Variations in people's hearing responses

[SilverEars]

There's at times contradictory impressions of headphones. Also, we hear about people getting their hearing checked, and you'd get plotted out the level of sensitivity. I've also seen on this page( https://www.nuraphone.com/) that shows people having considerably different hearing responses. Is this representative of what is going on with majority of the people out there? Are there significant differences in people's hearing responses or do majority have vary similar responses?

Anybody have links to papers or sites showing statistical data of people's hearing response?

 

[bigshot]

People have all sorts of hearing degradation. But that doesn't mean some people hear better than normal optimal hearing too. The thing about hearing damage is that you hear the whole world with that same curve, so you acclimate to it. I'm not sure if coming up with a correction curve to compensate would necessarily improve perceived sound quality.

Audiologists just focus on the range of human speech. They don't measure the full spectrum of hearing usually. So a checkup with an audiologist will just tell you if your hearing is major league borked. It won't tell you everything that might impact listening to music.

 

[castleofargh]

audiograms are considered normal within 20dB of variations, despite how it is indeed a limited part of the audible range that is being tested. so yes hearing varies a lot from people to people.
to that you have to add lifestyle and age related hearing loss.
the fact that people are used to listen to music at different listening levels, so the equal loudness contour plays a part. anytime I participated in an audio meeting of sort, I was amazed by how loud some people listen to music when trying some headphone after them.
depending on the gears(IEM/headphone/speakers) there will be some extra variations coming from individual HRTF, were it's also typical to have audible variations from one person to the next. the solution implemented by Nuraphone is only an attempt by the headphone at getting the shape of the ear and compensate for it somehow.

and of course the star of the show is the guy describing his impressions of a product. how many will use a poor reference they consider "neutral" because that's what they used for the last years? so when they call something bassy, it's only because it has more bass than the stuff they're used to. how many will end up calling 1khz treble for lack of learning those stuff, or use some weird poetic lingo? and sometimes the way they fit a headphone or IEM will also result in serious variations in FR.

in the end, subjective impressions very much deserve to be called subjective.

 

[Zapp_Fan]

I saw the Nuraphone in person and it's a very interesting piece of equipment, as I understand it, it's mainly compensating for differences in shape of pinnae, the "hearing test" portion of the calibration, I think is relatively coarse. For what it's worth, the Nuraphone has different presets built-in (or at least they had some for the demo) and they did make a big change in the sound. So if that is representative of what things sound like behind differently-shaped ears, then yes, there is plenty of room for individual variation in sound quality.

If a headphone completely lacks bass, everyone will hear that. But how people experience mids and treble really does vary quite a bit. Whether our brains compensate for this (and thus bring the ultimate subjective experiences more in line with each other or not) I have no idea.

 

[amirm]

Are there significant differences in people's hearing responses or do majority have vary similar responses?

Anybody have links to papers or sites showing statistical data of people's hearing response?

Sure. This has been studied extensively by Harman when it comes to speaker evaluation. The answer is that we are remarkably similar to each other when it comes to such assessments.

Please see the peer reviewed Journal of AES paper, Differences in Performance and Preference of Trained versus Untrained Listeners in Loudspeaker Tests: A Case Study, by Dr. Sean Olive

The data and statistical analysis in the paper is quite extensive. But here is the summary data you are asking about:

The horizontal axis is the different group of listeners performing the double blind listening test of four different speakers. Horizontal axis is different listening groups. HAR for example is Harman's trained listeners. Reviewers is what it says: audio magazine reviewers.

The vertical axis is the preference score. As we see the scores for speakers within a group does not change from any other. For example speaker "M" is the lowest graph and is rates as such by every category of listeners.

There are variations on how harsh the scoring is done (e.g. Harman listeners are much more ruthless and give much lower scores to poor speakers), but overall ranking does not change.

In all, 268 listeners participated in the test. Here is the high-level breakdown:

In a follow up test conducted on younger listeners by Dr. Olive, similar outcome was seen:

Note that the only way to achieve these results is by hiding the identity of the loudspeakers. Once people see the product, preferences shift due to prejudices regarding brand, looks, design, etc.

So as a whole, we like similar characteristics in loudspeakers which bodes well for designings speakers that appeal to large group of people. It may seem non-intuitive at first but if you think about the fact that vast majority of us like chocolate, and steak, it is not that hard to imagine it being correct .

Now your question was about headphones. Again Dr. Olive (and Todd Welti) have conducted good research there. See AES paper The Relationship between Perception and Measurement of Headphone Sound Quality

This double blind listening test paper is also quite extensive. The results here are not quite as cut and dry as loudspeakers:

We see some agreement among listeners but also exceptions. Authors dig further and find that some listeners are just not consistent in their scoring. But also, rated headphones on other factors such as the fit.

