Is it time for a new Harmon Curve survey?

Dec 10, 2024 at 4:06 PM Post #106 of 148
Too bad the thread devolved into personal attacks having nothing to do with the Harmon curve.

How many people commenting here actually EQ their headphone playback? How many have used the AutoEQ site to actually EQ their headphones to a measured Harmon curve? And of those, how many find fault with what they are then hearing and change the settings further to get a more accurate to real life sound? Or, switch off the EQ as sounding worse than no EQ to an exact Harmon curve as I did?

I find that with my HiFiMan HE1000SE headphones that when EQed to the Harmon curve that there is way too much bass. Even after I cut the prescribed boost settings in half. And the highs are rolled off. And the lower mids are too prominent, leading to a muddy sound.

I would hope that if you have never EQed your headphones, that you would refrain from commenting.

Those that have actually EQed their headphones to the Harmon curve according to measurements from the AutoEQ site, and Preferred the new sound as it is, or altered the settings, or turned the new EQ right back off as I did, I would be happy to hear your experience.

https://autoeq.app/
I'll take some credit there. Sometimes I get a little salty and I apologize.

I used AutoEQ on my DCA E3 and Stealth headphones to establish a baseline and then I tweak individual bands. Yeah, it's not the same as using measuring equipment to EQ these headphones exactly, but all E3s and all Stealths are going to have really similar characteristics in their frequency response that makes tools like AutoEQ valuable. There's probably more variance in cheaper headphones so I won't comment there (I haven't tried EQing my Audio-Technica or Beyers).
 
Dec 10, 2024 at 4:13 PM Post #107 of 148
I dont think they would do it by ear, how would that work at all? They rely on autoEQ or squig or some equivalent software to generate the filters and they get the data from some measurement database such as the ones by oratory, Tyll, crinacle. Maybe someone made the database by using the same type of rig the research was using. Of course people using autoEQ wont have a BK measurement rig or any rig at all...
 
Last edited:
Dec 10, 2024 at 4:25 PM Post #108 of 148
I used AutoEQ on my DCA E3 and Stealth headphones to establish a baseline and then I tweak individual bands.
I agree that whatever unit to unit variation there may be between my headphones and the the measurements that AutoEQ is using are very small compared to the vast changes of several db that I would have to make to get the harmon eq to sound good.

This is the feedback I am looking for.

What changes do you make from the prescribed settings?

I cut the bass shelf that was prescribed by 3db and it was still way too much bass. The highs sounded rolled off. And the mids from 1-2k were slightly elevated compared to what I would say sounded natural.
 
Dec 10, 2024 at 4:31 PM Post #109 of 148
I dont think they would do it by ear, how would tjat work at all? They rely on autoEQ or squig or some equivalent software to generate the filters and they get the data from some measurement database such as the ones by oratory, Tyll, crinacle. Maybe someone made the database by using the same type of rig the research was using. Of course people using autoEQ wont have a BK measurement rig or any rig at all...
Sounds like a bit of Russian roulette then; I can't speak for other brands but over the years I have taken a number of Sennheiser headphones apart of supposedly the exact same model and revision, and it was obvious that internal changes are made gradually throughout the production years without necessarily advertising it or rebranding it as a mark-2 model. Different mesh materials, different felt materials, different membrane damping materials etc. At least the changes from 600Ohm to 300Ohm were mentioned on the box...

That's why some folks pay so much attention to seemingly innocuous details like whether the headband clips say made in Germany or made in Ireland. It is not that the headband matters per se but it is an indicator of the exact production year model you are looking at in the absence of serial number batch info...

Do those measurement databases go into that much detail, like e.g. different curves for different serial number ranges?
 
Dec 10, 2024 at 4:44 PM Post #110 of 148
I have no idea, I only ever check them for sensitivities. Tyll had one measurement for HD800 and one for HD800S. Not like it matters a lot because unit to unit variation is probably the smallest problem with this method. It's dwarfed by both the differences between the couplers (used by Harman and the databases) and the differences between those and your head. Oratory also shows confident intervals, essentially measurement to measurement variation with his rig. Which can be also bigger than unit to unit variations.

