A Crucial Question: Does flat response for speakers and harman curve for headphones tend to be individual’s favourite response?

[buz]

It's been a decade but if your estimate is unbiased, inference for how a specific individual might perceive something does not get better with higher sample size, no?

It gets better if you add more explanatory variables.

 

[Hipper]

I respect Toole and Olive’s research and I understand that these are preferred response in blind test, though I do think Olive’s sample size is too small (usually 200-300 people in his various papers and only 50-60 in each strata like region, age etc) to safely establish a solid standard/target.
I personally like flat speakers more than coloured speakers (though not the case for headphones, I always feel harman curve having too much bass, I don’t need to eq my HD800s’s bass at all.)
But Are these preferred frequency response necessarily close to individual’s actual favourite response? I mean is it possible that an individual prefers flat and harman but at the same time prefer some different response more?
If that’s the case, wasn’t visiting shops and having as many demo as possible a more effective way than spending time reading measurements and analysing deviation from target(flat/harman) to find your favourite speakers and headphones, especially for non-professionals?
Then it circles back to just listen to the device...

I wouldn't obsess about these or any other curves. It is you that should enjoy what you hear. Use these curves as a guideline perhaps but tinker with them to your liking.

The best place to demo a speaker is using your amp and in your room.

Measurements offered on here are also guidelines that lead you to, perhaps, a selection of speakers you might consider. Speaker measurement may have the air of an exact science but you should enjoy a speaker because you like it, not because someone tells you you should like it or even that they like it.

 

[dasdoing]

True. It is enough to draw meaningful conclusions.
But the statistical inference from the small sample to the general public's preference is weak.
In other words, the conclusion itself is valuable and meaningful, but its power to apply/predict individual's preference is so weak.
Their research is more like small sample size social science instead of more reliable, repeatable, predictable physical science.

another problem is that it's their preference in that specific room, at that specific distance

 

[Sancus]

I always thought mixers and recording engineers used something like the Genelec 8341A SAM™ for near-field listening, aren't the high frequencies like playing Russian rolette with headphones?

Yeah, mixing on headphones is controversial I think because it often doesn't translate well to anything other than other headphones. Checking mixes on headphones is of course normal, but I doubt many serious professionals are using them as their primary tool during the process.

Given that some of the most common studio headphones like the Sony MDR-7506 do generally follow the target though, it's possible that many engineers agree with it.

 

[LightninBoy]

If that’s the case, wasn’t visiting shops and having as many demo as possible a more effective way than spending time reading measurements and analysing deviation from target(flat/harman) to find your favourite speakers and headphones, especially for non-professionals?

I'll address only the speaker half of this question. IMO, showroom demos are not more effective because listening to speakers in a showroom is not the same as listening to speakers in your room. Also you are limiting your speaker options to what is available in your local showrooms which is a tiny fraction of the overall market. Finally, and hopefully temporarily, we've got things like the pandemic to consider.

Buy based on measurements, then tweak to room/taste using EQ is the most effective approach, IMO.

 

[jss]

At the end of the day, freq and freq responses is just one of the many aspects of sound and sound waves that go into our ears. There are also the reproduction of consonance and dissonance (like in rock and heavy metal), amplitude (not just it relates to dB/power/density of sound perceived but placing a sound at "right" amplitudes in a reproduction, or collection of sounds - aka contrast), timbre (how are the harmonics and partials reproduced, giving you a truer sense of realism or not?), and quite a few other attributes as well (e.g. sound wave interference - like binaural beats; and smearing, rhythm, and masking; etc.) How one's gears and headphones deal with all these collectively "defines" a gear's profile and characteristics as all music are played out integrally and dynamically over time.

