Harman preference curve for headphones - am I the only one that doesn't like this curve?

[Frank Dernie]

Well, this is proposed, but remember that people who grew up on using smaller speakers and/or headphones might not be accustomed to the amplified bass due to body response.

Well I have had big full range speakers since about 1970 and whilst I have had to work on positioning them to minimise their excitation of room modes I have enjoyed extended clean bass from the ones I have chosen since then. I have auditioned a fair few rubbish-in-the-bass speakers over the years though.

 

[Dealux]

IEM bass boost is controversial, yet not quite unfounded - it is based on research concerning how to compensate for lack of bodily bass perception in IEMs.

Issue is it creates a thunderous sub bass (like an out of control sub in a speaker system) that muddies up the bass in many songs.

 

[watchnerd]

I find the excess bass unpleasant and false on the sort of music I listen to, mainly classical, but I understand most people listen to pop music and like extra bass. I was like that for at least the first 10 years of my listening life.

Ditto.

I don't like the Harman curve on classical or acoustic jazz.

It's mediocre on fusion and classic rock.

It shines on synthetic pop and techno.

 

[watchnerd]

The curve is a mix of seasoned listeners and general folk (after being screened and whatnot), and it’s not a surprising conclusion that those people that put 18” subs in their SUVs and blast out the neighborhood are likely to like more bass than the people that visit audiophile forums.

It's the audio equivalent of The Olive Garden restaurant.

 

[blse59]

Hi all

With some incredible help from fellow forum members here (@pkane , @Robbo99999 , @dasdoing ) I've been able EQ my headphones to the Harman preference curve.

I started with my own favourite Focal Elegia and just today I've EQ'd my friends Aeon 2 Closed and HD800-S , all using oratory1990 EQ parameters to achieve Harman curve.

With all these headphones I really hate the Harman curve (please don't delete my account @amirm )

It sounds nothing like my nearfield desktop Genelec monitors (1m from my head).

I thought maybe it was my Focal Elegia but I have the same tonality when I eq Aeon 2 closed and HD800-S

A flat headphones curve sounds closer to the nearfield speakers to me.

It's possible that as I continue to tweak target curves and listen, that I end up preferring something between these 2 extremes of flat vs Harman curve - the journey continues, so please don't shoot.

Am I alone here in really not liking Harman preference curve? Like really not liking it

A flat frequency curve is my personal preference. It's what I would call true hi-fi or high fidelity.

 

[watchnerd]

The idea the Harman curve is some universal 'right way' for everyone and every headphone is more religion than science and encouraged by places that worship the research and gather together online to baske in their righteous virtue and wisdom.

The top selling beers in the world are:

#1 Snow
#2 Budweiser
#3 Tsing Tao
#4 Bud Light
#5 Skol

Ergo, a beer made the "right way" should taste like those pale lagers.

Now, someone will point out that beers are not reproducers, and rightly so. But it's illustrative of the dangers of equating "what sells best" with "what is right."

 

[watchnerd]

Here is the other challenge around being slavish to the Harman curve:

What does it mean when preferences change?

I don't think we should expect a preference curve to be immutable over time as mass musical tastes change and evolve.

 

[celroid]

I own the Sennheiser HD598 (the old version with brown and cream) and I tried out the updated oratory1990 EQ from reddit:
https://www.dropbox.com/s/flxjqweapsikqm8/Sennheiser HD598SR.pdf?dl=0
and then also the EQ from jaakkopasanen:
https://github.com/jaakkopasanen/AutoEq/tree/master/results/oratory1990/harman_over-ear_2018/Sennheiser HD 598

I prefer the EQ from jaakkopasanen over the new one from oratory1990. It sounds a lot more clear to me, and I like the boost in treble.

Edit: have to mention I replaced the original velour pads with some pads from aliexpress, as the old ones completely broke down over the years.

 

[celroid]

The top selling beers in the world are:

#1 Snow
#2 Budweiser
#3 Tsing Tao
#4 Bud Light
#5 Skol

Ergo, a beer made the "right way" should taste like those pale lagers.

Now, someone will point out that beers are not reproducers, and rightly so. But the it's illustrative of the dangers of equating "what sells best" with "what is right."

I much prefer Weißbier from a German Bräuhaus over store bought beer

 

[watchnerd]

I much prefer Weißbier from a German Bräuhaus over store bought beer

I have my weekend's stout in a reusable growler filled at one of the local breweries.

 

[bobbooo]

I own the Sennheiser HD598 (the old version with brown and cream) and I tried out the updated oratory1990 EQ from reddit:
https://www.dropbox.com/s/flxjqweapsikqm8/Sennheiser HD598SR.pdf?dl=0
and then also the EQ from jaakkopasanen:
https://github.com/jaakkopasanen/AutoEq/tree/master/results/oratory1990/harman_over-ear_2018/Sennheiser HD 598

I prefer the EQ from jaakkopasanen over the new one from oratory1990. It sounds a lot more clear to me, and I like the boost in treble.

Edit: have to mention I replaced the original velour pads with some pads from aliexpress, as the old ones completely broke down over the years.

Pads can have a huge effect on the frequency response of a headphone, sometimes even more than the driver, so those EQ settings likely aren't taking it to the Harman curve.

