• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Master Complaint Thread About Headphone Measurements

DualTriode

Active Member
Joined
Oct 24, 2019
Messages
250
Likes
144
I think the tested headphone positions should reflect the perhaps small differences in mounting that we might see when we put our own headphones on our heads, because you can kind of feel how they feel on your head (you can even take a quick look in the mirror to see if the earcups are too high or too low, but most of the time doing it by feel provides the same position). I think an average of those small differences in positions has got to be the most reflective. We can't do anything about the variation of our own ears to that of the mannequin head.
Hello All,

It looks to me that we need to spend quite a bit more effort in the selection and calibration of microphones used in flat plane headphone measurement. With some random in-ear microphone we will get some random calibration curve. With random calibration of the flat plane measurements we will have random comparisons with the calibrated 45CA curves.

We will not know if the measured curve changes are caused by the change of microphone, coupler and removal of the external auricle.

Perhaps little steps towards the goal. May be we keep the microphone and coupler of the fake ear and remove only the auricle.

https://www.bksv.com/en/Knowledge-center/blog/articles/sound/measurement-microphones

https://hbm.wistia.com/medias/0974h442d8



I see fun steps in learning.

First I will dig into the 45CA technology and see what microphone capsule is used in the fake ear.



Thanks DT

The 40AG microphone is used in the 45AC-9. The 40AG is a "pressure" microphone.
 
Last edited:

Robbo99999

Major Contributor
Forum Donor
Joined
Jan 23, 2020
Messages
1,446
Likes
1,018
Location
UK
Hello All,

It looks to me that we need to spend quite a bit more effort in the selection and calibration of microphones used in flat plane headphone measurement. With some random in-ear microphone we will get some random calibration curve. With random calibration of the flat plane measurements we will have random comparisons with the calibrated 45CA curves.

We will not know if the measured curve changes are caused by the change of microphone, coupler and removal of the external auricle.

Perhaps little steps towards the goal. May be we keep the microphone and coupler of the fake ear and remove only the auricle.

https://www.bksv.com/en/Knowledge-center/blog/articles/sound/measurement-microphones

https://hbm.wistia.com/medias/0974h442d8



I see fun steps in learning.

First I will dig into the 45CA technology and see what microphone capsule is used in the fake ear.



Thanks DT
I agree, I don't place much weight on user measurements of headphones with in ear microphones due to issues with inserting the microphone deep enough to reflect it being adequately close to the ear drum, and all the differences you mentioned too, plus people have a somewhat unique "Headphone" Transfer Function (how the headphone interacts with their own ears in terms of shaping the measured frequency response) on their own heads in comparison to others so it's not an accepted standard to evaluate a headphone against.
 

solderdude

Major Contributor
Joined
Jul 21, 2018
Messages
7,098
Likes
13,831
Location
The Neverlands
It looks to me that we need to spend quite a bit more effort in the selection and calibration of microphones used in flat plane headphone measurement. With some random in-ear microphone we will get some random calibration curve. With random calibration of the flat plane measurements we will have random comparisons with the calibrated 45CA curves.

We will not know if the measured curve changes are caused by the change of microphone, coupler and removal of the external auricle.

Perhaps little steps towards the goal. May be we keep the microphone and coupler of the fake ear and remove only the auricle.
Yep, I calibrated using the following method.

I suggest to use the ears and nearfield monitors in a large room.
You can even measure the speakers response at the listening position (1m away) with the speakers in front of you with a normal 'width'.
Then you put on a known reference headphone (Let me suggest HD600 or HD650) EQ that to the exact same tonal balance at the exact same SPL.
Measure the result on a test rig and compensate the resulting plot (in broad lines) to that of the used speaker and voila.
The horizontal line you then get is audibly flat on nearfield speakers (so not speakers in a listening room at several meters away).

You will find mics that are calibrated in free air behave differently on an infinite baffle and need to be compensated for this.
Slapping 'a microphone' into a baffle and assuming that what comes out of it is accurate is a flawed way of thinking.

Turned out I only had to compensate bass a few dB and reduce a treble peak from my particular mic capsule.
Copying my rig is only possible when the same materials, mic capsule (no longer available) and compensated pre-amp is used.
It does not give the same results when I put in an iMM6 with its calibration curve and lows compensation.
 
