[swspiers]

P.T. Barnum had a quote that related to that I think....

Man, if you could just let go of the snide commentary, it would be great. It seriously undermines the communication at times.

 

[Joe Bloggs]

I don't quite understand the correlation between raw frequency measurements and the HRTF.


The HRTF curve shows that the human body/head/ear amplifies signals at ~2- 5kHz. When a headphone is measured, are we looking for raw measurements that look similar to the HRTF curve? If the HRTF says ~2-5kHz are amplified, why would we want headphones that also have an amplified response in those frequencies? Shouldn't it have an attenuation at those frequencies such that when the HRTF is accounted for, we get a flat line?

The HRTF describes your head's response to a sound coming from far away - a perfectly flat response from a speaker will end up at your eardrum with more energy around 2-5kHz, and less energy elsewhere. Because of this, what sounds like a flat response is actually a boost in that frequency range (if measured directly at the eardrum). Headphones bypass much of that amplification, so to achieve that same response at the eardrum (with 2-5kHz emphasized), they must also emphasize that frequency range. This is especially true with IEMs, since they pretty much directly broadcast to the eardrum, and fully bypass the effects of the ear and head shape.

Oh I see. So headphone measurements should line up with the HRTF curve in order to emulate what we would hear as if we were listening to speakers with a flat frequency response. If that's the case, why do people who measure raw headphone responses, like Innerfidelity for example, not display the HRTF right next to the raw data in a different colour or something? If headphones are trying to mimic a frequency response that one would hear from a flat speaker response, wouldn't it just be common sense to publish the HRTF next to the raw data to make a direct comparison?

Innerfidelity shows raw data in gray on the frequency response graph. The red and blue lines are with a certain compensation curve dialled in (so that flat blue and red lines should indicate well-compensated headphones). However it seems that even Innerfidelity itself is resigned to the fact that their compensation curve is no good and resorts to describing the sound of the headphones relative to some other ideal reference line that the headphones should exhibit. That reference seems to consist of a rolloff in the extreme treble and a peak somewhere around 10kHz

 

[bigshot]

Man, if you could just let go of the snide commentary, it would be great. It seriously undermines the communication at times.

 
I have nothing against people who just don't know something and might be genuinely interested in finding out. I'm happy to share with people like that, and in fact, that is exactly why I'm here. But that kind of person can be the exception around here, not the rule. Too many audiophiles who wander into our forum stick to their predetermined conclusions like glue and grab at straws for justifications of their misconceptions, instead of trying to figure out how sound reproduction really works. When we take the time to share what we know with them, they get mad and go on the attack because feeding their ego is more important than actually finding out the truth. I don't have much respect for that, and I don't see any reason why I should. It's intellectually dishonest. My comment was pertinent to the question being asked... Why is there such a disconnect between sound science and the rest of the audiophile community?
 
I see conversation after conversation get derailed by people who aren't interested in discussing how to get good sound and how to recognize it when you get it. They push the subject into areas like "Science doesn't know everything! It used to think the Earth is flat!" and "But what about this tiny fraction of a percent? (that may be inaudible)" or "I don't recognize the validity of ABX testing." or "Subjectivity is just as valid as objectivity. I like what I like." Those are all smokescreen arguments. I don't have a lot of patience for them.

 

[luisdent]

yea, that's exactly what I'm saying. the fact is that the theory of gravity has an enormous weight of evidence...

wait... isn't the weight due to the existence of the gravity in the first place?

 

[Joe Bloggs]

wait... isn't the weight due to the existence of the gravity in the first place?

Is gravity responsible for this response because somebody threw this conversation into orbit 6 years ago and it landed back on your head just now?

 

[luisdent]

wow. didn't even notice. i guess it just keeps pulling people back.

 

[ScareDe2]

I am not sure over-ear headphones have treble peak by design. I think it's the reflection of the higher frequencies inside the cup that emphasizes the treble region. If I install pieces of fabric around the driver - leaving the driver section uncovered - then the treble aggressiveness is gone.

 

[ADUHF]

Old topic. But I'm not sure the OP's question was really ever answered.