 

[celroid]

Have the Sean Olive studies been reproduced by others? Has anyone found negative results? I personally hated statistics and I wonder if qualified people have looked at the studies, to make sure there was absolutely no way p-hacking was done for the harman preference to be “statistically significant”.

I do not in any way mean to insult the researcher/s involved, I am just very skeptical of academic journals in general, due to many reasons such as the reproducibility crisis, because of "publish or perish" and other reasons...

 

[watchnerd]

Have the Sean Olive studies been reproduced by others? Has anyone found negative results? I personally hated statistics and I wonder if qualified people have looked at the studies, to make sure there was absolutely no way p-hacking was done for the harman preference to be “statistically significant”.

I do not in any way mean to insult the researcher/s involved, I am just very skeptical of academic journals in general, due to many reasons such as the reproducibility crisis, because of "publish or perish" and other reasons...

I'm sure the stats are fine.

Harman has no incentive to game it, and in fact, wants to make sure it's correct.

After all, they're using it to make multi million dollar product development decisions.

No company wants to build products on inaccurate market research and then face bad ROI on product development monies.

But, at the end of the day, that's what it is.....very good market research.

 

[Jimbob54]

Have the Sean Olive studies been reproduced by others? Has anyone found negative results? I personally hated statistics and I wonder if qualified people have looked at the studies, to make sure there was absolutely no way p-hacking was done for the harman preference to be “statistically significant”.

I do not in any way mean to insult the researcher/s involved, I am just very skeptical of academic journals in general, due to many reasons such as the reproducibility crisis, because of "publish or perish" and other reasons...

Oh dear, you went there.

 

[watchnerd]

Oh dear, you went there.

As a marketer, the thing I would be curious to know about the Harman curve is if there is any correlation (positive or negative) between population demographics and the preference curve.

 

[celroid]

I'm sure the stats are fine.

Harman has no incentive to game it, and in fact, wants to make sure it's correct.

After all, they're using it to make multi million dollar product development decisions.

No company wants to build products on inaccurate market research and then face bad ROI on product development monies.

But, at the end of the day, that's what it is.....very good market research.

Well there are many in the business so what's the benefit for Harman to publish their findings and help their competitors?

Wouldn't it'd be in their interest to publish results with a preference curve that other manufacturers will follow, and so make bad sounding headphones based on that research? Maybe they could have a separate set of results that they do not publish to make better sounding headphones than the competition. Cartoonishly evil

I do not actually believe this, as I actually like the harman curve eq for my HD598 but it was a funny thought.

 

[watchnerd]

Well there are many in the business so what's the benefit for Harman to publish their findings and help their competitors?

To establish thought leadership. It's a pretty standard marketing tactic for technology firms.

Which has worked, if ASR is any evidence.

And it doesn't help their competitors much if the competitors aren't competing in the same market segments and can't execute at the same scale.

Very few companies can compete at the mass scales, volumes, and margins that Samsung operates at.

A smaller volume maker might look at the curve and say:

"Okay, we can't compete with Samsung / Harman on cost and value for the mass market, so let's target the users that are more niche audience outside the Harman preference curve. We'll cede the mass market territory to Harman, and sell to the 25-50% who prefer something different."

You see analogous behavior with DAC makers who need to invent differentiation with things like multi-bit/R2R DACs in order to create higher margin products at lower volume, instead of competing with mass producers who do pretty standard implementations of ESS/AKM etc and compete on price.

 

[Jimbob54]

As a marketer, the thing I would be curious to know about the Harman curve is if there is any correlation (positive or negative) between population demographics and the preference curve.

To establish thought leadership. It's a standard marketing tactic for technology firms.

Which has worked, if ASR is any evidence.

And it doesn't help their competitors much if the competitors aren't competing in the same market segments and can't execute at the same scale.

Very few companies can compete at the mass scales, volumes, and margins that Samsung operates at.

A smaller volume maker might look at the curve and say:

"Okay, we can't compete with Samsung / Harman on cost and value for the mass market, so let's target the users that are more niche audience outside the Harman preference curve. We'll cede the mass market territory to Harman, and sell to the 25-50% who prefer something different."

I'm far from convinced headphone sales in pretty much any price bracket are driven by anything to do with the sonics and everything to do with who wears them or the story of how they were created.

So doing much more detailed research on the demographics of sonic preference might not yield much from a sales angle.

 

[watchnerd]

I'm far from convinced headphone sales in pretty much any price bracket are driven by anything to do with the sonics and everything to do with who wears them or the story of how they were created.

So doing much more detailed research on the demographics of sonic preference might not yield much from a sales angle.

You just gave the reason why it matters:

Marketing and advertising

If listeners <35 years old prefer the Harman curve more than listeners >35 years old, that influences your marketing audiences and advertising content and promotion strategies.

That's why I said "as a marketer", I would want to know that.

 

[andreasmaaan]

A flat frequency curve is my personal preference. It's what I would call true hi-fi or high fidelity.

That's perfectly valid with respect to electronics and (to a limited extent) loudspeakers, but headphones present special difficulties that make defining a "flat" response highly contentious in the first place.

@Ilkless summarised some of the difficulties and attempted resolutions in this article.