Master Complaint Thread About Headphone Measurements

[PeteL]

So for the first bit, can I not adhere to this notion and say no EQ is thus reliable when it comes to IEMs (or wherever else you think this applies)? Thus making EQ basically a random tossup by this logic?

I think you missed the part that he is talking about sharp variation in upper treble. It is true that these are unreliable measurement since it will vary on individuals and measurement rigs.

 

[markanini]

Exactly because of my iPhone I got the Variations.

You payed $500 for a IEM. That would make many in your shoes cheer when less expensive IEMs appear to graph worse. However one big issue I see with Variations graph is a broad recession in the upper bass. That's a tuning move for achieving a fake clarity, at the expense of enjoyment to most listeners, as this is a important frequency range for voices, rock guitar and snare drum.

 

[Tks]

I think you missed the part that he is talking about sharp variation in upper treble. It is true that these are unreliable measurement since it will vary on individuals and measurement rigs.

I didn't, but the reason I didn't address it directly, is because he finds it weird we ignore 8kHz+ (as per his video). If he is aware of the fact of your second sentence, then I'm not sure why he would term our relaxed stance on high treble content as "weird". To be clear, I'm simply willing to grant him his position just so I could get my question answered.

 

[Sharur]

but seeing as how there's no point in this level of granularity

But I can audibly detect errors in this level of granularity, so surely there cannot be "no point"?

Let me ask you point blank to settle this ordeal to my satisfaction. You not only claim to hear 20kHz+ in isolation obviously with no issue at all - what you're also now seemingly telling me, is you can also train yourself on top of that, to hear it in actual musical content? And you do this by being able to discern some sort of speed things of sound?

Just to be crystal clear here - the VERY LEAST, you can hear 20kHz+ tones without any issue, while music content obviously is a bit harder for obvious reasons.

I just don't get it - do you not find that to be an insanely exceptional ability? I'm talking golden ears ACTUALLY being substantiated in reality.. Why not get yourself medically verified for such ability, this would be groundbreaking to my knowledge at your age? Or do you just take this to be something people are limited simply because they aren't taking the time to do this training of sorts?

I did not say I can hear it with no issue at all. I have to use a significant amount of my listening ability I have trained over the years to pass this test with 99.95% confidence. I also did not claim to be able to pass this test with actual musical content. My observations are limited to the specific test in question. The point is that, if I can hear over 20 kHz with this test, I can probably hear pretty high with actual musical content. But yes, for this test, I am able to discern the difference by the perceived crispness (or speed) to the sound.

I can hear up to the limit of this site, although I have to turn my MacBook HP jack volume up to slightly below half to be able to blind test it "without any issue." I think one of the biggest limiting factors is that people do these tests with transducers that do not extend high enough. For example, if I were to do this test with transducers that significantly dropped off after 16 kHz, like the Etymotic ER4S, I would not be able to hear frequencies past 20 kHz. Other than that, I think a lot of people are not trained enough to be able to hear past 20 kHz. I've been thinking about audio for several hours a day for years at this point with many ABX tests to train myself, so my brain is probably more in tune with what my inner ear hair cells detect than the average person.

 

[Tks]

But I can audibly detect errors in this level of granularity, so surely there cannot be "no point"?

I mean, I don't understand the point of using "high Q factor", that's what I meant by "no point", as in - no point in EQ'ing in such a way. Btw, how high are we talking just as an aside. And also show me what these "artifacts" are, where can I read up on this stuff, and how can it be teased out definitively?

I did not say I can hear it with no issue at all. I have to use a significant amount of my listening ability I have trained over the years to pass this test with 99.95% confidence. I also did not claim to be able to pass this test with actual musical content. My observations are limited to the specific test in question. The point is that, if I can hear over 20 kHz with this test, I can probably hear pretty high with actual musical content. But yes, for this test, I am able to discern the difference by the perceived crispness (or speed) to the sound.

