A proof of why Harman curve (or any "bass shelf") is bad

[castleofargh]

The idea that we should all EQ our headphones to the exact same generic curve

It's good then that nobody ever said that.

 

[bigshot]

The idea is to have a baseline calibration that acts as an anchor for your search for your own ideal curve. You can just randomly change things to the point where you have corrections negating other corrections and you have no idea where your goal is. It’s like navigating on the ocean without the stars to guide you. Having a trail of breadcrumbs leading back to the calibration keeps you focused and efficient. If something doesn’t work, or you lose your way and too much of the spectrum is scrambled, you can always reel it back to calibrated and take the steps that work again and find out where you went astray.

If you’ve never Eq’ed you might not realize the importance of this. It’s very easy to get confused and lose track of where you are. How effective EQ is at improving your sound is directly related to how organized you are in your search for the ideal curve. Randomly trying settings gives random results.

 

[2leftears]

The idea is to have a baseline calibration that acts as an anchor for your search for your own ideal curve. You can just randomly change things to the point where you have corrections negating other corrections and you have no idea where your goal is. It’s like navigating on the ocean without the stars to guide you. Having a trail of breadcrumbs leading back to the calibration keeps you focused and efficient. If something doesn’t work, or you lose your way and too much of the spectrum is scrambled, you can always reel it back to calibrated and take the steps that work again and find out where you went astray.

If you’ve never Eq’ed you might not realize the importance of this. It’s very easy to get confused and lose track of where you are. How effective EQ is at improving your sound is directly related to how organized you are in your search for the ideal curve. Randomly trying settings gives random results.

At the risk of sounding like Bernard from "Yes Minister": you can't leave a trail of breadcrumbs in the ocean. Because the wind, the currents, the seagulls, the fish...

My apologies @bigshot; couldn't help myself...

For reference to those unfamiliar with Bernard from "Yes Minister": 1:30 in the video:
​

 

[bigshot]

How about chumming bloody fish for the sharks to follow?

 

[gregorio]

The idea that we should all EQ our headphones to the exact same generic curve because a statistically insignificant group of non-diverse people were used to create an average preference profile is no different than saying we should EQ our cheeseburgers to a certain amount of cheese based on the preferences of a non-diverse group statistically insignificant group.

Great example of a strawman argument but why would you think a fallacious argument would be effective in this subforum? Do you think a science discussion forum would be more or less able to spot or accept a fallacious argument?

The only thing the Harmon studies tells us is what the people involved in the study prefer. Does that mean anything to you? It certainly doesn't mean anything at all to me.

Yes, it does mean something to me, it tells me what the subjects in the study preferred on average, under controlled conditions. That’s a whole lot more than we can get anywhere else, which is typically nothing more than a bunch of opinionated audiophiles under uncontrolled conditions providing incentivised impressions (that are virtually always contradictory in at least some regards)!

So why in the world do so many of you want to do that with your headphones?

You think that just re-phasing the same strawman argument will make it more effective in this subforum?

G

 

[goodvibes]

None will be spot on for everyone but low bass and upper mid gain is required for a neutral-ish balance for earphones. I suspect that much of the rising towards the bottom bass increase is due to lack of low frequency directionality in free air (long wavelengths) increasing the overall mount for summed channel ears (acoustic crosstalk). The mids boost and highs rolling off likely has to do with ear gain and high frequency being less able to carry outside of direct exposure. Earphones and IEMs are presenting those highs in a more direct way to the ear canal so likely sound more correct when frequency compensated with a gentle roll.

These trial and error tests are done with people listening to free air vs earphones and not to preference. Is this frequency louder or quieter sort of thing? etc I think they can do this to a fault so none will be perfect but there a reason for and overall consensus within a few DB. It's enough to accommodate whatever difference individuals may have in ear gain in HPs/earphones vs free air. It's why you see some minor differences in preference and reference curves but you rarely see them deviate by more than a couple DB from the lower mids to lower treble. There's so little info above 10k that you'll see odd curves without much correlation and bass will be a preference things to many that don't use acoustic live as a reference but given the same correct circumstances the curves for most any individual would look a lot like the Fletcher Munson commonly used.