Indeed that remains a problem with studying headphones as it is impossible to conduct a true blind test. Even though headphones were placed on the listener head by a proctor, listeners quickly learn which is which based on feel of headphones on their heads and that can and likely biases their scores. There is also the inconsistency of measurements (to correlate with subjective results).

Here is part of the conclusions of note:

#4 matches what the research says about loudspeakers. That smooth (not flat) frequency response is desirable.

Probably more than you wanted to know but I figured better be more complete than not.

 

[castleofargh]

this is about preferred sound, I don't think it's the same thing as differences in hearing.
for example, we spend our life experiencing sounds transformed by our HRTF, so somehow we get used to that and to sounds changed that way. it's IMO a big factor in why we can all enjoy a relatively flat speaker sound. because it matched our reference of sounds(instruments, voices...) the way we always perceive them. I have the idea that it's probably more a matter of getting matching pattern with our life long stock, than a sign that we actually hear the same way.

once we move on to headphones or IEMs, as we bypass or alter part of our very own HRTF, neutral is not neutral anymore. and different people can need different corrections to get back to the response they're used to. I love those Harman papers, but we must keep in mind that a good deal of those tests were done with limited options for correction. be it picking within only a few headphones(I think the one everybody liked was a LCD2?), and the ultimate curve was obtained starting with a fixed signature and 2 knobs, one for bass and one for treble. not exactly HRTF friendly. plus taste can be affected by habits and subjective stuff more than actual hearing. it's only logical that Harman would care about preference, because they want to sell something in the end so preference matters the most.

 

[bigshot]

People have various degrees of compromised hearing. No human hears better than the generally accepted limits of human perception. I read somewhere that the record for high frequency hearing was around 23kHz, which is barely a single note on the musical scale. The ability to hear extends downward, not upward.

 

[HotIce]

Agreed, hearing responses and hearing preferences are two separate things.
But could also be intermingled.
Do I like "more highs" because I simply like more highs, or because they compensate my hearing deficiency on the highs?
I'd tend to think the former, but it's hard to prove.

 

[amirm]

love those Harman papers, but we must keep in mind that a good deal of those tests were done with limited options for correction. be it picking within only a few headphones(I think the one everybody liked was a LCD2?), and the ultimate curve was obtained starting with a fixed signature and 2 knobs, one for bass and one for treble.

Harman research is quite extensive. Here is an example of measuring no less than 61 In-ear Monitors at all price points to predict their fidelity:

I also must admit I didn't understand the rest of your post. Isn't OP asking how people go about evaluating fidelity of headphones? Aren't they presenting their preference or lack thereof for another? None are describing how they hear. They are expressing what they hear. Too much treble, bass, etc. are key characteristics that impact preference for both loudspeakers and headphones.

 

[bigshot]

I haven't found any correlation between price and sound quality. Expensive transducers can sound great or sound awful. Cheap ones can sound awful or pretty good. The best headphones I've heard are expensive, but that doesn't mean that all expensive headphones sound good. I don't care for in ear monitors though. Maybe it's different there than with headphones or speakers. I can't read that chart though. The vertical axis has two different definitions. I'm guessing the horizontal axis is different headphones, but what is expected and what is actual? and are the green lines price or quality?

 

[SilverEars]

I'm interested in measured ear responses, not preferences. I'm aware Dr Olive does studies on listener preferences. I'm interested in finding out if people's ear response(their actual sensitivity, not preference) across the audible spectrum is more or less similar or not, and how much it can deviate, etc..

 

[amirm]

I'm interested in measured ear responses, not preferences. I'm aware Dr Olive does studies on listener preferences. I'm interested in finding out if people's ear response(their actual sensitivity, not preference) across the audible spectrum is more or less similar or not, and how much it can deviate, etc..

If by that you are including the inner ear/cochlea/organ of corti, as far as I know any research there has been performed by removing said organ from dead people and testing it for resonances and such. Kind of hard to probe those electrical signals in live people I imagine . For this reason we resort to subjective testing where the person gives us the feedback of what is or is not heard, not objective measurements of the organ itself.

Now if you just want the outer ear, then there are in-ear microphones to measure such but that is not remotely the full story.

We are also excluding the cortex/central nervous system which overrides/interprets what you hear.

Anyway these types of questions leave the domain of audio science and travel into medical science. For that reason, it is above my pay grade. So hopefully others chime in with data.

 

[sonitus mirus]

Out of curiosity, I did a bit of digging and stumbled across some reference material that is somewhat applicable.

I found it to be interesting.

How and Why Does Spatial-Hearing Ability Differ among Listeners? What Is the Role of Learning and Multisensory Interactions?