Well, look I won't call you a liar, but I'll ask again since you seem to not want to answer any questions directly. Am I missing something when I say when I claim you're basically a golden ear medical specimen if you can actually hear 20Khz+ content? Why not actually get this verified? The reason I keep saying this, is because it's THAT stunning to me if actually true. If you actually get something like this verified, I'd be inclined to just accept basically any claim you make about sound quality going forward, with more trust than even FR plots since high treble content isn't really something most measurement rigs can reliably measure anyway.

I can hear up to the limit of this site, although I have to turn my MacBook HP jack volume up to slightly below half to be able to blind test it "without any issue." I think one of the biggest limiting factors is that people do these tests with transducers that do not extend high enough. For example, if I were to do this test with transducers that significantly dropped off after 16 kHz, like the Etymotic ER4S, I would not be able to hear frequencies past 20 kHz. Other than that, I think a lot of people are not trained enough to be able to hear past 20 kHz. I've been thinking about audio for several hours a day for years at this point with many ABX tests to train myself, so my brain is probably more in tune with what my inner ear hair cells detect than the average person.

Again that'd be great. But we still don't know what this training entails tbh (maybe someone here can enlighten me, I know of training to discern MP3 vs FLAC, but to hear 20kHz+? It's just a big question mark to me). On top of the exceptional notion of someone beyond their early teens being able to hear 20Khz at all.

---------------------------------------------------------------------------------------------

But just going back to my original comment since there has been a massive tangent about your abilities. I'll ask again, given that these IEM's are ~$50, THD vanishingly low (which I'll assume you have no problem with, as low THD in my book is by definition always a good if we're talking headroom for EQ'ing for instance). Are there any sound problems one would still find on these once they do EQ them? And please, leave out this odd "high Q factor artifacting" ordeal, since I've never been exposed to this, nor do I know precisely what you're refering to. But if we stay below these "high Q factor values" (So like ~1? max) - is there any possibly complaints with sound with respect to this IEM once you EQ it?

 

[markanini]

I didn't pay $520, I traded it for an used Dt1990 for which I had paid $300 years ago. But to the point, the upper bass recession is, as far as I can judge, exactly how a subwoofer would kick in in a loudspeaker system. It appears very natural to me and sounds great.

So you're saying you like something that looks wrong on the graph, a product that costs $520 new. Did you carry out blind tests against other products to verify your preference?

 

[CedarX]

…….our ears adapt to basically anything after a few minutes, besides honky harsh treble garbage I suppose).

Don’t want to derail the debate… but to me this is one of the largest elephant in the EQ room. I own HPs that are being trashed for their significant deviation from whatever target, yet I can absolutely get used to them and they don’t make my tracks less enjoyable…

Here’s my complaint about headphones measurement: what deviation, value, threshold, or any other characteristic, may (will?) result in these “honky, harsh, (treble or other) garbage” that we can’t adapt to?

 

[amirm]

My position is not that the high frequency measurements are 100% accurate, but that if you understand the features the measurement rig is imparting, you can gather useful information.

Well you can't. Until you have done dozens of measurements yourself on these test rigs, you don't know what you don't know. Touch the IEM and response can change drastically. The fixture remains the same so your argument doesn't hold. This is on top of smoothing, averaging, etc. that folks do to publish measurements.

 

[amirm]

Here’s my complaint about headphones measurement: what deviation, value, threshold, or any other characteristic, may (will?) result in these “honky, harsh, (treble or other) garbage” that we can’t adapt to?

There is no way to answer this because it highly depends on the spectrum of music you are listening to. If a track is already bright and you listen to it with a headphone that has exaggerated highs, you likely won't tolerate them long enough to get used to them. Then again if the variation is small, then you might and be OK with them.... until you hear a more accurate headphone which then resets your standards.

 

[amirm]

can hear up to the limit of this site, although I have to turn my MacBook HP jack volume up to slightly below half to be able to blind test it "without any issue."