Not all graphs, or perhaps more to the point, measurement systems show a peak at 8-10 kHz. This seems to be a feature though of the Head Acoustics systems used by sites like Rtings, and formerly Inner Fidelity. And is most likely the result of a resonance in the simulated ear canal (and possibly also the concha) at that frequency, at least on the over-ear headphone measurements.

 

[ADUHF]

The HRTF describes your head's response to a sound coming from far away - a perfectly flat response from a speaker will end up at your eardrum with more energy around 2-5kHz, and less energy elsewhere. Because of this, what sounds like a flat response is actually a boost in that frequency range (if measured directly at the eardrum). Headphones bypass much of that amplification, so to achieve that same response at the eardrum (with 2-5kHz emphasized), they must also emphasize that frequency range. This is especially true with IEMs, since they pretty much directly broadcast to the eardrum, and fully bypass the effects of the ear and head shape.

I don't know about IEM's. But on over-ears, I believe most of the boost at ~3 kHz is also caused by the shape of the ear. Particularly the ear canal, as shown in the above illustration and below. So it's not really in the headphones per se.

 

[ADUHF]

Oh I see. So headphone measurements should line up with the HRTF curve in order to emulate what we would hear as if we were listening to speakers with a flat frequency response. If that's the case, why do people who measure raw headphone responses, like Innerfidelity for example, not display the HRTF right next to the raw data in a different colour or something? If headphones are trying to mimic a frequency response that one would hear from a flat speaker response, wouldn't it just be common sense to publish the HRTF next to the raw data to make a direct comparison?

An excellent idea! (I wish I'd thought of it myself. )

Perhaps with the new 5128 measurements systems, with more accurate ear anatomy, something like this would in fact be possible. It would be nice to have this as a reference though even with current 711 systems.

Sure, you could do that. It is somewhat complicated by the fact that everyone has a slightly different HRTF, but you could at least use a fairly standard one as your point of comparison.

Yes indeedy!

 

[ADUHF]

A flat response is the correct response curve, but that isn't as easy to do as it sounds. Expensive headphones can do a nice flat response without resorting to tricks.

https://www.head-fi.org/threads/study-headphone-price-and-sound-quality.956469/

 

[ADUHF]

No headphone is flat past ~1k because basically. . . **** happens and it's complicated.

True dat.

 

[bigshot]

Yes, that study surprised me. That's why I posted it!

 

[ADUHF]

Yes, that study surprised me. That's why I posted it!

Fwiw, the title of your topic is a bit misleading imo. If they really wanted to do an in-depth study of the differences in sound quality between higher and lower cost headphones, then they should probably be looking at some other characteristics as well, such as driver symmetry, distortion, speed, spectral decay, extension, etc.. And not just the basic frequency response. Whether an untrained listener would be able to discern or appreciate those other differences though is probably debatable. That is where alot of the energy, $$, and research seems to be invested though in some higher cost cans.

Correct me if I'm wrong, but I seem to recall reading somewhere (possible in one of Tyll's excellent Inner Fidelity articles) that most headphones, including even relatively inexpensive ones, have measurably lower distortion than even the best in-room transducers. Which might be viewed either as a strength by some (ie the ones who like alot of clarity and detail), or as a weakness by some others, who prefer a more fuzzy speaker-like sound.

I like headphones with good detail, stereo imaging, bass and treble extension, speed and clarity though, as well as comfort and an accurate frequency response. Which is why I will consider all the above characteristics when buying new headphones. There are some lower cost headphones that have been making some pretty good strides in these other areas as well though, in addition to the improvements in the accuracy and extension of their frequency response.

 

[bigshot]

To me, response is the bulk of the difference between headphones. (Obviously comfort and usability matters too.) I doubt if I could hear differences in distortion, speed, decay, etc. In all the cans I've heard, that stuff is way too small to really matter. Bass extension is an issue though. There are cheap headphones with limited low bass, and that is clearly audible. They address that in the study. More expensive cans tend to have lower bass. That seems to be the one aspect where money does buy quality. But I'm guessing that there are mid range cans with good low bass that could be EQed to do anything one might want. No reason to spend a whole lot. Just shop wisely and figure out the response curve you are looking for.