Of course despite this, autoeq will still nudge the headphone toward the target especially if it was far from it to begin with but without measuring the result (where you will run into the problems already mentioned, isnt that funny?) there is no way to know how much the difference is exactly, one can only guess.
 
Last edited:
Dec 10, 2024 at 9:00 PM Post #111 of 148
I agree that whatever unit to unit variation there may be between my headphones and the the measurements that AutoEQ is using are very small compared to the vast changes of several db that I would have to make to get the harmon eq to sound good.

This is the feedback I am looking for.

What changes do you make from the prescribed settings?

I cut the bass shelf that was prescribed by 3db and it was still way too much bass. The highs sounded rolled off. And the mids from 1-2k were slightly elevated compared to what I would say sounded natural.
If you're truly used to how your headphones sound (say you only have one pair of headphones that you use exclusively), I don't think any EQ change is going to sound "for the better" simply because your reference is the same pair of headphones you're already wearing. I increased the sub bass (so 20-40Hz) and moved the high cut out a couple thousand Hz (I think it's centered at 10kHz generally).
 
Dec 10, 2024 at 11:18 PM Post #112 of 148
I agree that whatever unit to unit variation there may be between my headphones and the the measurements that AutoEQ is using are very small compared to the vast changes of several db that I would have to make to get the harmon eq to sound good.

Sample variation within the same make and model of headphone can vary as much as +/- 3dB. More expensive headphones are more expensive because their quality control for consistency is tighter. I was told this by the headphone designer who designed Oppo’s PM-1s. I suspect that the PM-2s were PM-1s that didn’t meet tolerances.
 
Last edited:
Dec 10, 2024 at 11:23 PM Post #113 of 148
Sample variation within the same make and model of headphone can vary as much as +/- 3dB. More expensive headphones are more expensive because their quality control for consistency is tighter. I was told this by the headphone designer who designed Oppo’s PM-1s. I suspect that the PM-2s were PM-1s that didn’t meet tolerances.
At least the drivers I’m sure. I thought they had fairly different “chassis” right?
 
Dec 10, 2024 at 11:26 PM Post #114 of 148
The engineer told me that the most expensive part of the PM-1 was the pivot that connected the cups to the headband. Apparently, it is made from a more durable material and was machined to precise dimensions to not wear out or break over time. The drivers in both models were the same.
 
Dec 11, 2024 at 6:59 AM Post #115 of 148
The engineer told me that the most expensive part of the PM-1 was the pivot that connected the cups to the headband. Apparently, it is made from a more durable material and was machined to precise dimensions to not wear out or break over time. The drivers in both models were the same.
Bonkers really; all to make the headphones look a bit more 'modern stylish' and such that you can wear a hoody over them. Perfectly adequate, very durable and cheap solutions for this have been around since the 80's. But they don't look as good. Nowadays there are quite a few overly complex and mechanically unnecessarily stressed headphone pivot designs out there.

But this is OT.
Sample variation within the same make and model of headphone can vary as much as +/- 3dB. More expensive headphones are more expensive because their quality control for consistency is tighter.
No kidding; in the past I have returned headphones to the shop because L/R matching was too poor; exchanged it twice until I got a sample that was fine to my ears. And these were top-of-the-line Sennheisers.
 
