• 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!

Is Flat Bass or Harman Bass Better for Mixing on Headphones? (Professional Mixing Engineers Only)

Is Flat Bass or Harman Bass Better for Mixing on Headphones? (Professional Mixing Engineers Only)


  • Total voters
    31
OP
Sharur

Sharur

Senior Member
Joined
Apr 10, 2021
Messages
476
Likes
213
My understanding of the current Harman curve is that it is essentially the diffuse field curve with added bass.
There are some key differences in the treble which @thewas's comment helped me better understand. Ideally, the Harman target should be somewhere between diffuse field and free field. As can be seen in the Harman 2019 target, the extra energy from 5-10 kHz shouldn't be there nor should it drop like a rock after 10 kHz.
graph - 2021-09-14T193648.103.png

graph - 2021-09-14T193723.451.png

graph - 2021-09-14T193654.290.png

graph - 2021-09-14T193650.776.png
 
Last edited:

Shazb0t

Addicted to Fun and Learning
Joined
May 1, 2018
Messages
643
Likes
1,230
Location
NJ
There are some key differences in the treble which @thewas's comment helped me better understand. Ideally, the Harman target should be somewhere between diffuse field and free field, if I'm not mistaken. As can be seen in the Harman 2019 target, the extra energy from 5-10 kHz shouldn't be there nor should it drop like a rock after 10 kHz.
View attachment 153409
View attachment 153411
View attachment 153410
View attachment 153412
Regardless of whether you want to use the diffuse field or etymotic research curve, my answer regarding your question about added bass still stands. The problem with the other choice in your poll being "flat" is what represents flat? Lol.
 

Zim

Active Member
Forum Donor
Joined
Jul 9, 2021
Messages
107
Likes
212
Location
Kangarooland
Wouldn't the "Harman bass" be considered the flat-bass-equivalent in speakers based on the research?
 

Shazb0t

Addicted to Fun and Learning
Joined
May 1, 2018
Messages
643
Likes
1,230
Location
NJ
Wouldn't the "Harman bass" be considered the flat-bass-equivalent in speakers based on the research?
I don't think so. My understanding is that it has added bass based on preference testing where listeners increased bass above "flat bass" to levels that they preferred. IMO the prior revision of the Harman curve, that was the base curve for the preference testing, is what you're thinking of. Alternatively, the diffuse field curve is one of a few other close alternatives.
 
Last edited:

ADU

Major Contributor
Joined
Jul 29, 2021
Messages
1,584
Likes
1,086
Harman's 2018 Over-Ear Headphone target versus Oratory's 2021 Diffuse Field compensation curve, in case anyone's interested....

HARMANPLUSDIFFUSE.jpg


To make the difference a little bit easier to see and contemplate, I've traced over the two curves. The Harman curve is in black, and the DF curve is in green.

HARMANPLUSDIFFUSE2.jpg


Imo, Oratory's DF compensation curve is probably a bit too depressed in the low treble, and maybe also a tad too bright in the 2 to 3 kHz range to accurately reflect the DF response of his specific gear, pinnae, and measurements.

The DF curve above is probably not completely accurate for the rig that Harman used as well. (To the best of my knowledge, they never published any DF measurements for their specific rig and pinnae.) The upshot is that you have to take some of the differences in the levels between the two curves, esp. in the upper mids and low treble, with a grain of salt.

The Harman curve (in black) is also generally too rolled off in the high treble to be considered neutral in that range. That is why a few of the headphones on the above plot are excursing above the Harman target in the high treble. That is normal imo for a neutral headphone.

The dip at around 9-10 kHz in the treble is also a normal feature for a neutral headphone on Ora's plots imo. And I believe a neutral headphone should also have a bit more of a dip in the upper mids than the Harman target, in approximately the 1.5 to 2 kHz range. That is why most of the headphones on the above graph are sagging down a bit in that area.

I chose the 5 headphones above specifically for their general similarity in shape to the Harman curve btw. That does not necessarily mean that I believe they have the most neutral responses though. Because I don't think the Harman target is completely neutral. (And I'm not referring specifically to the bass.)

Imo, the headphone on the above graph that is probably closest to a neutral response is the Senn HE-1, which is a very expensive e-stat. And the next closest might be the Onkyo A800, which is a dynamic driver headphone that is no longer in production (and has an unusually wide headband). The Onkyo is an open headphone though, so it rolls off significantly in the bass below 50 Hz. And the HE-1 probably drops down a bit more that necessary in the sub-bass as well. It is also an open headphone btw.

The dips or notches in the low treble at around 4k on the Beyer Custom One and AKG K371 are not normal features of a neutral response imo, btw. The notch on the Beyer is so narrow though, that I might just leave it as is. And not bother trying to correct it. And the two bumps in the bass on the Beyer and PSB are also not neutral.