Doesn't mean you can't prefer something else but it came from good tech and proper testing.

 

[bagwell359]

Yes, it does mean something to me, it tells me what the subjects in the study preferred on average, under controlled conditions. That’s a whole lot more than we can get anywhere else, which is typically nothing more than a bunch of opinionated audiophiles under uncontrolled conditions providing incentivised impressions (that are virtually always contradictory in at least some regards)!


G

Talk about strawman. Why must the other set compared to the people that were involved in the Harman data set be audiophiles (self proclaimed)? My prior point was I'd trust a subset of trained musicians that played primarily or only acoustic instruments over any other sub group or wad of people. Democracy is a wonderful thing - but I wouldn't want people that think Cold Duck and Boones Farm is good stuff opining on fine wines. Why would I want people with little or no experience with live unamplified music and say death metal and EDM fans deciding what is correct?

Also if this is a proper test (double blind) why would these audiophiles you have conjured up be influenced by appearances, brand names, etc? Seems like a bias has un-moored you from objectivity. Its double blind for all, and yes I would expect different results from a subset of trained musicians than the current Harman. Would that be a non optimum result?

 

[bigshot]

I doubt if you’d trust any study that disagrees with your preconceived conclusion. It’s funny to nitpick a carefully conducted study when you won’t even apply the most basic controls to your own listening tests. I don’t think you’re qualified to comment on how well tests are conducted.

Score: comparing audio to fine wines = minus ten points

 

[gregorio]

Talk about strawman. Why must the other set compared to the people that were involved in the Harman data set be audiophiles (self proclaimed)?

You state “Talk about strawman” and then in your very next sentence you argue a strawman, brilliant! Don’t you even know what a strawman argument is, or that you’re using one? I did not state “the other set compared” must be audiophiles, you just made that up false argument, hence a strawman! I actually stated in effect that “the other set compared” is “typically” audiophiles, because audiophiles are a subset who most commonly opine on headphone preferences.

My prior point was I'd trust a subset of trained musicians that played primarily or only acoustic instruments over any other sub group or wad of people.

That is so ridiculously wrong on so many levels!

1. What prior point? You haven’t made a prior point in this thread, as far as I can see this is your first post to this thread. I therefore obviously wasn’t responding to you, I was responding to pblogic007’s question: “Does the Harman study mean anything to you?”.

2. This is the Sound Science subforum and therefore what’s valued and relevant is objective evidence, not what “you’d trust”!

3. What objective study (under controlled conditions) is there of headphone frequency preference with a significant sample size comprising only of a “subset of trained musicians” as you defined? If there is not such a study, how on earth can I compare, trust or prefer it to the Harman study?

4. Although irrelevant, I personally would not especially trust that “subset of trained musicians” because their experience/reference of their acoustic instrument is what it sounds like from a few inches/feet away, not what it sounds like either in the audience of a live performance or necessarily on any particular recording.

5. I’m a highly trained musician who only played acoustic instruments, I studied at a top music conservertoire for several years and for quite a few years subsequently I was a professional orchestral musician. So, I fulfil all your conditions for what “you’d trust” and yet you’re not trusting me, you’re doing the exact opposite and arguing with me. You’re contradicting yourself!

Also if this is a proper test (double blind) why would these audiophiles you have conjured up be influenced by appearances, brand names, etc? Seems like a bias has un-moored you from objectivity.

Firstly, do you really think I “conjured up” all these audiophiles, their impressions, reviews, opinions and preferences of HP frequency response on Head-Fi? You obviously must have reliable evidence for such an accusation? Secondly, what “proper test (double blind)” of audiophiles’ HP frequency response preferences? Give me such a proper test/study of a decent sample size and I’ll consider that too but if there isn’t one (or if I’m unaware of any), then how on earth can I objectively consider it? Seems like you’re “un-moored” from not only objectivity but also logic!

