Not sure I like Harman curve

[Compact_D]

Could anybody kindly explain me what it is all about, really?

In my limited understanding, music should sound like it sounds on studio monitors in the studio. I am listening all my music on a pair of near-field studio monitors with a room-equalized subwoofer, and that is absolutely the best sound I can get for the available budget and given listening room. I have compared, and my setup sounds satisfactorily similar to that what *I* hear in the real studio. I could be completely mistaken, but I do not need better sound, except maybe closer to that in the studio. Maybe closer to the sound in a "better" studio.

Headphones should also sound as close to studio monitors as possible. Are there even any other options?
I could imagine to compare not to the studio monitors but to the live performance, but somehow I think that is a job of the mastering engineer, not my job as a listener.

So, why does the discussion of Harman curve always mentions "Preference"? I understand that there should be a "curve" but that should be a purely technical curve for headphone measurements to match the measurements of the speakers.
And, there should be no even any mention of this curve in any discussion! Same as if your measurement microphone would have a roll-off say at the low frequencies, one would apply the appropriate correction curve, but not discuss it or even mention it!

Am I seriously misunderstanding something? Thanks.

 

[Chromatischism]

Am I seriously misunderstanding something? Thanks.

Not all speakers or headphones sound the same, so the target curve brings them in line (to the extent possible) with the majority of the bell curve, who, prefer uncolored sound.

For speakers that means matching to a target that assumes a typical room.

For headphones that means taking into account the shape of our ears.

I don't know why we sometimes get bad results trying to EQ our way there. It could be limitations of the software, hardware, or both.

 

[solderdude]

So, why does the discussion of Harman curve always mentions "Preference"?

Because humans have preferences ?
Yours (as well as mine) is nearfield monitor sound but others may have different preferences, as shown by Harman research.

And, there should be no even any mention of this curve in any discussion!

Because there are different target curves it is important to know (and thus mention and perhaps also show) what target has been used.
Otherwise the generated plots are worthless.

 

[Chromatischism]

Because humans have preferences ?
Yours (as well as mine) is nearfield monitor sound but others may have different preferences, as shown by Harman research.

Preference is a funny thing. There really shouldn't be anything but the music that was created by the designer.

Preference is like saying I like my pictures with a color filter on them. Ok, then find a photographer who uses a lot of filters, because maybe this artist isn't your style.

 

[solderdude]

The world would be a boring place if there were no preferences.

 

[markanini]

Lets imagine that the Harman curve matched closely to some unanimously agreed on godawful sounding headphones.

Now let's come back to reality and it turns out Harman is a good match to a number of well regarded headphones. That should tell you enough about validity, at least over previous published curves.

So where does the motivation to criticize the research behind it come from? Logically if there was a problem with the Harman target it would be clearly articulated by now and the discussion would be about HOW to solve that hypothetical issue.

For example the people behind the USound target were able to describe an issue and, what causes it, and offer a solution:

Additional research performed by USound and IEM showed that in the presence of background noise (such as distant traffic), listeners tend to prefer more bass and slightly less treble compared to the original Harman Target

In this discussion is people are attempting to attack the Harman target from various angles, often exposing their lack of knowledge and sometimes megalomania. The topic will keep coming up probably but that's why nothing substantial will come out of it.

 

[Compact_D]

I thought that a proper "target curve" was already agreed, and it is that of a flat frequency response in an anechoic chamber. So all we need is to determine a frequency response for a headphone measurement that will produce the same sound for a headphone. No "preference" involved.
Should be not a "good speaker in a room" or other marketing bs, but exactly a studio monitor in an anechoic chamber.

If there is any additional "equalization" required due to specifics of human hearing, that should be done at the mastering process. Alternatively, most playback devices have tone controls for individual preferences.

P.S. Re attacking the Harman target - I have read most of what I could find regarding this "research", but what I see is not resembling a proper scientific research too much.

 

[solderdude]

I thought that a proper "target curve" was already agreed,

Not really. Harman proposes several target curves depending on OE, speaker, IEM.
Others have proposed other target curves so there is not a single one that is agreed upon (at least not for headphones).
Amir uses Harman target.

So all we need is to determine a frequency response for a headphone measurement that will produce the same sound for a headphone.

If that were a simple task then why hasn't this be done already ?
Ah... isn't that what Harman research set out to do (to produce a target the majority of people would prefer).

When you do not see the Harman research as proper research you are makining it rather difficult for yourself.
Which other target do you feel is based on proper research and better than Harman and why ?

 

[Compact_D]

Not really. Harman proposes several target curves depending on OE, speaker, IEM.
Others have proposed other target curves so there is not a single one that is agreed upon.

If properly done, there should be a direct link between the studio and listening room. That link is a flat studio monitor in a studio environment.
Target curve for a speaker should account for expected listening room, for headphone it should account for anatomy of ear, all at the end producing the same reference sound, ore as close to it as possible.

Problem with Harman curve is that it doesn't even claim to achieve this, instead involving a "preference". Preference of untrained listener turns everything into joke. Let that untrained public listen to studio monitors for a period of time, and their "preference" will change significantly.

 

[solderdude]

If properly done, there should be a direct link between the studio and listening room.

Why and what is properly ? A studio is not the same as an anechoic room and the average listening room does not exist.
For speakers in a room there are several factors that play major roles in what we can hear and what we can measure (also using different techniques)

That link is a flat studio monitor in a studio environment.

What studio, what monitor, in what positioning etc... too many variables right there.

