Not sure I like Harman curve

[jae]

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.

There is no way for harman to know your own personal physiology/HRTF to give you some kind of curve that is most accurate in timbre in these semireflective field conditions so I'm not sure what you want them to reveal. A variety of binaural amplitude and time-domain components of sound that affect timbre in a room (even nearfield) are effectively removed when using headphones. You can come extremely close to accurate timbre with FR compensation alone, but it will never be perfect which is why there will always be large amounts of variance because of physiology, listening experience, age etc. even if you ignore whatever "preference" component. The end result at your ear drum will also vary by a number of other practical things like the design of the headphone, uniform seal, if your canal is occluded or not, perhaps incidence of the soundfield, other factors like bone conduction etc. The various iterations of the Harman curve can essentially be viewed as an averaging over a relatively small sample size for those field conditions, with "psychoacoustic" and "preference" considerations mostly in the bass region. If you want what I assume is the objective "on paper" measurement averages for the harman target, I figure it will be more or less identical to Harman curves without the bass preference (so flat), maybe (+)/-2 dB or so in the ear gain region relative to the lower bands.

 

[Compact_D]

There is no way for harman to know your own personal physiology

That is exactly why they should not ask my preferences but compare what I hear while comparing with some reference.

there will always be large amounts of variance because of physiology, listening experience, age etc. even if you ignore whatever "preference" component.

I expect those will somewhat cancel out if instead compared two sounds. I assume that variations will be smaller if comparing two sounds instead of looking for preferences.

A variety of binaural amplitude and time-domain components of sound that affect timbre in a room (even nearfield) are effectively removed when using headphones.

That is *exactly why comparing headphones with "speaker in listening room" seems as unscientific as anything can be. That is exactly why headphones should be compared with near-fields, or if that is still not good enough, compare with speaker in anechoic chamber. But I think that near-fields are enough.

Regarding the "good listening room". Average music, like say chamber music recorded ambience will sound "good" on near-fields, presumably close to original. It *may* sound "better" in a good listening room because ambience of the listening room is somewhat added to the sound.
But take organ music in a large cathedral and the story is completely different. Record will still sound great on near-fields, but any listening room will inevitably make sound worse, unless listening room is a cathedral itself. This is unless the ambience of the "good listening room" is inaudible.

Regarding statements of Mr. Toole and others regarding how bad everything sounds in an anechoic chamber, I think this is more of a brain trick, where we are so used to hear reflections that in a sudden absence of those everything seems "weird" and we somehow lose orientation and hearing therefore is severely affected. I am sure that if we stay in the anechoic chamber for longer, things would certainly start to sound completely different. I am saying this based on my experience of having to work in a near absolute silence for a while (not completely anechoic, but close). It is quite unbelievable how perception of everything can change in a relatively short time, it feels like you have to re-learn the very basic senses.

 

[solderdude]

One can EQ nearfields to certain 'room curves'
Will all studios NOT EQ their monitors to certain standards ?
Do they just connect studio monitors, set all adjustments flat and hope for the best ?

 

[trevintheaudioboy]

Is harman target a summation of listener preferences? Or is it intended to give headphones a “speaker-like” sound? I’ve seen both these claims made. Either way, not everyone likes harman’s target. Personally that much bass is distracting and muddies the sound for me.

 

[Doodski]

Is harman target a summation of listener preferences? Or is it intended to give headphones a “speaker-like” sound? I’ve seen both these claims made. Either way, not everyone likes harman’s target. Personally that much bass is distracting and muddies the sound for me.

I'm the same way. I prefer a high boost for the upper frequencies all the way to the limits of my hearing range.

 

[solderdude]

Is harman target a summation of listener preferences? Or is it intended to give headphones a “speaker-like” sound?

It is an attempt to give headphones a speaker-in-a-room alike sound where the bass boost is largely based on listener preference at a specific listening level.
All of this in an attempt to create headphones (and speakers) that sell well to the largest possible market. Makes perfect sense from this p.o.v.
Personally I have gotten used to less bass boost for most well made recordings but like a Harman-type bass boost when going portable/outdoors or with some pop/rock recordings that sound a little 'thin'.

Thank the audio-god(s) for tone control that is quickly to adjust.

 

[BoredErica]

It is an attempt to give headphones a speaker-in-a-room alike sound where the bass boost is largely based on listener preference at a specific listening level.
All of this in an attempt to create headphones (and speakers) that sell well to the largest possible market. Makes perfect sense from this p.o.v.
Personally I have gotten used to less bass boost for most well made recordings but like a Harman-type bass boost when going portable/outdoors or with some pop/rock recordings that sound a little 'thin'.

