• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Theory: If I EQ headphones to Harman AE/OE curve, what other attributes should I look for?

Fluffy

Addicted to Fun and Learning
Joined
Sep 14, 2019
Messages
856
Likes
1,423
A target curve is also loudness (amplitude/SPL/volume) dependent.
When you listen at low volume(SPL) it differs from when playing at loud volume(SPL) .

Some headphones with a specific curve thus sound 'better' on low listening volume(SPL) and others sound good at realistic or loud volume(SPL) .
I've suspected this to be true for a long time, but I didn't really see anyone addressing that subject directly. I have this theory that I prefer headphones with curtain curves because of the loudness I'm used to listen at, which is generally higher than most people. And this is maybe why a lot of high end headphones sound bad to me, because they are tuned to be listened to at lower volumes.
 

flipflop

Addicted to Fun and Learning
Joined
Feb 22, 2018
Messages
927
Likes
1,240
This is cool, do you have a link to the source?
Got it from this post. Looks like a photo taken by Amir during a presentation by Sean Olive.

And it certainly won't sound the same in headphones. With headphones there is no room, so they should be adjusted to flat.
I hope you're not seriously suggesting using a flat raw response for headphones.

A target curve is also loudness (amplitude/SPL/volume) dependent.
When you listen at low volume(SPL) it differs from when playing at loud volume(SPL) .
The problem is dealt with by using reference playback levels as discussed in this thread.
The SPL for the 2018 Harman AE/OE target curve is also clearly defined in the paper:
2.7.PNG
 

flipflop

Addicted to Fun and Learning
Joined
Feb 22, 2018
Messages
927
Likes
1,240
Do they address what happens if someone listens at an SPL different from what they tested with? Does the target curve change?
The paper doesn't address it, but the answer is 'yes' due to how our hearing works. For instance, lower listening levels would require a target curve with more bass and vice versa. This also applies to loudspeaker listening.
1920px-Lindos1.svg.png
 

Fluffy

Addicted to Fun and Learning
Joined
Sep 14, 2019
Messages
856
Likes
1,423
If this hasn't been tested like the regular target curve, then what you are saying is merely a conjecture. Finding a target curve at a specific SPL doesn't mean you can directly scale it up with volume using an equal-loudness correction. It would be reasonable to assume so, but without directly testing it's nothing but a hypothesis.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
f-m-overlayed.png


This shows the correction needed for different average listening SPL levels to perceive the same tonal balance (right plot) .
85dB average would be between the blue and green line and would be very similar to that used during mixing (with speakers).
 

Fluffy

Addicted to Fun and Learning
Joined
Sep 14, 2019
Messages
856
Likes
1,423
As I said, this sounds reasonable. But as long as you don't remake the experiments that lead to the Harman curve, and playing at a higher\lower SPL while applying that loudness correction, you can't know if the target curve persists into different sound levels. There could be hidden variables besides simply compensating for loudness.
 

Soniclife

Major Contributor
Forum Donor
Joined
Apr 13, 2017
Messages
4,500
Likes
5,417
Location
UK
I don't know if the measurements are wrong or this Harman curve is even less ideal than I suspected, but those curves sound awful.
The Harmon methodology makes sense, but they test the response in a different way to the sites on the web, so I don't see how using those measurements to produce EQ is likely to be successful.
 

pwjazz

Addicted to Fun and Learning
Forum Donor
Joined
Sep 21, 2018
Messages
507
Likes
746
Ok, testing time. Just for amusement, I applied the curves from that list to the Audeze LCD2C and Focal Clear, which I own. And OMG NO. What the hell is this? I don't know if the measurements are wrong or this Harman curve is even less ideal than I suspected, but those curves sound awful. They are light years apart from the curves I made on my own for this headphones, even completely the opposite in some cases.

In the LCD2C they decided to put a whopping 10 DB increase at 4060hz! Of course, this is unlistenable and sounds insanely harsh at any volume other than very very low. I actually bought the lcd2c because of their very relaxed and warm sound, and at most I put a general +0.5db upward tilt to bring back some presence. This intense boost in mid treble is totally unjustifiable. The Bass rise is actually similar to what I use (they put +5 db low shelf at 110hz, I use +4 db at 130hz), and the overall curve lowers the very irritating mids that I find to be too boxy, so I also apply a reduction there. But the overall sound signature with their curve is hollow and extremely harsh.