Those tests are very unreliable because your default audio pipeline in Mac/Windows creates artifacts that can be audible. You need a bit-exact pipeline which is not going to happen in a browser.

 

[DualTriode]

I can hear up to the limit of this site, although I have to turn my MacBook HP jack volume up to slightly below half to be able to blind test it "without any issue."

Sample size N := 1

You may hearing amplitude modulation, distortion at its worst, not 20khZ+ frequency.

I believe that you are confabulating.

 

[solderdude]

Interesting. Might you tell us which model it was?

I don't know if they want people to know. It was a Massdrop model.

Sure, but if some very well trained listeners with good high-frequency hearing report deficiencies (not talking about me, unfortunately), they are relevant, even if they cannot reliably tested on a rig.

If multiple trained listeners do then yes.

 

[solderdude]

You do not know how a flat anechoic speaker will be heard (individual eardrum response) by yourself or others, but this is independent from fidelity as natural sound perception is accounted for by each individual's features and an effort is not made to EQ the speaker to match one's HRTF.

I do know how 'anechoic flat speaker' is heard by me. In fact I could listen to that all day thanks to a measurement mic. Most headphone listeners don't have that reference though and assume that their headphone with this or that computer generated EQ is how it should sound. In the ballpark they might be right but above several kHz they might be wrong.
No way to tell without a reference.

 

[PeteL]

I didn't, but the reason I didn't address it directly, is because he finds it weird we ignore 8kHz+ (as per his video). If he is aware of the fact of your second sentence, then I'm not sure why he would term our relaxed stance on high treble content as "weird". To be clear, I'm simply willing to grant him his position just so I could get my question answered.

OK, what is "our" relaxed stance on high high treble content? Is it that measurements are unreliable up there, which I agree with, or is it that any issue in the treble, either measured or heard, is non existent and nothing can be done about it with EQ?

 

[Jeromeof]

I think the lack of reliable high treble measurements allows the subjectives to talk about "technicalities" whether they are real or not.

 

[Sharur]

Those tests are very unreliable because your default audio pipeline in Mac/Windows creates artifacts that can be audible. You need a bit-exact pipeline which is not going to happen in a browser.

Mac OS over 15 years ago had all resampling distortion below -150 dB. I doubt it's worse now than it was then. I do not know if you are talking about something else though.

The tone generator thing is tangential to the test I was posting results for, which is an ABX test for 20 kHz band limited white noise vs 22 kHz "full range" white noise. https://www.audiocheck.net/blindtests_frequency.php?frq=20

 

[markanini]

Come on, who says it looks wrong on the graph? If you ask me, what looks wrong is the Red, not the Variations (compare the region 100 - 600Hz).View attachment 287234

I'm not telling you what to do. Reviewers might insist that you just trust them. I think I might have done some blind testing to still my curiosity. Especially to justify a $520 IEM as an objectivist.

BTW Red is a product that's differs from Harman IE target by design, especially in Bass+ mode. That's confirmed by the graph you shared.

 

[PeteL]

I think the lack of reliable high treble measurements allows the subjectives to talk about "technicalities" whether they are real or not.

OK, I wanted to talk about earphones, You wan't to talk about types of listeners. I'll take "whether they are real or not" as the key part, although I do not fully know what "technicalities" really means.

 

[markanini]

how possibly could you

Well, I'm not a reviewer so I wouldn't, even if you seem fixated on that framing. I'm a user in a discussion with another user(I assume?) about the objective merits of a $520 IEM. According to Harmans predictive model the following IEMs have a higher preference score according to Harmans predictive model.

 

[Jeromeof]

OK, I wanted to talk about earphones, You wan't to talk about types of listeners. I'll take "whether they are real or not" as the key part, although I do not fully know what "technicalities" really means.

Neither do I - but allowing part of the frequency range to be "free form" means people will put their own spin on that part of the spectrum and therefore probably justify why one IEM might cost $500 and another only cost $55 with effectively the same FR.