Is the Harman target the most neutral headphone target?

[Soria Moria]

Hi. In the headphone and IEM space I've noticed that people tend to refer to headphones with flat measuring bass as neutral, but from my own listening using Oratory's EQ presets on various headphones and the Quarks DSP, any headphone that alligns with the Harman target sounds a LOT more like a pair of reference speakers than any so-called 'neutral' headphone I've heard. Headphones I've owned and used that people have referred to as neutral and mixing headphones are ones such as HD800S and LCD-X, but both of these sound very coloured to me and I would definitely make a terrible sounding mix with them un-EQed.

 

[GaryH]

I think it is well established that flat bass in a headphone or in-room response of a loudspeaker is NOT neutral. We’ve had trained listeners draw the perceived spectral balance of these targets and they are perceived as not flat.
Toole spent 10 years having listeners rate loudspeaker based on perceived fidelity/neutrality. When we switched to preference, the loudspeakers ratings didn’t suddenly change. There is a high correlation between fidelity/neutrality/ preference.

Our headphone targets do not deviate significantly above 200 Hz from a anechoically flat speaker measured in our reference room at the DRP. For the AE/OE target it’s within 2 dB of the bass of the in-room speaker target. For the IE target it’s higher, but there are data to support it needs to be higher to be perceived as equivalent

 

[Soria Moria]

Very thankful!

 

[kemmler3D]

The harman curve(s) are meant to be neutral (compared to box speakers in-room) on average, but can't account for any quirks of your own HRTF or hearing in general. And it's hard to say whether any given headphone keeps to a given curve above 10khz (at best) anyway since measurements are not very reliable there. So IMO some EQ is probably necessary for everyone to get a neutral-to-them sound, but how exactly to achieve that is not super straightforward, since you have to rely on your ears to do it.

 

[fpitas]

Just like with speaker response, one size does not fit all. Some people like more bass. Some like less. Harman simply wanted a reasonable middle-of-the-road curve to make the most customers happy.

 

[kemmler3D]

Just like with speaker response, one size does not fit all. Some people like more bass. Some like less. Harman simply wanted a reasonable middle-of-the-road curve to make the most customers happy.

With speaker response, you can at least be confident that the response of the speaker is what people hear. They come to the speaker with their entire HRTF intact.

With headphones, you're effectively attaching the driver to a custom-built horn with a microphone at the small end of it. You are not allowed to know the specifications of the horn in advance, just rough dimensions. So you don't know what any given person actually hears, let alone what they would PREFER to hear. Now, optimize the response at the microphone.

The Harman curve does a good job outlining the average parameters of this "horn" but cannot account for every variant, that's up to the owner.

Someday I assume we will be able to get high-res laser scans that are used to generate personal FIRs for HRTF filtering and beyond... and all headphones can just be ruler-flat out of the box... until then you need to EQ with stock curves plus ears.

 

[fpitas]

With speaker response, you can at least be confident that the response of the speaker is what people hear. They come to the speaker with their entire HRTF intact.

With headphones, you're effectively attaching the driver to a custom-built horn with a microphone at the small end of it. You are not allowed to know the specifications of the horn in advance, just rough dimensions. So you don't know what any given person actually hears, let alone what they would PREFER to hear. Now, optimize the response at the microphone.

The Harman curve does a good job outlining the average parameters of this "horn" but cannot account for every variant, that's up to the owner.

You're right about the wild card of your ear canal etc, but speakers get placed in rooms of all different types, too. In both cases some tweaking is a good idea if you want it just right for you.

 

[kemmler3D]

You're right about the wild card of your ear canal etc, but speakers get placed in rooms of all different types, too. In both cases some tweaking is a good idea if you want it just right for you.

Correct about rooms. But if I am developing EQ curves... I'd rather deal with the acoustics of a large-ish rectangular prism than an ear...

 

[julian_hughes]

Hi. In the headphone and IEM space I've noticed that people tend to refer to headphones with flat measuring bass as neutral, but from my own listening using Oratory's EQ presets on various headphones and the Quarks DSP, any headphone that alligns with the Harman target sounds a LOT more like a pair of reference speakers than any so-called 'neutral' headphone I've heard. Headphones I've owned and used that people have referred to as neutral and mixing headphones are ones such as HD800S and LCD-X, but both of these sound very coloured to me and I would definitely make a terrible sounding mix with them un-EQed.

