelas]

What do you mean by 'the end all'?

I've cut the corners of thoughts too fast; I mean that some people are merely trying to fight the idea of there not being an ultimate curve - and that's all fine, since there won't be any in the future, never, but this idea is not inherent to the Harman curve itself.

 

[watchnerd]

@watchnerd seeing how far the variance in HRTFs goes, I'd be really, REALLY, careful interpolating what the rest should use, simply because the variance is much bigger than merely "3-6dB more in bass and 1dB more of treble" - this just doesn't add up!

Well, even aside from the HRTF....

"3-6 dB more in bass" doesn't tell the shape of the curve needed, anyway.

 

[watchnerd]

I've cut the corners of thoughts too fast; I mean that some people are merely trying to fight the idea of there not being an ultimate curve - and that's all fine, since there won't be any in the future, never, but this idea is not inherent to the Harman curve itself.

What do you mean by an ultimate curve?