As for the Clear, it's tamer, but still wrong for my ears. They for some reason put a -4db reduction at 200hz, a region I usually boost up because those headphones just lack in the bass guitar region. It gives a really anemic presentation. They also seem to think this headphone doesn't have enough treble energy because they put some boosts there, which makes it even more piercing than usual. I do agree on the general reduction in energy around 3500hz that I also apply myself, but they boost it back up at 4300hz which defeats the purpose. Again, this sounds to me a bit hollow and quite harsh.

Finally, I tried putting on pink noise and listening through the suggested curve of each headphone, to see if it actually results in a similar tonal balance. It's not, and I could even dare to say they sound more dissimilar with the curves then without them.

So my conclusion is that based on this very rough experiment, the Harman curve is just probably not for me. I could see how it would be appealing for short term listening because of its emphasis on presence and impactful lower bass, but for long term it seems to me extremely fatiguing and lacking in warmth.

On the other hand, this is the same sound signature I heard in numerous high-end headphones (maybe with tamer bass though), that I generally despise. So maybe there is a large percentage of the population that do appreciate this sound. Well good for them, I say. But I still don't think every manufacturer should aim for the same response. Otherwise where is the diversity and opportunity for surprise?

Glad I'm not the only one who doesn't like these presets. I find Sonarworks better for my taste.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
The whole idea of the 'bass correction' which is seen in the Harman curve ultimately is based on preference of a lot of listeners.

The real question is why 'flat measuring' isn't perceived as 'flat' as in speakers in a room when both are at 85dB Phon.
There is a simple answer to that... the absence of tactile feel.
At 85dB average with speakers in a decent room one not only gets 'input' from the ears but also from the body.
This info is missing.
A second one is that 'flat' as in speakers in a room will almost always have boosted lows due to the room.
This too is absent with headphones.

On top of this there is the question of preference of listening SPL but this is something that adds to the, for the above mentioned reasons, needed correction.

Not all people want the same corrrection (possibly based on their speaker references or real life experience) so not everyone may prefer Harman correction even though some do and feel that such correction must be applied.

R'tings, Sonarworks and Goldenears all have the 'bass correcttions' already applied (so do I but prefer a 6dB/oct and less extreme correction.
So 'flat' measurements in the above mentioned plots already 'incorporates' the extra bass correction and 'flat' in those plots of 'flat' at normal listening levels.
 
OP
N

nhunt

Member
Joined
Oct 6, 2019
Messages
53
Likes
51
The paper doesn't address it, but the answer is 'yes' due to how our hearing works. For instance, lower listening levels would require a target curve with more bass and vice versa. This also applies to loudspeaker listening.
View attachment 35550
Yeah, I believe that Bose Quietcomfort 35 headphones actually actively adjust the bass with your spl when noise canceling is turned on. Similar to dynamic EQ for Audyssey room correction and similar loudness settings.

All that said, I typically don’t listen to headphones at reference spl and have never given it much thought like I do with loudspeakers. Curious to play around with it and look at some measurements now.
 

DivineCurrent

Active Member
Forum Donor
Joined
Mar 13, 2019
Messages
109
Likes
135
I find that with EQ, the less peaky and more even the headphone's response is the better. I find especially the HD600/650 to sound much better with EQ than with the Focal Elex, I think there is some ringing and resonance issues with the Elex that EQ simply can't completely fix.

Also, I've found equalizing by ear using sine sweeps works very well. I usually start with a comfortable volume at 1 kHz, and then do a slow sweep up to 10 kHz, and then from 10 to 20 kHz. If there is any range of frequency that sounds much louder (or softer) than the original 1 kHz tone, I adjust EQ until the sweep sounds relatively even in volume.

Everyone's ears are shaped differently and have different resonances, so I think that's why listening preferences differ so much. I find most IEMs to be much too forward in the upper mids/lower treble from 2-6 kHz, so I always use EQ with IEMs. Using the sine sweep method, I've actually made my Etymotic ER2-SE to have a very similar tonality to the HD600.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
I think there is some ringing and resonance issues with the Elex that EQ simply can't completely fix.

Haven't heard the Elex but with the other Focals this surely is the case.
 

Soniclife

Major Contributor
Forum Donor
Joined
Apr 13, 2017
Messages
4,500
Likes
5,417
Location
UK
Also, I've found equalizing by ear using sine sweeps works very well. I usually start with a comfortable volume at 1 kHz, and then do a slow sweep up to 10 kHz, and then from 10 to 20 kHz. If there is any range of frequency that sounds much louder (or softer) than the original 1 kHz tone, I adjust EQ until the sweep sounds relatively even in volume.
That seems a similar approach to the neutralizer app, but rather more expert level. I've always wondered if several people did this with the same headphones would you get the same results, or does the individuals hrtf come into play.
 