 

[Chromatischism]

At the end of the day, freq and freq responses is just one of the many aspects of sound and sound waves that go into our ears. There are also the reproduction of consonance and dissonance (like in rock and heavy metal)

What are these and how do they not show in the frequency response?

amplitude (not just it relates to dB/power/density of sound perceived but placing a sound at "right" amplitudes in a reproduction, or collection of sounds - aka contrast)

This is literally frequency response.

timbre (how are the harmonics and partials reproduced, giving you a truer sense of realism or not?)

This is also frequency response.

and quite a few other attributes as well (e.g. sound wave interference - like binaural beats; and smearing, rhythm, and masking; etc.)

Do you have references to literature where these terms are defined and quantified?

 

[jss]

My point is simply that there are aspects of a sound that a FR curve won't capture/reflect ..

 

[Chromatischism]

My point is simply that there are aspects of a sound that a FR curve won't capture/reflect ..

I realize that is what you posted, but so much of what goes into modern measurements has been studied, with papers and books written. I'll need to see some published sources that define and quantify those things otherwise it's like hearing common audiophile tropes like "pace, rythm, and timing".

 

[Geert]

My point is simply that there are aspects of a sound that a FR curve won't capture/reflect ..

Yes, but not the ones you mentioned. You're confusing how sound is produced with how it's being reproduced via a speaker. The reproduction part is defined by a few technical parameters, frequency response being one the most important (other being directivity, distortion and maybe phase).

 

[ZolaIII]

I don't honour Harman or any other curve for that matter. I mean why should I other than bass response they are pretty much the same thing and bass whose corrected in Harman hedaphones one for at least couple of times till now. And neither does Samsung who own them and their IP (check 5 different "prefence curves" implementeded in their recent TWS for example).
Main thing to understand is that response will change with loudness level and none of so called curves include compensation for it. So for instance most hedaphones that blindly follow Harman curve at 72 dB will have raised bass response for +2 dB when you crunk it up to 84 dB and more so on as you rise it future more and I am keeping this discussion on hedaphones as what you get from speakers under 100 Hz will be deaply influenced by the room.
The moment bass overlaps with mids and up to 5 KHz range you will lose a lot of information as that's the range where 70% of information is. So every clame how any of the "curves" is the one and final doesn't hold water and can be discarded as genuine snake oil.
Personal preference will be subjectively influenced with whatever and the state of your hearing depending of deficits from aging or illness like Tinnitus and while at least partially you can compensate for some for others you can't. Luckily this is being added lately in a vogue averaged form recently in many player's and even streaming services (where you choose compensation for your age). But more work regarding this neads to be done by demographic, gender and even race (in the cultural meaning) in order to have a better and more precise compensation but even so it won't be individually fit for everyone ever.
So my advice is before baying equipment test your hearing and not blindly but medically (audiogram) and pick the one that best can scope with such and your personal preference preferably still having some headroom for compensation in those areas. Would be very interesting to do a research between preference and cultural influences on such. When EQ-ing take in consideration desired loudness listening level and include equal loudness normalization for such. This is easy thing to do with speakers as most of amplifier still included equal loudness normalization controls in the form of bass and treble control's with a 750 Hz cross (if implemented correctly of course), sadly with hedaphones this is usually unavailable (it's available if you use it trogh power amplifier hedaphone out but as those are usually lo quality one's almost on one does and besides one implemented for speakers dose differ a bit from one for hedaphones [ISO 226]).
That will put the end to many such discussions in the future (or at least I hope so) which I think is a good thing.
At least that's my humble opinion to where future studies should be pointed at and what needs to be adresed better in the future.

 

[Geert]

I don't honour Harman or any other curve for that matter. I mean why should I other than bass response they are pretty much the same thing

Which confirms they are correct. And why wouldn't they as they are based on a flat anechoic response, which means neutral sounding.

Main thing to understand is that response will change with loudness level and none of so called curves include compensation for it.

It's normal for a frequency response preference curve to show only two variables, frequency and prefered level. If you want to see curves for different loudness levels you introduce a third variable. You can make such a curve yourself by combining the preference curve with an equal loudness contour curve.