Last edited:
Dec 11, 2024 at 7:02 AM Post #116 of 148
Ummm because its more likely I know what an oboe should sound like on a very good recording more than a bass head that listens to EDM only.
Given that extreme, then very probably. However there’s two obvious points: Firstly, who is more likely to know what an oboe on a particular recording should sound like; you or another audiophile, a rock drummer, a professional sitar player or, the engineer who recorded it (or another engineer who works in the orchestral field), a conductor, a classical composer, a classical producer, an oboe maker, an orchestral venue acoustician, etc.? Even given your specific example of an oboe, then the “pro acoustic musicians” and “audiophiles with live acoustic music experience” are not a “good idea” or at least not as good as others would be.
Secondly, using your same examples but flipping the question around, who is more likely to know what a non-acoustic bass “should sound like” or notice features of a non-acoustic bass, you or the bass head who listens to EDM? Not all music is acoustic and features an oboe, if fact the opposite, the vast majority of consumed music recordings are not of purely acoustic ensembles featuring an oboe, the vast majority features a non-acoustic bass!
I am not interested in suppressing the opinions of others per se, however, I believe the sub sets I have identified are likely to yield results closer to what my perceptions are than a wad.
How is a suggested testing methodology that eliminates/ignores certain other demographic groups not effectively “suppressing the opinions of others”? And, the Harmon curve is a test based on the preference of humans, not a test of your personal perceptions. It would not be a “good idea” to base a test on your personal perceptions or to bias a test so it “yields results closer to what your perceptions are” except for you personally and why would anyone apart from you fund a test/study specific to your personal perceptions?
[1] Trying to suck me into a straw man argument..... [2] I posit that different self identified sub groups would end up with different results. [3] I would place more trust in the ones I specified, but, I would still verify by ear the accuracy of the settings.
1. I did not employ a strawman argument, I used your argument just taken to it’s logical conclusion.
2. And yet they don’t! The scientific evidence indicates age, gender and to a limited extent trained listening are the differentiators, not the sub groups you posit, with the latter more in terms of consistency rather than preferences.
3. You can place trust in whatever you want but in this subforum we obviously place trust in the scientific evidence or at least more trust than in what an audiophile posits, especially when their own example contradicts what they’ve posited!
The Hall is the same (basically), of course the players and composer are different.
The hall is the same but the sound is not, it varies very significantly depending on the location within that same hall. So, that’s none of the determining variables that “experience of live acoustic music” will tell you!
Still a higher level of accuracy than Led Zeppelin in sports arena.
And who said “accuracy” when creating a recording is the goal or was ever the goal? Maybe that’s what you’re looking for but that would be “a fool’s errand” as it’s virtually never the goal of those creating the recordings.
The subject of music, recordings, and playback has not been specified, so that's an open question.
Exactly, because the Harman curve is designed as a general listening preference of humans irrespective of the music or recording/mixing. And certain recordings may still need adjustments because not all recordings are perfectly balanced. However, you did specify the music and recordings because the BSO in a concert hall implies orchestral music and a specific (orchestral) style of recording and mixing. But even on the basis of you closing the question hugely, to this particular genre and recording style, still your assertion is incorrect, as I explained above.
While it is a good example of 'modern' recording of pop music - it’s not live, it is amplified. Not my reference. I'd be looking at 3 ribbon mike recordings of acoustic music played in a good hall with minimal post processing as being more accurate.
So your reference is a handful of recordings (out of countless millions of commercial releases) from half a century and more ago, which were already outdated even in their own time and have hardly ever been used since the early/mid 1970’s? Your reference excludes almost all recordings and almost everyone!
Have not read it. Age is a factor obviously.
Why would you argue and/or make counter “posits” in a science discussion forum without reading/knowing the science?
It’s not either/or. It’s a preference measurement - that appears to work for many. In the forms I have sampled it. I do not accept the results as acceptable as is.
You haven’t read the results but nevertheless you “do not accept the results as acceptable”?

G
 
Last edited:
Dec 11, 2024 at 8:15 AM Post #117 of 148
Bonkers really; all to make the headphones look a bit more 'modern stylish' and such that you can wear a hoody over them.

It really wasn’t style. It’s the functionality and structural integrity of the pivot. That affects the comfort, ease of folding up the cans, and strength of the weakest point of any headphone. It doesn’t look any different than any others, but I haven’t ever experienced many midrange cans that feel like this. The Apple AirPods Max sort of do. It falls under the category of build quality.
 
Dec 11, 2024 at 11:31 AM Post #118 of 148
It really wasn’t style. It’s the functionality and structural integrity of the pivot. That affects the comfort, ease of folding up the cans, and strength of the weakest point of any headphone. It doesn’t look any different than any others, but I haven’t ever experienced many midrange cans that feel like this. The Apple AirPods Max sort of do. It falls under the category of build quality.
OK, my bad; maybe not bonkers. Foldability and portability of headphones is something I didn't consider as it has no relevance whatsoever to my use of headphones, but I can see it will be quite relevant for others.
 
Dec 11, 2024 at 7:30 PM Post #120 of 148
I don’t know what you’re talking about. There are different Harman curves for in ear?
 

Users who are viewing this thread

Back
Top