Thank the audio-god(s) for tone control that is quickly to adjust.

Little OT but in terms of EQ, the main thing I'm missing is track-based EQ. I wanna play something in Foobar and for specific tracks to play a different EQ than whatever I'm normally using with EqualizerAPO.

Between normal EQ based on spinorama, volume based EQ, and EQ based on track, that's a heck of a lot of adjustment and makes my head swirl. xD

 

[Compact_D]

Personally I have gotten used to less bass boost for most well made recordings but like a Harman-type bass boost when going portable/outdoors or with some pop/rock recordings that sound a little 'thin'.

Sennheiser was justifying their bass boost of Momentums with exactly this, and it's a totally valid point. I wouldn't argue with Harman's bass boost also, if it was explained like this.

volume based EQ

'Volume based EQ' does not work for me at all, it just sounds wrong. My brain must be doing necessary EQ for me if that makes sense, like vision adapts to colour temperature changing during the day and we do not need to wear different colour glasses to compensate.

 

[Chromatischism]

'Volume based EQ' does not work for me at all, it just sounds wrong. My brain must be doing necessary EQ for me if that makes sense, like vision adapts to colour temperature changing during the day and we do not need to wear different colour glasses to compensate.

Maybe so. It's pretty obvious to me that bass and treble get quieter faster when I turn the volume down, and that bass can easily get overwhelming when I turn it up. The balance is just not there without adjusting the curve.

 

[V17]

I have question which, as far as I've seen, hasn't been raised in this thread.

Does anybody else have a problem with the 250 Hz dip?

From what I've seen in this post, it seems like the dip was created by taking the average HRTF (which is the basis for the Harman curve) and adding a low shelf filter with a few dB to compensate for a perceived lack of bass. However the corner frequency of the low shelf is relatively low, so there's now a dip between the bass and the classic headphone mid-treble hill.

For me the Harman target curve overall sounds great (though I usually add even more "tilt", gradually slightly reducing treble), but the 250 Hz dip is the one thing that simply sounds wrong to me. I have a comparison with speakers that I built and measured to be relatively smooth and the 250 Hz dip sounds similar as if I scooped the 250 Hz on those loudspeakers.

So the one fix that I do when EQing headphones (which otherwise works tremendously well) is adding about 3 dB with Q=1 at 250 Hz. This is unrelated to the overall preferred amount of bass.

I was really wondering if other people hear this issue because intuitively I'd say this is not caused by a difference between an average HRTF and my real HRTF and it seems to be caused by Harman deciding to put the low shelf too low (but why?).

 

[markanini]

I have question which, as far as I've seen, hasn't been raised in this thread.

Does anybody else have a problem with the 250 Hz dip?

From what I've seen in this post, it seems like the dip was created by taking the average HRTF (which is the basis for the Harman curve) and adding a low shelf filter with a few dB to compensate for a perceived lack of bass. However the corner frequency of the low shelf is relatively low, so there's now a dip between the bass and the classic headphone mid-treble hill.

For me the Harman target curve overall sounds great (though I usually add even more "tilt", gradually slightly reducing treble), but the 250 Hz dip is the one thing that simply sounds wrong to me. I have a comparison with speakers that I built and measured to be relatively smooth and the 250 Hz dip sounds similar as if I scooped the 250 Hz on those loudspeakers.

So the one fix that I do when EQing headphones (which otherwise works tremendously well) is adding about 3 dB with Q=1 at 250 Hz. This is unrelated to the overall preferred amount of bass.

I was really wondering if other people hear this issue because intuitively I'd say this is not caused by a difference between an average HRTF and my real HRTF and it seems to be caused by Harman deciding to put the low shelf too low (but why?).

Are you talking about Harman OE 2018?

 

[V17]

Yes, but it seems that the other versions have a variation of dip as well.

 

[markanini]

Yes, but it seems that the other versions have a variation of dip as well.

I might have some input on this but I'd rather be clear about which target is being discussed first. Certainly before discussing HRTFs. No offence, there's a lot of low effort critique of Harman target that muddles legitimate critique and I can't tell if you have an agenda.

 

[Peterinvan]

Harman for Golden Ears Only?

I am now 75 and my audiologist reports that my upper frequencies drop several DB starting at about 6KHz. Using SineGen.exe, I cannot hear anything beyond 10KHz.

I therefore find that brighter equipment suits me better. My listening enjoyment factor is still great.

