Headphone Measurements Using Brüel & Kjær 5128 HATS

[solderdude]

AKG K92. The two channels were massively out of balance. Wonder if this is behind people saying imaging changes between headphones.

What a mess! I don't like listening to these at all but they are comfortable to wear.

My conclusion as well. There probably is something wrong with one of the drivers. I didn't measure such channel imbalance.

Here too I am glad to see my $3.- mic and cheap ass and incorrect rig isn't that far off in general and it is good to see that the treble response is quite similar to my FP. Of course, between 1k and 5k my measurements are a bit 'off' and because I do not compensate bass nearly as much as Harman bass response shows a difference as well.
The green channel seems to be the correct one. Bass should pan unnaturally to one side in your K92.

 

[Mad_Economist]

"pleased to see that cheap DIY contraption isn't far off from $ 41k official HATS."

So what's the point of the HATS? Let's see how close something like a miniDSP Ears gets to it. If it's really close maybe Amir should save a boat-load of money and get that.

I have done such comparisons before on Headphones.com as well as a comparison against an earless flat plate system here in dialogue with @solderdude

 

[crinacle]

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

View attachment 78385

Geomean:

View attachment 78386

Tyll measurements:

View attachment 78387

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.]

Here is the Stax SR-007 with SRM-313 solid state amp:

View attachment 78390

(this was the last position in the previous measurement).

The hump is still there.

Looks like sub-optimal seal to me. Staxes tend to have this problem on a lot of measurement rigs and probably some of the hardest headphones to measure for this reason (that and certain closed-back dynamic driver headphones like Focal's). I've measured a few Staxes and could only eliminate the front-volume-decoupled resonance with extra clamping force (presumably getting a more optimal but also less accurate seal in the process):

(Per usual, ignore 1k+ Hz data due to differing pinna and such)​

 

[amirm]

Amir, have you tried any iem's?

Thanks for the reminder. Here is a disaster:

My calibration is at 200 Hz and I was pulling my hair out trying to figure why it had so little output. Then I put it on my ear and it has the same problem.

Tried the other one (with obvious difference in fit):

This one in green is better. Let me try some other ones.

 

[Mad_Economist]

You are definitely not getting robust coupling with the canals, or the diaphragms of those IEMs are physically punctured.

For fitting headphones with strong seal dependency at low frequencies, continuous test signals and RTAs can be useful; I use periodic noise and an FFT, Tyll used a square wave and an o-scope, etc.

 

[amirm]

Looks like sub-optimal seal to me. Staxes tend to have this problem on a lot of measurement rigs and probably some of the hardest headphones to measure for this reason (that and certain closed-back dynamic driver headphones like Focal's). I've measured a few Staxes and could only eliminate the front-volume-decoupled resonance with extra clamping force (presumably getting a more optimal but also less accurate seal in the process):

Yeh, it has an asymmetrical pad and it rotates as well. The clamping force/earphone height is not adjustable which makes this harder.

 

[Fluffy]

This was probably already said, but all the compensated plots have a major peak around 10khz. That doesn't look right.

 

[RayDunzl]

Grins:

I stuck the UMIK-1 under the earpiece of my HD650 while it was on my head:

Pink noise

 

[crinacle]

(Btw one of my posts that was quoted by Amir got caught in a spam filter for whatever reason, feel free to delete this post once it's rectified)

 

[amirm]

You are definitely not getting robust coupling with the canals, or the diaphragms of those IEMs are physically punctured.

Ah, figured it out. Not used to sticking an IEM in "someone" else's ear. Stand by.

 

[Mad_Economist]

This was probably already said, but all the compensated plots have a major peak around 10khz. That doesn't look right.

There was something of a mis-cue regarding the compensation to apply - I'm not sure that @amirm has tried the present compensations yet.

 

[amirm]

View attachment 78408
(Btw one of my posts that was quoted by Amir got caught in a spam filter for whatever reason, feel free to delete this post once it's rectified)

Sorry about that. Your first 10 posts get heavy scrutiny.

 

[amirm]

Earstudio HE100 and @Mad_Economist latest targets:

Wow! Maybe this rig shines a lot better with IEMs!

 

[solderdude]

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. Using a much earlier measurement that varied somewhat from this, I made this EQ:

Thanks. What I am seeing is bass roll-off but don't see the midbass hump people are always whining about. Most people don't find the bass lacking in the HD650 but rather miss the 'rumble' which is below 40Hz. Also I never heard any complaints about the HD650 sounding harsh or sharp in the treble but rather the opposite that it needs a little 'lift' in the treble. +5dB as a peak at 10kHz would bring an unnatural emphasis to instruments (hyper detailed) which the HD650 isn't famous for. Instead the opposite is the most heard complaint.
A very often heard complaint of the HD600 (and lesser so the HD650) is that the HD650 is too 'forward' and has too much 'energy' around 3kHz.
I don't hear it that way though and the slight depression shown seems 'accurate' to me.

below the $ 3.- mic with cheap DIY incorrect flatplate measurement as a comparison.

There, of course, is a difference in lowest frequencies because of the Harman target.
It is interesting that crinacle's measurement of the HD650 shows a dip around 9kHz. My plot seems to be nicely in between.

It's an interesting exercise this high dollar HATS and one should expect to get a high accuracy for that amount of money.
When I EQ 30Hz +10dB one does get an impressively thumping bass. While impressive at first it would annoy me after a while.
That's a personal thing though. I personally find the amount of bass 'average' people prefer not to be the same as mine.

 

[YSC]

nice trial of the new dummy head gear! as mainly a head-fi guy I really hoped you can't resist the temptation and acquire that! but in the headphone world I am more interested in some classic ones being praised so much, e.g. HE6, HE500, HD800, AKG K701, Bayer T1 as well as the latest and greatest offerings from focal etc.

 

[solderdude]

This was probably already said, but all the compensated plots have a major peak around 10khz. That doesn't look right.

When you look at Tyll's plots you see this in all of them (except the few that have a dip there)
Crinacle's plot of the HD650 actually shows a dip there. Different compensation or different measurement ?

 

[Eetu]

This was probably already said, but all the compensated plots have a major peak around 10khz. That doesn't look right.

And the Harman computed targets all seem to have a 10khz dip. Is it accurate/representative? For example Oratory seems to use a Harman AE/OE target that has a smooth downward treble slope (without the 10k dip)?

 

[crinacle]

It is interesting that crinacle's measurement of the HD650 shows a dip around 9kHz. My plot seems to be nicely in between.
View attachment 78413

When you look at Tyll's plots you see this in all of them (except the few that have a dip there)
Crinacle's plot of the HD650 actually shows a dip there. Different compensation or different measurement ?

Do note that the IEF targets were extrapolated against data from GRAS KEMAR, the pinnae of which causes a dip around those regions (notably for the HD6X0 and the HD800/S). The raw data of my rig (which uses the EARS pinna) does not have the dip.

 

[amirm]

Here is HE100 one more with compensated curve:

So I took a shot at boosting the lows, knocking off that resonances at 170 Hz and boosting the region around 500 Hz. Here is the outcome:

Man, this is like the difference between eating frozen fish and fresh!

This IEM was one of my motivations for wanting to get into measuring headphones. Out of box it just sounds so dull and uninteresting. Lack of low and upper bass/lower mid-range seems to have been the problem.

Question is whether I can get better an IEM for $41,000!!!

 

[JohnYang1997]

Deleted the post.
This from naver ear-fi for reference.