The Harman curve describes a statistical, aggregated listening *preference*. It's not an attempt to describe a neutral target.

 

[GaryH]

The Harman curve describes a statistical, aggregated listening *preference*. It's not an attempt to describe a neutral target.

Read the quote I posted above from Dr Sean Olive.

 

[julian_hughes]

Read the quote I posted above from Dr Sean Olive.

No, I refuse to do so.

 

[GaryH]

 

[dlaloum]

Read the quote I posted above from Dr Sean Olive.

If you go back to that thread (from which your quote came) - it is clear that Dr Olive was referring to work identifying PERCEPTION.

ie: most people's PERCEPTION of Neutral (which as we know, is impacted by sound levels and by loudness curves, etc...) - aligns with the Harman curve (as it was developed based on an average of the tested peoples perceptions!!

True Measurable "Neutrality" - is a whole different beast - and this is particularly indicated by Dr Olive's specific reference for a different curve when used for Binaural and/or HRTF purposes (where, I would suggest, true "measurable" neutrality" would be a requirement).

Also when considering that this therefore becomes a discussion about perceived neutrality, focusing on stereo recordings, listened to through headphones (without HRTF application...) - the volume level at which the listener chooses to use the headphones will in turn impact the listeners perception of neutrality as per the standard loudness perception curves...

Which of course implies, that different headphones will sound best at different volume levels - and that different users preferences, in terms of headphones, may also be driven by the levels at which they listen.

If you combine loudness curves, with a measurably flat/neutral headphone response - and set the listening level to the average used for the Harman tests that defined the Harman Curve.... do you end up with roughly the Harman Curve?
If not - then there is something else going on as well... if so then we have an understanding of what drives preferences, in a consistent system.

 

[julian_hughes]

If you go back to that thread (from which your quote came) - it is clear that Dr Olive was referring to work identifying PERCEPTION.

Yes! Thank you.

 

[GaryH]

If you go back to that thread (from which your quote came) - it is clear that Dr Olive was referring to work identifying PERCEPTION.

ie: most people's PERCEPTION of Neutral (which as we know, is impacted by sound levels and by loudness curves, etc...) - aligns with the Harman curve (as it was developed based on an average of the tested peoples perceptions!!

True Measurable "Neutrality" - is a whole different beast - and this is particularly indicated by Dr Olive's specific reference for a different curve when used for Binaural and/or HRTF purposes (where, I would suggest, true "measurable" neutrality" would be a requirement).

Also when considering that this therefore becomes a discussion about perceived neutrality, focusing on stereo recordings, listened to through headphones (without HRTF application...) - the volume level at which the listener chooses to use the headphones will in turn impact the listeners perception of neutrality as per the standard loudness perception curves...

Which of course implies, that different headphones will sound best at different volume levels - and that different users preferences, in terms of headphones, may also be driven by the levels at which they listen.

If you combine loudness curves, with a measurably flat/neutral headphone response - and set the listening level to the average used for the Harman tests that defined the Harman Curve.... do you end up with roughly the Harman Curve?
If not - then there is something else going on as well... if so then we have an understanding of what drives preferences, in a consistent system.

Perceived neutrality is the only neutrality that matters.

 

[julian_hughes]

That's only your perception.

 

[kemmler3D]

Semantics, here:

Perceived neutrality = I am perceiving something that is known to be objectively neutral?

or

Perceived neutrality = I am perceiving something that is subjectively similar to something else that's objectively neutral?

or

Perceived neutrality = I am perceiving something that sounds subjectively neutral to me, without reference to objective neutrality?

 

[julian_hughes]

I am perceiving my navel.

 

[dlaloum]

Semantics, here:

Perceived neutrality = I am perceiving something that is known to be objectively neutral?

or

Perceived neutrality = I am perceiving something that is subjectively similar to something else that's objectively neutral?

or

Perceived neutrality = I am perceiving something that sounds subjectively neutral to me, without reference to objective neutrality?

- Although there was mention of "experienced listeners" defining it as "neutral"

But as we already know, even experienced listeners can easily be tricked in double blind tests - with minor volume level variations