Last edited:

Dreyfus

Active Member
Joined
Feb 7, 2019
Messages
225
Likes
249
Location
Germany
Microphone calibration is a crucial topic, indeed. It presets the trend of the HF response. Depending on the sound field your treble might get boosted by up to 10 dB due to the pressure build up on-axis to the capsule.

The 60318-4 type of ear simulator as much as the flat-plate-like cheek simulators come with a microphone which is tuned to measure flat in a 180° sound field (pressure zone). Usually this is ensured by comparing the microphone with an electroacoustic actuator. In a controlled environment you could also compare the mics under FF anechoic conditions. Measuring within DF is more of a guess since we have only little certainty on the exact sound field and on-axis cancellation in front of the capsule. That is one of the reasons why I would not trust too much in DIY flat plate rigs, including my own, of course.

I once asked a local sound engineer if he could calibrate my EM 273 mics to the pressure zone response within his actuator. He asked for 180 Euros per piece. That is seven times the price I would pay for a simple 0° calibration. :rolleyes:
I am not mad enough, yet ...

@Robbo99999:
The HE4XX measures quite even because the pressure and seal is better than average on my head. Some other headphones (thinking of you, AKG!) might introduce larger errors. Also, according to Hammershøi and Møller, the blocked ear canal reponse adds less HRTF (here: PRTF) variance than you would see measuring with the open canal or at the actual drum reference point. There will be more interactions once you include a human ear canal. See @amirm's experiences measuring with the B&K 5128 which is closer to the human anatomy and its imperfections than a standard ear and cheeck simulator, not to mention the simple flat plate rig.

By the way: one thing to note is that the design of the ear simulator cavity and especially the anthropometric ear canal add to the diffusion of the sound field. A flat plate rig that places the capsule right in front of the driver will see less random incidence that would be produced by an asymmetrical ear. That is also why the flat plate is less dependend on the exact position of the headphone on the fixture, of course. As already mentioned, it has less reactivity in the sound path.
Ideally, this should also be accounted for when calibrating the mic. My guess is that the flat plate calibration asks for a relatively high treble attenuation due to the more dominant effect of on-axis phase superimposition. This effect will probably vary with the design of the headphone however. Both the shape of the headphone housing (inner cavity) guiding the waves and the angle of the driver can change the directivity. That is why I think measuring with a pinna is a mandatory element when gathering a larger group of comparative data. Headphones are designed to work in diffuse, asymmetrical sound fields.
 
Last edited:

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
10,837
Likes
10,311
Location
Riverview FL
Suggestion:

@amirm

Dig into your junk box and measure a pair of the no-name ear buds or IEM or whatever that came with something bought in the past, sometime when you have nothing better to do.

I tried the ones that came with my Moto phone a few days ago (after a year), and they were awful. No highs. I can't hear highs anyway so they must have really been bad.

(My HD650 sounded fine but not real loud plugged into the phone)

Just for grins, of course.

Or, don't.
 

DualTriode

Active Member
Joined
Oct 24, 2019
Messages
250
Likes
144
@solderdude,

I liked where you started with the near-field monitor speakers and matching the headphone frequency response to the speakers. You lost me with the multi-color flat line plot. That is not calibration that is marketing. The multi-color flat line plot has nothing to do with calibrated Sound Pressure Level. It is a transfer function sort of. At a visual level it works well at another level it is missing real calibrated SPL numbers.

@Dreyfus,

You are intent on circling back around to human anatomy. I agree that anatomy is important. I will also say that anatomy is largely variable. Even if we use a standard fake ear it will add peaks and dips to our high frequency measurements that will confuse the measurement and comparison of the headphone hardware. There is much to measure prior to adding 39 types of ear into the mix.

One place where you and I appear to have a completely different view is if headphones operate in a Diffuse-Field or Pressure-field.

My view is that speakers operate in a Diffuse-Field. Headphones are designed to operate in a pressure-field and sound as much as possible like speakers in a Diffuse-Field. This is a difficult challenge.