Fluffy

Addicted to Fun and Learning
Joined
Sep 14, 2019
Messages
856
Likes
1,423
I've always wondered if several people did this with the same headphones would you get the same results, or does the individuals hrtf come into play.
I bet there would be similarities, but you will never get the exact same results.

I also take a similar approach when equalizing a new headphone, but after I've done the "dry" eq pass according to the sweep, I take a listen and see if it actually improves anything. Most of the time you can't rely only on EQing based on just sinewaves. Your brain doesn't listen in single tone sinewaves, and you need to see how it affects the music perception as a whole.
 

DivineCurrent

Active Member
Forum Donor
Joined
Mar 13, 2019
Messages
109
Likes
135
That seems a similar approach to the neutralizer app, but rather more expert level. I've always wondered if several people did this with the same headphones would you get the same results, or does the individuals hrtf come into play.
Huh, that's pretty cool. I didn't realize there was an app that does that.

There are a few things I've noticed while doing sine sweeps however. Almost always, 5-6kHz needs to be lower in level in order for the music to sound natural. I read here that the Axel Grell, the engineer from Sennheiser responsible for the HD580, 600 and 650, purposefully put notches at 5kHz and 16kHz to acheive a subjectively more natural sound for the 650. Not sure if that plays any part in Harman tuning, but definitely something to consider.
 

mitchco

Addicted to Fun and Learning
Audio Company
Joined
May 24, 2016
Messages
640
Likes
2,397
Personally, I enjoy the Harman target curve. I find most headphones to be thin sounding. For sure it is all about preference, but as Sean Olive has had literally hundreds of participants over the years take part in headphone perception testing, most tend to have the same preference.

Here are two PDF slide presentations from Sean Olive and team that document "how" they performed the controlled subjective listening tests, the measured objective results and conclusions:

The Perception and Measurement of Headphone Sound Quality: Do Listeners Agree on What Makes A Headphone Sound Good?
and
Perception and Measurement of Headphone Sound Quality: Is There a Preferred Target Response.

The presentations summarise 5 years of research and if you are really interested, it is worth not only reading the presentations, but getting the 14 AES papers that provide the details.

To date, this is the best headphone listener preference research available as far as I know and far surpasses anything else I have seen in this area (and for loudspeakers too as you will read in the presentations). The methodology and approach is very interesting (to me anyways).

Finally, there is no correlation of price versus headphone frequency response as I found out in this review.

I hope folks spend the time to understand the research and conclusions as I found it educational and to my ears, in agreement.
 

JJB70

Major Contributor
Forum Donor
Joined
Aug 17, 2018
Messages
2,905
Likes
6,148
Location
Singapore
The correlation between frequency response and price, or rather the lack of correlation, is no surprise. In general I really don't think most expensive headphones (I would use the Sennheiser HD660 as a threshold for defining that point) are worth it in terms of SQ. They may be worth it in terms of premium materials and build etc) but rarely for SQ.
 

Hipper

Addicted to Fun and Learning
Joined
Jun 16, 2019
Messages
753
Likes
625
Location
Herts., England
Hi! Wanted to pose a little thought experiment:

Assume I can theoretically get a frequency response curve for a pair of headphones that matches the Harman 2018 target AE/OE curve (published at http://www.aes.org/e-lib/browse.cfm?elib=19436, and available at various places like RTINGS ratings and Oratory1990's EQs over on reddit). What other factors should I look at in a pair of headphones presuming I can EQ them to an ideal FR? For example, isolation would be one of the obvious ones.

Many of the factors listed by Flipflop in post 4 are worth looking at, of which I would highlight comfort, both short term and for longer listening sessions. In addition being able to clean and/or replace ear pads is useful.

For playback, as well as EQ, you might consider a crossfeed processor (there's a debate to be had on this), and perhaps a balanced headphone cable with balanced headphone amp (another, maybe more heated debate!), or at the very least, perhaps a better headphone terminal then the ubiquitous 6.5mm jack. I don't necessarily recommend any or all of these, just that these are available options to consider. There are furthermore other processing abilities available, such as those that can mimic speaker type listening experience like 'Out of Your Head'.
 
Top Bottom