So for instance most hedaphones that blindly follow Harman curve at 72 dB will have raised bass response for +2 dB when you crunk it up to 84 dB and more so on as you rise it

Changing loudness doesn't invalidate frequency response preference curves. As explained above your mixing up different variables.

And how would a headphone compensate for different loudness levels? Should they make a different subtype for each level?

So my advice is before baying equipment test your hearing and not blindly but medically (audiogram) and pick the one that best can scope with such and your personal preference preferably still having some headroom for compensation in those areas.

Do you realy think people can find headphones with a frequency response optimized for their personal hearing deficiencies, and fulfilling other requirements they might have (like budget)? I think you'll have more success compensating for hearing deficiencies via EQ. Especially since these deficiencies can vary between left and right.

This is easy thing to do with speakers as most of amplifier still included equal loudness normalization controls

Actually, most amplifiers don't. Most high end amplifiers don't have tone controls. Nowadays software EQ-ing is the way to go, and that's a good thing.

 

[ZolaIII]

@Geert not to quit qout chop you (as I hate such).

"It's normal for a frequency response preference curve to show only two variables, frequency and prefered level. If you want to see curves for different loudness levels you introduce a third variable. You can make such a curve yourself by combining the preference curve with an equal loudness contour curve." Which no one does so far and as stated should be included. Thing is you won't listen at even remotely the same level as it will be both subjectively and objectively influenced by many things like; environment noise, desired lv, mood... so yes different levels of competition on different loudness level and included in EQ equations and if possible for particular analog transponder.
I didn't think that you would find an exact match to your hearing state but how buying one's with lo distortion in deficit areas (while being avare of such) and there for more headroom to add more there (to a reasonable extent) with EQ, and of course you can do it per channel.
Most amplifiers from most representative brands actually did in the past 50 years or so and such still do. Of course this day's and age you will do it in a digital domain and everything else of course.

If you (any of you) have a better idea I am eager to hear it!

 

[Geert]

Which no one does so far and as stated should be included.

Again, a graph that shows the relationship between 2 variables is the most basic visualisation of a specific effect being researched, in this case frequency curve preference. You can't say it's no good, it's actually essential. If you want to derive other graphs from it to include a third variable you can. Only when you're the CEO of Harman you can demand such a graph. Others should be happy Harman makes this research publicly available.

And if you still question Harman's research, this is their latest study where they had 130 people test 31 headphones https://www.docdroid.net/enYZsTS/a-...-around-ear-and-on-ear-headphones-pdf#page=10

The study again confirmed: "The headphone equalized to the new Harman OE/AE target curve was preferred to 28
of the 31 models tested when combining results of both trained and untrained listeners. Four other models were equally preferred to thet arget headphone".

And "A statistical model based on these deviations can predict listeners’ preference ratings with about 86% accuracy with 6.7% error". This while some people want to make us believe Harman's research is only good for the minority of people.

 

[ZolaIII]

@Geert please stop that doesn't even qualify as scientific research and even if it did it's far from conclusive one. I understand scientific methodology (methodology and logic in general) a little better than most (thanks to my education). And that whosent the point even to start with. 130 test subjects from which half received training or were employed by them with out proper hearing check of course and conducted by them and not independently. If you don't see bias there you probably never will. One hedaphones on which they simulated others while some physical characteristics can not be faithfully simulated (like impulses response of hedaphones xwz that's 3 ms on hedaphones used for sim where its longer).
By the way this day's you need to be a CEO at Samsung as they own them.

Have a nice time.

 

[Geert]

@Geert please stop that doesn't even qualify as scientific research and even if it did it's far from conclusive one.

Then maybe point us to better research on the topic. And what's not conclusive is mentioned in the paper.

130 test subjects from which half received training or were employed by them with out proper hearing check of course and conducted by them and not independently. If you don't see bias there you probably never will.