It’s double blind for all, and yes I would expect different results from a subset of trained musicians than the current Harman. Would that be a non optimum result?

Let me get this straight, you’re telling me that I shouldn’t take anything from a published/verified objective study but I should instead rely on your personal expectation? Specifically; the results that “I [you] would expect” from a hypothetical study that doesn’t exist (to my knowledge)? And yes, that is definitely “non optimum”, pretty much as “non optimum” as I can even imagine, in fact so non optimum it actually made me laugh!

G

 

[bigshot]

All musicians don’t have the same preferences any more than all regular people do. I don’t know why a tuba player in a German oom-pah-pah-pah band would necessarily automatically prefer the same response curve as a banjo player in a Dixieland jazz band.

 

[bagwell359]

You state “Talk about strawman” and then in your very next sentence you argue a strawman, brilliant! Don’t you even know what a strawman argument is, or that you’re using one? I did not state “the other set compared” must be audiophiles, you just made that up false argument, hence a strawman! I actually stated in effect that “the other set compared” is “typically” audiophiles, because audiophiles are a subset who most commonly opine on headphone preferences.

So sorry, you were either making a strawman argument or you were slandering "audiophiles". Which is it?

1. What prior point? You haven’t made a prior point in this thread, as far as I can see this is your first post to this thread. I therefore obviously wasn’t responding to you, I was responding to pblogic007’s question: “Does the Harman study mean anything to you?”.

It was on another thread - "Is it time for a new Harman survey?" You responded to it just over 2 weeks ago.

2. This is the Sound Science subforum and therefore what’s valued and relevant is objective evidence, not what “you’d trust”!

Please. You trust data, so do I. But since hearing is a subjective sense, its not all measurements, one must evaluate the output - is that how the Harman data is expressed by those following the criteria?

3. What objective study (under controlled conditions) is there of headphone frequency preference with a significant sample size comprising only of a “subset of trained musicians” as you defined?

I never said that there was one. I am suggesting such a study would yield different results from the current data set of Harman. By what logic do you suppose someone raised on recorded music via boom boxes/transistor radios/walkman/ipods would recognize a collection of well recorded acoustic music as well as musicians who work in that space?

If there is not such a study, how on earth can I compare, trust or prefer it to the Harman study?

Talk about concrete. I assume you understand the process of posing problems, before there is any data one way or another, and then to design a method to test this hypothesis. Just because it hasn't been done, it doesn't mean it shouldn't be done, and I posit that the results would be different than Harman 2018. None of that is remotely illogical OR argumentative.

4. Although irrelevant, I personally would not especially trust that “subset of trained musicians” because their experience/reference of their acoustic instrument is what it sounds like from a few inches/feet away, not what it sounds like either in the audience of a live performance or necessarily on any particular recording.

Perhaps, but such people whom I have known throughout my life listen to other performers a very great deal over time.

5. I’m a highly trained musician who only played acoustic instruments, I studied at a top music conservertoire for several years and for quite a few years subsequently I was a professional orchestral musician. So, I fulfil all your conditions for what “you’d trust” and yet you’re not trusting me, you’re doing the exact opposite and arguing with me. You’re contradicting yourself!

One persons opinion/experience isn't a proof. I'd like a larger set. I neither take or not take your opinion. I note it, anymore than that would be counter to fair inquiry in my opinion.

Firstly, do you really think I “conjured up” all these audiophiles, their impressions, reviews, opinions and preferences of HP frequency response on Head-Fi? You obviously must have reliable evidence for such an accusation?

When one spends so much time battling what they consider ignorance/stupidity, its a natural human trait to start to paint anyone suspected of holding some or all of these traits with the same brush. Denigration would be the word I would use, and there is no shortage of that to be found in this thread. Not to say that many audiophiles have made mistakes of all sorts, and acted with a similar level of denigration. Is there now a chance of addressing my hypothesis shed of these knee jerk reactions?

econdly, what “proper test (double blind)” of audiophiles’ HP frequency response preferences? Give me such a proper test/study of a decent sample size and I’ll consider that too but if there isn’t one (or if I’m unaware of any), then how on earth can I objectively consider it? Seems like you’re “un-moored” from not only objectivity but also logic!