Target curve for a speaker should account for expected listening room,

A 'standard' listening room is ....

for headphone it should account for anatomy of ear,

If it were only that simple... but it isn't.

Problem with Harman curve is that it doesn't even claim to achieve this, instead involving a "preference". Preference of untrained listener turns everything into joke. Let that untrained public listen to studio monitors for a period of time, and their "preference" will change significantly.

Harman research is what it is. Do you prefer other research ? What would you feel is proper research ? What target curve(s) came from that ?

 

[PeteL]

If properly done, there should be a direct link between the studio and listening room. That link is a flat studio monitor in a studio environment.
Target curve for a speaker should account for expected listening room, for headphone it should account for anatomy of ear, all at the end producing the same reference sound, ore as close to it as possible.

Problem with Harman curve is that it doesn't even claim to achieve this, instead involving a "preference". Preference of untrained listener turns everything into joke. Let that untrained public listen to studio monitors for a period of time, and their "preference" will change significantly.

The thing to keep in mind tough, the mastering engineers themselves go trough all sorts of compromises. Their job is not only to deliver the most amazing production on their "perfect" listening environment. They have to make sure the music they put out will be listenable in laptop and Iphones speakers, on headphones and earbuds, in cars, post the extra compression of radio stations, etc. What you are getting is not even what is their own preference if there was none of those limitations, their job is for it to work with anything, if they were specifically mastering and mixing for your own listening room, for you personally, they may have had put out something completely different.

 

[someguyontheinternet]

Problem with Harman curve is that it doesn't even claim to achieve this, instead involving a "preference". Preference of untrained listener turns everything into joke. Let that untrained public listen to studio monitors for a period of time, and their "preference" will change significantly.

This is doesn't appear to be true, because Harman research found no correlation between listener experience and preference:

 

[Compact_D]

Thanks, nice video perfectly showing how flawed their concept really is.

Only thing that "scientists" forgot to do is - apply their "target curve" to the studio equipment after mastering engineer has just finished his job and see what their employees prefer. Clearly missed opportunity to instantly improve the work of a mastering engineer by whole 5 points, LOL.

But if seriously, with headphones there was a real chance to listen to the music "as mastering engineer intended" (which is not easily possible with speakers for many reasons) but that chance is missed by "scientinsts" trying to sell more headphones. Anybody really does not understand what the real goal of the whole research was?

P.S. I am not saying their target is wrong, it may as well be perfectly correct, even it does not appear so according to what I hear. I will still eq my headphones by matching them to my monitors which *to me* sound sufficiently similar to those in the studio.

 

[Jimbob54]

I will still eq my headphones by matching them to my monitors which *to me* sound sufficiently similar to those in the studio.

I wasnt aware anyone was suggesting you couldn't.

But unless you can draw that FR as it applies to headphones as measured on a particular rig its of no use to anyone else.

 

[tiramisu]

I like the Harman curve, I prefer the Diffuse field curve for monitoring. I would prefer that our baseline measurements were against diffuse field because setting average user preference as the baseline of anything to me seems subjective and is a move in the wrong direction.

 

[solderdude]

Anybody really does not understand what the real goal of the whole research was?

The goal of the Harman research was to find out what 'sound' the vast majority of people preferred so they could 'tune' headphones in that direction and create more sales.

If indeed all sorts of listeners, in general, have a preference of a certain tonal balance/target using many different recordings then that preference can not be much off.
Of course the research also shows that not all people prefer that target. You seem to be one of them. This is not weird and makes neither you nor the research 'wrong'.
You just don't prefer the bass boost 'the majority of people' in general prefer.

 

[James-F]

One thing though, the Harman curve is ever evolving and is tied to the flavor of the week of their control groups.
They may create a better ear model or a test group of people that lean towards certain sound, and suddenly we have a new "Harman 2023" guideline.

I'm not saying that the current Harman curve is not close, judging by K371 vs calibrated monitors it's pretty darn close but lacks some very high 'air' (10kHz+) frequencies compared to near field monitors imo.
I do wish Harman created another model of a professional studio room with near field monitors 4 feet apart at an equilateral triangle, from what I see in the presentation their current listening room is a big home theater, not a studio.

A home theater as their reference listening room..... anyone mixes and masters their music in one?

 

[markanini]

View attachment 215987

Look at amount of diffusion panels. Most diffuse field target proponents are ignorant about what a diffuser looks like and what it's function is acoustically so here's to owning your self. Harmans reference room literally has more diffusion panels than the average mixing room.

 

[Compact_D]

This is not weird and makes neither you nor the research 'wrong'.

You are of course 100% correct in what you mean.

But - I do not even care if I like that target or not, and there is no point discussing that. Important is how close headphones sound to the reference speaker, not if I or anybody else likes it or not. Only speaker that can be reasonably called reference IMO is a studio monitor, better near-field, especially if we are talking about headphones. Also, listening should be done by experts, using properly selected test tracks.

For some reason, Harman does not want to reveal this information, even I am sure they have the proper non-marketing target. It is of course also a bit weird to talk about "science" then come up with this - it is "what seems to satisfy most tastes", but may be it's just me.

 

[solderdude]

The Harman curve is based on how well measuring speakers in a somewhat conditioned room sounds like.
Not on how near field monitors sound in a studio.
Productions from studios are produced in a way that it sounds good in home conditions. At least the well made recordings.
Pop music is produced so it sounds good on a phone, at home, in a car, on boomboxes. It is a compromise.