When the experts (Harman) develop their "ideal" FR chart, what age group are they using, and what quality ears?

 

[V17]

I might have some input on this but I'd rather be clear about which target is being discussed first. Certainly before discussing HRTFs. No offence, there's a lot of low effort critique of Harman target that muddles legitimate critique and I can't tell if you have an agenda.

I do not have an agenda that I know of. I like the work Harman is doing because the increase in quality I've gotten from EQing my headphones has been great.

The important thing in my post is that the 250 Hz dip sounds wrong to me compared to flat-ish sound from loudspeakers and I'm trying to find if somebody else has that problem or why it exists. The rest is mostly speculation based on things I read here and there.

 

[sejarzo]

Harman for Golden Ears Only?

I am now 75 and my audiologist reports that my upper frequencies drop several DB starting at about 6KHz. Using SineGen.exe, I cannot hear anything beyond 10KHz.

I therefore find that brighter equipment suits me better. My listening enjoyment factor is still great.

When the experts (Harman) develop their "ideal" FR chart, what age group are they using, and what quality ears?

I wonder about that too, as I'm 65 with a scoop-out from 6-8kHz in my left ear and cannot hear above 12-13kHz using various sine wave generators and Ety ER3SEs, but find the Harman curves have way too much treble boost. <insert shrug emoji>

 

[markanini]

I do not have an agenda that I know of. I like the work Harman is doing because the increase in quality I've gotten from EQing my headphones has been great.

The important thing in my post is that the 250 Hz dip sounds wrong to me compared to flat-ish sound from loudspeakers and I'm trying to find if somebody else has that problem or why it exists. The rest is mostly speculation based on things I read here and there.

I don't hear at 250Hz in OE 2018. At 150Hz I hear a 2-3 dB suck out.

 

[Keith_W]

These are excellent questions gents, and for a while I have been looking for a copy of the original paper published by the AES in 2014 and the follow-up study in 2019. From what I know, they collected data from 238 participants of varying age, gender, and had a mixture of trained and untrained listeners in four countries. They were given a headphone with an approximation of the Harman curve and allowed to freely adjust bass and treble to their preference.

In particular, I would like to see the standard deviations and the 95% confidence interval as well as how the population of test subjects was collected. With no disrespect to either of you, I would hope that anybody with hearing loss or even partial hearing loss would be excluded from the study, because it would unfairly skew the results - so I would also be interested to see if hearing loss was taken into consideration and what cutoffs were used. I would also like to see if they repeated the experiment with a second or third headphone to see if the test subjects were internally consistent in their selection.

I suspect that a line graph with standard deviations like this one shown below (this is a random image found on Google) would be a better representation of the study data rather than the line graph alone typically shown in all publications including Amir's graphs:

I also think that rigid adherence to the line curve is too dogmatic given that there are likely standard deviations in preference, as well as standard deviations in the measurement setup. For example, the standard deviations are likely wider at extremes of the curve (i.e. bass and treble) for two reasons - human hearing is less sensitive to larger deviations in these frequencies, and measurement setups also show larger variations. Of course this is speculation since I have not actually seen the standard deviations. A curve that lies within the standard deviations collected in the study would be more scientific.

IF they excluded individuals with hearing loss as part of their study (as they hopefully should have) then your target individual target curve should lie outside Harman. I would suggest you tune your EQ to your preference and just leave it at that.

 

[Chromatischism]

Harman for Golden Ears Only?

I am now 75 and my audiologist reports that my upper frequencies drop several DB starting at about 6KHz. Using SineGen.exe, I cannot hear anything beyond 10KHz.

I therefore find that brighter equipment suits me better. My listening enjoyment factor is still great.

When the experts (Harman) develop their "ideal" FR chart, what age group are they using, and what quality ears?

Listeners were plotted on a bell curve. Naturally those with degraded hearing fall toward the outsides of the curve. The meat of the bell is where a company can place their bets as far as developing products that will please the most customers.

 

[markanini]

also think that rigid adherence to the line curve is too dogmatic given that there are likely standard deviations in preference, as well as standard deviations in the measurement setup. For example, the standard deviations are likely wider at extremes of the curve (i.e. bass and treble) for two reasons - human hearing is less sensitive to larger deviations in these frequencies, and measurement setups also show larger variations.

Some will have literall tendencies that they will apply to everything, I don't want to give such people too much credit. That's why I don't think that represents an accurate frame. It's no secret that the target represents an average, at 85dB SPL, and assumes successful fit.