The 28 trained listeners were tested for normal hearing. They didn't show substantial different preferences than the other people. Given the 86% accuracy of the model being tested you can also conclude there probably weren't that many people involved with serious hearing deficiencies.

Next to that the test was a blind test, so they didn't know when they were listening to the Harman curve. So what bias?

One hedaphones on which they simulated others while some physical characteristics can not be faithfully simulate

Again, the principle of changing only 1 variable when testing. If you understand scientific methodology you should know.

 

[jae]

I've found that I also prefer a couple of dBs less bass than Harman on headphones. So somewhere between Diffuse Field/Etymotic and Harman. The 'bass boost' is an average, so if you feel like it's too much don't force it. View attachment 106172

Not only amplitude but the bandwidth/q is significant. Psychoacoustically, boosted sub/lower bass will mask other bass and mid frequencies less than an equivalent boost let's say beyond 120-150hz would mask elsewhere. Boosts past this point are almost universally considered bad or deleterious to accurate timbre, usually result in "muddy" sound is the term that's always thrown around, and in contrast you can get away with much more in lower frequencies lower than that before things like intelligibility become a major problem. This will obviously present itself more obviously in some genres, for example in a tune with vocals it will be more obvious to almost everyone.

The fact that harman iem target requires more gain in bass than their oe target leads me to suppose that the lack of low frequencies interacting with our body/pinna/room (which still doesn't explain the quantity of gain- bone conduction?) or perhaps the psychoacoustic effect of partial/full occlusion has an effect on how these frequencies are processed. I would like to assume that completely flat has to be "objectively" wrong as far as timbre accuracy in lieu of reflections, and maybe even as far as general preference is concerned for sounds with incidence close to the ear/different angle than is expected normally. As far as I know variance in low frequency transfer functions between individuals is quite low relative to other bands, so individual physiology explains that even less like it does with higher frequencies. There could be a lot of mechanisms, maybe since internal/infrasound may be more perceptible with occluded ears, we need the higher gain to mask or interpret these frequencies better?

 

[Tom C]

So, the way I’m reading the OP is, maybe Harman sounds good to most people, but could it be that some individuals may prefer another FR curve over the Hartman curve. Well, of course, the answer to that is yes. I once had a friend who’s favorite playback system was in his car, which had tone controls (this was 1978), and he would turn the bass all the way down to zero, and turns the treble all the way up to max. I didn’t like the way it sounded, and never met anyone else that did anything close to this extreme, but Walter really enjoyed it…
Let’s talk 2022. Brick and mortar has all but vanished. The last 7 speaker purchases I have made were for units I never heard before buying. When I was a rank beginner, I didn’t know what I wanted to listen for anyway. Without some measurements and expert, unbiased guidance (unavailable inside a dealer’s shop), a buyer is a ship at high seas without a rudder. Even in a major, world class city it is going to be hard to find a dealer who has everything you might want to hear, so you’re going to end up visiting multiple dealers, and you’re still going to only hear what those dealers feel they can make the most money selling. If you have measurements, you can at least have a reliable yardstick for judging the quality of the product. And if it’s a quality product, you can use EQ to hit any target frequency response you like. Even Walter’s.

 

[dfuller]

The Harman OE curve is a preference curve - it's got way more bottom end than a neutral response would. Don't consider it accurate in the bottom end - that's "fun" low end.

 

[Sokel]

What are these and how do they not show in the frequency response?

That's musical terms mostly (an a subject of controversy,but that's an other talk).

You can also use it by ear to tune strings, against another string,the more you're closing the frequency the slower is the repetition of the "pulses" (forgive my lack of engineering terminology) until it gets perfectly in sync.


In speakers and headphones probably means that the right and left plays close but not absolutely close so the phenomenon occurs.

(but I can't understand how,a sine wave is a sine wave,it may be different in amplitude but the freq should be the same.because,that's the difference that make this happen,a slightly different freq)