You start off with a fair request, but then cannot help yourself yet again.

Let me get this straight, you’re telling me that I shouldn’t take anything from a published/verified objective study but I should instead rely on your personal expectation? Specifically; the results that “I [you] would expect” from a hypothetical study that doesn’t exist (to my knowledge)? And yes, that is definitely “non optimum”, pretty much as “non optimum” as I can even imagine, in fact so non optimum it actually made me laugh!

G

Take anything? Certainly the results are usable - I use them, and as I have stated a number of times on HF, I find the bass under 100 if not 120 to be too elevated, as well as the 2.5-4k area, so adjust for them when using PEQ. There are others who agree who came to that conclusion on their own. Interesting, but not of course conclusive of anything.

Again, cannot help yourself, what a treat you must be in person. Well, fire away any way that satisfies you. I'm now up to 4 people on ignore. Have a nice day.

 

[bigshot]

You’re not doing well, I’m afraid.

 

[Davesrose]

Well it seems this thread has de-evolved as most threads on Sound Science do. With the premise of the Harman target curves.....that's what they are: a target curve extracted from sampling preferences of various people. You can try the current headphone/speaker set to see if its your cup of tea, and you might either find it is or isn't your cup of tea. Another thread was asking if we need to have another revision of the Harman curve: well there have been a few revisions with very slight adjustments. There might be something with the given sampled listeners, that they favor acoustic music. If you listen to synth and are a bass head, then maybe there's not enough of a boost for you. No one is stopping you from EQing a curve that you will find most enjoyable: but also don't be surprised that other people will have their own preferences. As someone who is also involved with computer graphics and software development: I will sometimes find it funny when I run into people claiming there's just one ultimate computer OS for everyone. As someone who works with Mac, Windows, Linux...I find strengths and weaknesses with each of them. The world would be a better place if you find what you like, but realize you won't convert everyone to your same subjective tastes.

 

[gregorio]

So sorry, you were either making a strawman argument or you were slandering "audiophiles". Which is it?

So that’s a “yes” then, you don’t know what a strawman argument is!

It was on another thread

So no prior point in this thread then!

But since hearing is a subjective sense, it’s not all measurements, one must evaluate the output …

The Harman study isn’t a measurement, it was a controlled study of subjective preferences!

I never said that there was one. I am suggesting such a study would yield different results from the current data set of Harman.

Exactly my point. There isn’t such a study for me to consider, so I should instead consider what you’re suggesting the results of a hypothetical study (that doesn’t exist) would be!

I assume you understand the process of posing problems, before there is any data one way or another, and then to design a method to test this hypothesis.

I do understand that process. Don’t you understand that unless that process has been completed (or even started!) then there is nothing for me to consider, except your unsubstantiated suggestion/expectation?

Perhaps, but such people whom I have known throughout my life listen to other performers a very great deal over time.

Yes, we do but not nearly so much as we listen to ourselves from just a few inches away. Assuming that you mean “such people” to be professional musicians who’ve spent many hours a day for years practicing.

When one spends so much time battling what they consider ignorance/stupidity, it’s a natural human trait to start to paint anyone suspected of holding some or all of these traits with the same brush.

So that’s a “no” then, you have no reliable evidence to support your accusation of me “conjuring up these audiophiles” and instead you’re going to argue a different argument. Do you know what that’s called?

Again, cannot help yourself, what a treat you must be in person.

I cannot help myself? You seem to have forgotten that you were the one who responded to me with a bunch of BS and now you’re all butt-hurt because I have the nerve to respond and call out the BS you aimed at me. What a treat you must be in person, if you can make-up BS and expect everyone to just accept it!

G

 

[bigshot]

Well it seems this thread has de-evolved as most threads on Sound Science do.

It started out with a misconception about what the Harman Curve is and what its purpose is and went downhill from there. It started out as axe grinding, the misconceptions were corrected and ignored… over and over. That’s common for threads on Sound Science.