Headphones couple directly to your head, a good seal is required. If the seal leaks, much of the bass Sound Pressure Level leaks with it. The headphone cup is a small cavity in relation to the frequency wavelength. This is the definition of “Pressure-field” not Diffuse-Field.

“Pressure-field” SPL measurements are much less dependent on external ear anatomy than Diffuse-Field measurements.



Thanks DT

Read about Pressure-Field and Diffuse-Field.

https://www.bksv.com/en/Knowledge-center/blog/articles/sound/measurement-microphones
 

solderdude

Major Contributor
Joined
Jul 21, 2018
Messages
7,098
Likes
13,831
Location
The Neverlands
You lost me with the multi-color flat line plot. That is not calibration that is marketing. The multi-color flat line plot has nothing to do with calibrated Sound Pressure Level. It is a transfer function sort of. At a visual level it works well at another level it is missing real calibrated SPL numbers.
Can I ask you what you are on about... There is no reference to any multicolor flat line plot anywhere in my calibration procedure.
I merely outlined why shoving a mic in a flat piece of wood is NOT going to give accurate results and, just like HATS manufacturers have to make define their 'corrections' it is also essential for a flatbed rig. Just like an in ear mic or test rig like the EARS also needs different calibration and why HATS from different manufacturers or different configurations also need different and individual corrections. A mic in a piece of wood or whatever material is not different.
Of course the measured level is also calibrated (give or take a dB or so) but that's another story and easy to do when you have an SPL meter lying around.

Regarding the 'marketting' plot.... I assume you mean this drawing:

It has absolutely nothing to do with any calibration nor exact levels. It merely is an indicator of what 'sonic' effect a bump or dip has on an overall sound signature. The severity of the effect is determined by the how much dB variation there is and one can easily have combinations or much wider bumps or dips. It also is about 'audibly flat' being represented by a horizontal line. My measurements all are 'audibly flat' when the line is horizontal, see it as a target at 'normal' listening levels to me.

As mentioned... it has absolutely nothing to do with any calibration procedures.
 

Dreyfus

Active Member
Joined
Feb 7, 2019
Messages
225
Likes
249
Location
Germany
You are intent on circling back around to human anatomy. I agree that anatomy is important. I will also say that anatomy is largely variable. Even if we use a standard fake ear it will add peaks and dips to our high frequency measurements that will confuse the measurement and comparison of the headphone hardware. There is much to measure prior to adding 39 types of ear into the mix.
Well, then let's just agree that we have different philosophies about headphone measurement.
I think that everything up to this point has been sufficiently elaborated. The reader and operator shall decide which route to follow.

This is the definition of “Pressure-field” not Diffuse-Field.
My definition of a pressure field configuration is a microphone being flush mounted with a boundary layer which exposes the diaphragm of a pressure microphone (omni) to a 180° sound field.

Here is the definition shared by GRAS:
A pressure microphone is for measuring the actual sound pressure on the surface of the microphone’s diaphragm. A typical application is in the measurement of sound pressure in a closed coupler [here: ear simulators] or the measurement of sound pressure at a boundary or wall [here: flat plate rig]; in which case the microphone forms part of the wall and measures the sound pressure on the wall itself.
Side note: The flush mounted variant is also being referenced as "Boundary Layer Microphone" (BLM) and should not be confused with Crown's concept of the "Pressure Zone Microphone" (PZM) which is a 90° configuration applied onto the surface instead of flush mounting into the surface.

One place where you and I appear to have a completely different view is if headphones operate in a Diffuse-Field or Pressure-field.
[...]
“Pressure-field” SPL measurements are much less dependent on external ear anatomy than Diffuse-Field measurements.
A diffuse field mic configuration where the capsule is exposed to theoretically 360° of sound has nothing to do with headphone measurement per se. As already mentioned, I am talking about pressure field configurations where the mic is mounted flush into the surface or into an ear simulator.

All of that does not mean that there cannot be diffuse sound incuded by reflections inside of the ear cup, though. That was the point I was refering to considering the pressure add up in front of the capsule for the 0° incidence versus "random incidence" up to +/- 90°.