Ora's graphing tool used for the above plots: https://headphonedatabase.com/oratory/headphones
 
Last edited:

ADU

Major Contributor
Joined
Jul 29, 2021
Messages
1,584
Likes
1,086
index.php


This is exceedingly difficult for alot of people to grasp. But the reason that the Harman curve (and headphones with a similar response) appear tilted by about -1 to -1.5 dBs per octave compared to the diffuse field curve is because the DF curve represents the HATS measurement rig's raw response to a spectrally flat diffuse sound field, that has equal amplitude in the bass, midrange and treble.

That is not what you get when you put two anechoically flat loudspeakers into a typical semi-reflective room though, like in your home. The air, and wider dispersion of the speakers in the lower frequencies, and the boundary gain in the room will all tend to boost the amplitude of the lower frequencies relative to the higher frequencies. Giving the speakers a darker overall sound or tilt in a room. That is perfectly normal and neutral btw. And it's how we are used to hearing sound in a typical semi-reflective room.

You can see this tilt in the estimated sound power and in-room responses of the spinorama plots of some of the better loudspeakers here...

https://pierreaubert.github.io/spinorama/

Average Estimated Sound Power of 5 Well-Extended Neutral Loudspeakers:

AVGSOUNDPOWEROF5SPEAKERS.jpg


That explains the tilt in the Harman curve and headphones on the above raw plots relative to the diffuse field curve. And also on diffuse field compensated graphs, like the ones I've been using here...

https://www.audiosciencereview.com/...ng-the-250-ohm-beyerdynamic-dt-770-pro.26249/

The Harman headphone target and headphones are all attempting to approximate the more darkly tilted in-room response that anechoically flat neutral loudspeakers acquire when they are used in a typical semi-reflective room.

If you have not yet seen it, this video presentation by Floyd Toole of Harman does a much better job of explaining why speakers behave this way in a room. And it also explains a little better how to read and interpret the spinorama plots in the first link above.

 
Last edited:

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
Wouldn't the "Harman bass" be considered the flat-bass-equivalent in speakers based on the research?

yes except that headphones target curves are not as standardized as speaker target curves.

The “Harman bass” sounds the best to users but that doesn’t mean it would translate the best to speakers or even other headphones when it comes to mixing.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
index.php


This is exceedingly difficult for alot of people to grasp. But the reason that the Harman curve (and headphones with a similar response) appear tilted by about -1 to -1.5 dBs per octave compared to the diffuse field curve is because the DF curve represents the HATS measurement rig's raw response to a spectrally flat diffuse sound field, that has equal amplitude in the bass, midrange and treble.

That is not what you get when you put two anechoically flat loudspeakers into a typical semi-reflective room though, like in your home. The air, and wider dispersion of the speakers in the lower frequencies, and the boundary gain in the room will all tend to boost the amplitude of the lower frequencies relative to the higher frequencies. Giving the speakers a darker overall sound or tilt in a room. That is perfectly normal and neutral btw. And it's how we are used to hearing sound in a typical semi-reflective room.

You can see this tilt in the estimated sound power and in-room responses of the spinorama plots of some of the better loudspeakers here...

https://pierreaubert.github.io/spinorama/

Average Estimated Sound Power of 5 Well-Extended Neutral Loudspeakers:

View attachment 153444

That explains the tilt in the Harman curve and headphones on the above raw plots relative to the diffuse field curve. And also on diffuse field compensated graphs, like the ones I've been using here...

https://www.audiosciencereview.com/...ng-the-250-ohm-beyerdynamic-dt-770-pro.26249/

If you have not yet seen it, this video presentation by Floyd Toole of Harman does a much better job of explaining why speakers behave this way in a room. And it also explains a little better how to read and interpret the spinorama plots in the first link above.


Isn’t the goal of all listening rooms is to aspire to become a perfect Diffuse field at all frequencies?

I’m not sure about reproduction but for production that would be preferred.
 

ADU

Major Contributor
Joined
Jul 29, 2021
Messages
1,584
Likes
1,086
Isn’t the goal of all listening rooms is to aspire to become a perfect Diffuse field at all frequencies?

I’m not sure about reproduction but for production that would be preferred.

It probably depends on the application, abdo123.

I think there will still be a tilt in the response though due to the beaming and narrower dispersion in the speakers' tweeters, and the wider dispersion in the lower frequencies. Even if the room itself reflects back all frequencies equally from any given direction.

The room contributes, but I believe it is largely the dispersive or directivity characteristics of the speakers themselves that cause a timbral shift toward the lower frequencies. Floyd Toole explains this better than I can in the above video though, which is well worth the hour or so it takes to give it a listen.

Most (though not all) loudspeakers for home and studio use become more narrow in their dispersion as they go up in frequency. And that is mostly what causes the tilt in their in-room versus anechoic/direct response.
 
Last edited:

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
It probably depends on the application, abdo123.

I think there will still be a tilt in the response though due to the beaming and narrower dispersion in the tweeter, and the wider dispersion in the lower frequencies. Even if the room itself reflects back all frequencies equally from any given direction.

