Headphone Measurements Using Brüel & Kjær 5128 HATS

[Mad_Economist]

@amirm The high frequencies in pretty much all these headphones are a sea of wild peaks and troughs. Are you confident this rig isn't reflecting/cancelling at these HF frequencies?

Are headphones really that bad at HF?

Consider the wavelengths of frequencies around the band where things become...eccentric. Then consider the dimensions of the earpads, pinna features, canals, etc - you may find your answer here.

 

[amirm]

@amirm The high frequencies in pretty much all these headphones are a sea of wild peaks and troughs. Are you confident this rig isn't reflecting/cancelling at these HF frequencies?

Are headphones really that bad at HF?

You know how the room creates big peaks and valleys in low frequencies? Well, if you shrink your room to the distance between the headphone driver and your ear, then the problem area becomes the high frequencies. Reflections happen and mix with different phases.

Fortunately our hearing discrimination is poor in high frequencies so what we hear is not exactly what we see.

 

[cistercian]

Consider the wavelengths of frequencies around the band where things become...eccentric. Then consider the dimensions of the earpads, pinna features, canals, etc - you may find your answer here.

I have wondered the same thing as restorer john. Given differences in pinna, ear canal volume, head shape, etc I
no longer wonder why some people prefer one design over another.

 

[amirm]

The compensated hd650 doesn't look right...

The compensation is work in progress. The focus should be on the measurements.

 

[zenki]

Keep it up

 

[phoenixdogfan]

Looks like a crash dummy that was hastily stuffed into an American Tourister.

 

[trl]

As promised, here is the HD-650 results:
[...] Re-measuring the same but having the Audio Precision software equalize the measurements using inverted Harman computed one we get:

View attachment 78328

This indicates too little bass below 100 Hz or so. There is also lack of energy between 2 and 6 kHz as noted.

Inversely, there is too much around 7 to 10 kHz. [...]

Similar, but not identical with measurements done by Frans here and Roman here, I'm speaking mostly about the differences at around 5KHz and also after 9KHz too. Could this be from a different dummy head used or it's just a

 

[rxp]

Just another source of GRAS data: https://www.0db.co.kr/

 

[trl]

As promised, here is the HD-650 results:
[...] Re-measuring the same but having the Audio Precision software equalize the measurements using inverted Harman computed one we get:

View attachment 78328

This indicates too little bass below 100 Hz or so. There is also lack of energy between 2 and 6 kHz as noted.

Inversely, there is too much around 7 to 10 kHz. [...]

Similar, but not identical with measurements done by Frans here and Roman here, I'm speaking mostly about the differences at around 5KHz and also after 9KHz too.

Taken from diyaudioheaven
​

Could this be from a different dummy head used or it's just a different or probably modded pair of headphones?

 

[amirm]

Here are consistency tests for HD-650. Note that I was really rough and random with these to find the limits. On the other hand, the HD-650 sticks to this "head" like nobody's business. It is as if it is made for this Hats.

Looks like above 10 kHz it is a crap-shoot.

Here it is the geometric mean:

 

[Mad_Economist]

Looks like above 10 kHz it is a crap-shoot.

In general, this is a pretty safe assumption with headphones - sometimes it's a crap shoot lower down, but low variation past 10khz is pretty atypical barring niche cases.

 

[amirm]

Here is the AKG N700NC with its much smaller cups which is hard to place:

Seems geo mean generates reasonable results though: (graph label is wrong -- should say N700NC)

 

[Mad_Economist]

Note that as variable as the N700NC's low frequencies are here, it'd be worse in passive mode - the DSP is actively trying to keep things around the same level there.

 

[amirm]

Here is the Sony MDR-7506:

Quite a mess. Only area that is reliable is from 2 to 4 kHz!

Seems any target for bass would be wrong for this one due to high variability.

Wonder where that dip at 200 Hz is coming from.

 

[amirm]

Note that as variable as the N700NC's low frequencies are here, it'd be worse in passive mode - the DSP is actively trying to keep things around the same level there.

Yeh, seems a lot better than the Sony which has a similar cup size.

 

[BYRTT]

@amirm The high frequencies in pretty much all these headphones are a sea of wild peaks and troughs. Are you confident this rig isn't reflecting/cancelling at these HF frequencies?

Are headphones really that bad at HF?

Know you got some good answers already but also its a fun target as seen below that can be extracted from spreadsheet in post 1 and a thanks to @Mad_Economist for that, as far i understand Harman has researched that specific head phone curve is equal to a good speaker in a good room setup so transducer needs possible not be very flat from the start to design head phones, and the close distances to boundarys and pressurization gain must be the helper that add the bass performance because looking at impedance charts for head phones most of them are resonating around 100Hz area..

 

[Mad_Economist]

Know you got some good answers already but also its a fun target as seen below that can be extracted from spreadsheet in post 1 and a thanks to @Mad_Economist for that, as far i understand Harman has researched that specific head phone curve is equal to a good speaker in a good room setup so transducer needs possible not be very flat from the start to design head phones, and the close distances to boundarys and pressurization gain must be the helper that add the bass performance because looking at impedance charts for head phones most of them are resonating around 100Hz area..
View attachment 78382

I think people slightly miss what Sean was going for regarding "good headphones sound like a good speaker in a good room" - the main conclusion of Olive, Welti, & McMullin 2013 is that when you equalize a speaker flat at the listening position with an omni microphone (array), measure it with a head, and equalize a headphone to the same eardrum response on that head, people prefer roughly the same EQ adjustments to both.

Edit: I may have mis-parsed your comment - if you're talking about why headphone response remains linear or falls from a lower frequency than Fs, it is indeed due to the pressure chamber effect; hold a headphone next to a HATS' head sometimes, that's where different principles of design really show their differences!

 

[BYRTT]

Here is the Sony MDR-7506.....Wonder where that dip at 200 Hz is coming from.

Think culprit is they try probably acoustic or whatever trick to design a low end boost starting at 200Hz as in the Harman target curve then trick cost a physical dip before it starts raise, it looks you benched L-ch here because on page 1 we can see the dip is a bit smoother cost compared for the R-ch.

 

[amirm]

Stax SR-007 Vintage headphone driven by SRM-007t differential tube amplifier:

Geomean:

Tyll measurements:

Wonder if the low frequency peaking is the tube amp. Let me replace it with solid state and see what I get.....

[this is a 20 year old headphone so probably not the way it was then as far as pad and elastic.]

 

[imagidominc]

I'm very excited to see headphones getting measured. They're what I use everyday anyway. That price for the equipment is steep though.
I would love to see the Focal Elex and the HD800 reviewed, but that's probably not gonna happen if you have to return it. There are tons of opinions of those headphones online, but I don't trust anyone's opinion on this stuff as much as Amir's.

Zeos said the Elex have a wide soundstage. I love my Elex, but wide? No...