Admittedly, I have to correct one aspect:
The pressure add up increasing the higher frequency response does only apply for a mic in the free field. As soon as the mic is mounted flush into the boundary the critical frequency is lowered due to the increase of the boundary layer's surface which acts like an enlargement of the front of the capsule.

Pressure doubling formula:
f = c / d

For a 10 mm capsule that would be around 34.3 kHz for the full wavelength and around 8.6 kHz for the 1/4 wavelength.
If we now increase the area to say 100 mm we would end up with 3,4 kHz (full) and 857 Hz (1/4). A 150 mm plate would end up with 2,3 kHz (full) and 572 Hz (1/4). And so on ...

The formula is expecting +3 dB for the 1/4 wavelength and +6 dB for the full wavelength in a 0° incidence where the incomming and the reflected wave overlap with each other. In theory, everything above the calculated frequency would be boosted by up to 6 dB. In some cases (without proper damping of the front grid) the effect is said to take even up to 10 dB initially and then drop to 6 dB calculated above.

I am not sure by how much this effect has to be addressed with the larger surfaces in a headphone measurement situation. I assume that it largely depends on the directivity of the excited sound field where a lot of on-axis (0° +/- a few) resonance could raise the level up from the calculated frequency whereas a lot of off-axis (diffuse sound field up to +/- 90° on a flat plate) might mask the effect. On the other hand, headphone drivers are not working as point sources and add a lot of foam on the side damping such resonances. So there might be different rules for this specific case.

Maybe our colleague @Mad_Economist has any info on that topic for headphone measurement situations on flat plate rigs?
 
Last edited:

DualTriode

Active Member
Joined
Oct 24, 2019
Messages
250
Likes
144
@Drayfus,

Sifting through your posted words I am seeing words like, "Headphones are designed to work in diffuse, asymmetrical sound fields." less often. I am seeing that words like pressure, pressure microphone, and pressure zone are gaining frequency.

I like it, greater accuracy.

DT
 
Joined
Feb 15, 2019
Messages
25
Likes
22
Review more content creation headphones, because this will have a profound on effect on the end product, regardless of driver tech.
There's not that many models on the market suitable for the task, a fairly representative selection of avalaible models would be: Shure SRH440, AKG K371, Sennheiser PXC550-II and NAD Viso HP70.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
30,297
Likes
89,383
Location
Seattle Area
Thread Starter #151
@amirm

Dig into your junk box and measure a pair of the no-name ear buds or IEM or whatever that came with something bought in the past, sometime when you have nothing better to do.

I tried the ones that came with my Moto phone a few days ago (after a year), and they were awful. No highs. I can't hear highs anyway so they must have really been bad.
I plan on these these but first need to get my head above ground with respect to all the on/around ear headphones that are here/been sent. FYI I have some super expensive IEMs here as well.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
30,297
Likes
89,383
Location
Seattle Area
Thread Starter #152
There's not that many models on the market suitable for the task, a fairly representative selection of avalaible models would be: Shure SRH440, AKG K371, Sennheiser PXC550-II and NAD Viso HP70.
I have the K371 and will test that soon. Don't have the others. If members have them and want to loan them to me, I would love to measure them.
 

Dreyfus

Active Member
Joined
Feb 7, 2019
Messages
225
Likes
249
Location
Germany
Take such complaints to the complaint thread: https://www.audiosciencereview.com/...adphone-measurements.18451/page-8#post-627138

I have issued a reply ban for you as your comments are derailing the thread per our new guidelines.
I'm sorry?

You asked me to try the EQ. I replied by commenting that it doesn't work for me and shared my own one as much as my personal listening impression.

Then you reply with a derogative comment that ignores both the preceding points of the discussion as much as my observations about the X2HR in a measured vs perceived manner. Then I get a ban from the thread.

I agree that parts of the discussion would have better belonged to this thread. Yet, your interaction on the thematic level is really unprofessional. You selectively discard diverse, subjective opinions on devices which are made for listening to music. Yet, you share your own listening impressions and stress that they match well with what yout measured with your GRAS equipment. So as it looks, you appear to accept and have an idea of a correlation between measurements and subjective listening evaluations. Still, you appear to ignore my words which - first of all - claim to keep an open balance between "isolated" measurements on test fixtures and subjective listening instead of solely and dogmatically relying on just one of them.