The room contributes, but I believe it is largely the dispersive characteristics of the speakers themselves that cause a timbral shift toward the lower frequencies. Floyd Toole explains this better than I can in the above video though, which is well worth the hour or so it takes to give it a listen.

Most (though not all) loudspeakers for home and studio use become more narrow in their dispersion as they go up in frequency. And that is mostly what causes the tilt in their in-room versus anechoic response.

I think it's better to define what small room acoustics are and what studio acoustics are. That's the issue i think this discussion is having.

If i'm sitting less than a meter away from my (anechoicly flat) speakers then i think the downward tilt would not be as significant since the ratio of direct to reflected sound is very high. specially if the boundaries of the room are also far enough that the amplitude of the reflections are even lower than what would be the case in a typical small room situation.

From my point a view a studio should have the phase cancelations and resonances of an anechoic chamber with the ambiance and 'busy-ness' of a typical small room. Mission impossible for sure but there are some state of the art studios that come close to that ideal.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
diffuse field is for speakers which is not the same as headphone target.
 

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,422
Likes
2,407
Location
Sweden
So what are the static and dynamic signal room responses 5-20000 Hz of an average control room? Main monitor speakers.
 

Koeitje

Major Contributor
Joined
Oct 10, 2019
Messages
2,292
Likes
3,880
Headphones are only useful for checking minute details, everything else should be done on monitors.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
A DF target for measuring sound in a specific reverberant room emulated in an acoustical dead and controlled condition using HATS is simply not the correct target for a headphone mounted on a HATS.
So DF is not the correct target for headphones used for mixing/mastering.

What's obscure about this and why does it not have anything to do with the discussion ?
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
A DF target for measuring sound in a specific reverberant room emulated in an acoustical dead and controlled condition using HATS is simply not the correct target for a headphone mounted on a HATS.
So DF is not the correct target for headphones used for mixing/mastering.

What's obscure about this and why does it not have anything to do with the discussion ?

it's obscure because you don't explain why a DF target is not correct.

Lets say I have built a room that everyone on earth agrees that is the best room in the world for music reproduction, If I place a measurement rig inside that room then whatever that rig measures should be the 'Target'. Even the accuracy of the measurement rig itself wouldn't matter because the 'target' itself would be act as a calibration since my room is the best room in the world. Unless the measurement rig is measuring the incorrect parameters.

Now Harman came up and said we think sub-bass should be boosted to compensate for the reduced physical bass sensation. My personal experience aligns with that and indeed I appreciate the Harman target.

However if i were mixing, and my mix will be reproduced live on a stage on speakers. I would rather use the 'target' my perfect room produced with the measurement rig. And have other people's headphones have the Harman Target.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
The reason is that DF is obtained in a acoustically dead room with a bunch of speakers in a certain pattern with a certain simulus.
There is a certain distance from those many speakers as well as angles. The whole HATS response is taken into consideration.
The response of this is substantially different from a small enclosed space around the pinna where angles differ from DF, the distance and the way it arrives at the ear entrance differs. On top of that it differs per type of headphone.
Therefore DF is NOT a proper representation of a headphone and thus the wrong one.

For Studio work, mixing and mastering (when there is no other way than using headphones) should NOT be done with Harman response.
This is for reproduction of recordings.
When emulating flat monitors at around 80-85 dB SPL in a studio with lots of acoustic treatment the sound equivalent is NOT the same as Harman.
It will be close to Harman but with the bass boost reduced to almost no boost.

The Harman headphone response is for headphones used at home/portable and music enjoyment usually not at studio levels.
So ... simply do not use a Harman response headphone in a studio (take the Stealth for instance) without simply reducing that bass hump and your mixes will be fine on reproduction at home or with headphones with Harman response.

Studio sound is not home/portable music reproduction and requires different treatment.
DF is not a proper target for usage of headphones in a studio.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
The reason is that DF is obtained in a acoustically dead room with a bunch of speakers in a certain pattern with a certain simulus.
There is a certain distance from those many speakers as well as angles. The whole HATS response is taken into consideration.
The response of this is substantially different from a small enclosed space around the pinna where angles differ from DF, the distance and the way it arrives at the ear entrance differs. On top of that it differs per type of headphone.
Therefore DF is NOT a proper representation of a headphone and thus the wrong one.

For Studio work, mixing and mastering (when there is no other way than using headphones) should NOT be done with Harman response.
This is for reproduction of recordings.
When emulating flat monitors at around 80-85 dB SPL in a studio with lots of acoustic treatment the sound equivalent is NOT the same as Harman.
It will be close to Harman but with the bass boost reduced to almost no boost.

The Harman headphone response is for headphones used at home/portable and music enjoyment usually not at studio levels.
So ... simply do not use a Harman response headphone in a studio (take the Stealth for instance) without simply reducing that bass hump and your mixes will be fine on reproduction at home or with headphones with Harman response.

Studio sound is not home/portable music reproduction and requires different treatment.
DF is not a proper target for usage of headphones in a studio.

In what situations is the Diffuse Field target relevant in your opinion?
 
Top Bottom