I am sorry for your reluctance. Maybe there is something lost in translation (english is not my native language obviously). But I feel that you are not willing to address my points in a serious conversation. If that is the point, I sure can leave and lower your perceived SNR in the threads.

As for the matter itself, there is no objectivity in subjectivity. Hence, a subjective impression cannot be raised to a common standard. Still, subjectivity is the gate through which we look into the world. Even when interpreting measurements. It's all human imperfection, no matter what you try. And I see no reason why a listener should distrust his own perception in favour of a generalized measurement which can be just as deceptive and spontaneous at times.
As I have said before: There is justification for both. So why not treat them equally? With care and conscious knowledge of the limits, but still equated.

I see your points for objectivity. But I don't share the thinking that the methods you apply offer enough certainty so that we could just look at graphs and numbers and judge over a product without actually listening to it. Headphone measurement was never that easy. And it will never be, as long as they are used by human beings with real ears and intuitive perceptions.
 
Joined
Apr 1, 2018
Messages
63
Likes
101
Location
Seattle
@amirm Have you re-measured the same suite of measurements after using your recommended EQ? I'd be interested to see if they cause distortion or clipping after EQ is applied. i didnt see these re-measurements occur in the few threads I've looked at. Would also be interesting to see how the FRs look after your EQ too.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
30,297
Likes
89,383
Location
Seattle Area
Thread Starter #156
@amirm Have you re-measured the same suite of measurements after using your recommended EQ? I'd be interested to see if they cause distortion or clipping after EQ is applied.
I did it for one headphone. Can do it for others but it just tells us what we can already predict from looking at the current distortion graphs. It requires rewiring and new setup to run the post EQ test so if is not needed, I rather not run it. :)
 

DualTriode

Active Member
Joined
Oct 24, 2019
Messages
250
Likes
144
I did it for one headphone. Can do it for others but it just tells us what we can already predict from looking at the current distortion graphs. It requires rewiring and new setup to run the post EQ test so if is not needed, I rather not run it. :)
@amirm,

I am attempting to sort this stuff out too.

It makes sense to me that you do a few tests off line and keep the results to yourself. Share online if the results are significant.

Assuming that 2%, 3% or 5% THD may not be audible, why report it?

On the other hand if a particular frequency band has 2% or 3% THD on the initial test and you boost / equalize that that frequency band by 6dB or 8dB there may be a significant bump in measured THD, even clipping, inquiring minds may want to want to know.

Thanks DT
 

solderdude

Major Contributor
Joined
Jul 21, 2018
Messages
7,098
Likes
13,831
Location
The Neverlands
@amirm
I noticed a few times you found it difficult/time consuming to position headphones.
Could it be an idea to either use some white noise and running a real time spectrum analyzer to view proper seating ?
Or as an alternative method listen to the output of the HATS while playing bassy music on the DUT ?
Maybe this could speed up seating difficult headphones a bit faster ?
 

pkane

Major Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
1,726
Likes
2,645
Location
North-East
@amirm
I noticed a few times you found it difficult/time consuming to position headphones.
Could it be an idea to either use some white noise and running a real time spectrum analyzer to view proper seating ?
Or as an alternative method listen to the output of the HATS while playing bassy music on the DUT ?
Maybe this could speed up seating difficult headphones a bit faster ?
White noise with an RTA view in REW is what I use to find the best/repeatable position (on a flat plate). Easy to see the difference as the headphones are shifted a bit a time, and I usually chose the position where the bass response is best, and higher frequency dips not as sharp or deep.

Here's a difference between two positions on the plate, about 0.5cm apart horizontally, with HE560 (red was the position I went with):
1610492336334.png
 
Last edited:

solderdude

Major Contributor
Joined
Jul 21, 2018
Messages
7,098
Likes
13,831
Location
The Neverlands
It seems rather averaged/smoothed for a white noise plot.
I meant using white noise while positioning and using a real-time analyzer.
Have made some BT headphone measurements using white noise (generated with a noise generator flat up to 100kHz, really easy to build)
I have to record at least 1min of noise and then analyze it to get something close to a sweep.

 

Similar threads

Top Bottom