Approaching Head Measurements

After Audeze clued me in to how much the different headphone measurement heads vary in measurement performance, I decided to undertake a little experiment of my own. Well...not all my own, I'd love for others to join in! Here's what I'm going to do:

I have been in contact with G.R.A.S. and B&K and will be receiving a loan of an appropriate measurement head from both companies. Exact timing isn't known yet, but it's a matter of a couple months at most.

I will be measuring five headphones on each system. I've picked circumaural headphones both sealed and open because I didn't want to deal with the variables of seal with on-ear types. They are: Audeze LCD-3F; NAD VISO HP50; Mr. Speakers Ether C; and Sennheiser's HD 600 and HD 800 (stock).

I will be measuring all five headphones on the G.R.A.S., B&K, and Head Acoustics heads, and then will spend some time comparing the data. Here's the part where you may come in:

On a couple of instances in the past I've had folks volunteer to give me a hand with complex math problems. Analyzing this data may very well be a complex math task. I'm half way decent with a spreadsheet...but I know darned well there are many InnerFidelity readers who are far more qualified than I regarding data analysis. So, I'm going to open up the process of analyzing the data to InnerFidelity readers. Those who choose to play around with the data will be able to post their thoughts in the comments when I write the final article; but some of you who are very expert and would like me to consider publishing your analysis within the final post are welcome to contact me privately to make arraignments. (tyll(at)tyllhertsens(dot)com)

I think I will be mostly looking at the differences in the raw frequency response curves. I'll likely do something like subtract the HD 800 from the HD 600 on each head and then look at how the differences differ from head-to-head. But readers are more than welcome to analyze the data in any way you see fit.

The purpose of this exercise is to determine:

  • What are the differences between measurement heads?
  • Are there any head specific gross measurement artifacts?
  • Are the heads equally sensitive to changes in the treble areas with changing headphones?

I just ran the five headphones through my test sequence for some fresh measurements. Excel spreadsheets for the five headphones measure on my Head Acoustics head can be downloaded here:

Audeze LCD-3F
Mr. Speakers Ether C
Sennheiser HD 600
Sennheiser HD 800

I will make the spreadsheets available for the other heads as soon as I complete the tests. Once both are tested it will take me a couple of weeks to digest the data and be ready to publish results, please contact me no later than one week after the second set of measurements are made if you'd like your analysis published. No guarantees here, I'll need to evaluate your submissions prior to agreeing to publish.

For those of you not interested/qualified in the data analysis, you may still want to download some of the spreadsheets above and have a peak at what they look like. This is my standard format for all headphone measurements.

These spreadsheets are copyright The Enthusiast Network 2016.

Feel free to ask questions, and make comments and suggestions in the comments below.

tony's picture

Big Sound revealed DACs and Cabling ( at least for me ) and that the HD800 continues to "float" up with the finest headphones in comparable performance levels. ( it always sounded outstanding to me )

Somehow, outa all this research, something un-expected will be uncovered, probably something important.

I'm stay'n tuned

Tony in Michigan

ps. I'd bet that you're taking this headphone thing further than even the US Military

mat's picture

Have you thought about including one or two universal fit IEMs in your test group? Presumably they would eliminate any effects the pinna and concha geometry may have on the circumaurals and isolate the sound signature differences effected by the various inner ear models. I think it would be interesting to see how they differ from the full sized cans, but I suspect you may have left them out for the same reason as the on-ear headphones.

Tyll Hertsens's picture
Lotta people asking for this in the comments. I think I might give a try to the new Etymotic ER4SR as they're pretty good at making IEMs that measure predictably...if you can get them in the ear canals well.
francisk's picture

I'm "eargerly" looking forward to your measurement of both new Etymotic IEMs, the ER4SR and the ER4XR :-)) Can't wait to hear how much better these 2 new Etymotic IEMs sound like. I've been a big fan of the ER4S and ER4B paired with the Etymotic red filter and grey triflange tips for some time now. My current favorite is the ER4B with Etymotic red filters and grey triflange tips which gives me a better high frequency detail and attenuates the 2-4kHz honkiness of the ER4S. Your measurement of the new ER4SR and ER4XR will greatly help many Etymotic fans like me to determine whether to upgrade to the latest models...until we get to test it ourselves, hopefully soon. Keep up the fantastic IEMs measurement Tyll, the best ever headphones and IEMs measurement website I've ever come across on the internet or anywhere else for that matter.

castleofargh's picture

egoistic suggestion:
would it be possible to get one or 2 of those headphones going on a tour at the end of the this affair?
send them to people who have a humble little measurement rig, so that they can if they so chose, align on one of your results, and use the headphone and the measurement to calibrate their own rig.

personally I'm more into IEMs so I would love it if you included a pair of IEM(or even better, one vented, one sealed), with maybe some information about the insertion depth. IDK how this is defined? distance to the coupler I guess?

back on topic, I can't help much, aside from fooling around with room eq wizard, I'm really a good for nothing guy. but definitely interested to follow this topic.

Tyll Hertsens's picture
Sending a few of the cans around would certainly be a possibility. Once we get to the end bring it up again and we'll see if there's enough interest.
Hjelmevold's picture

I have a Neumann KU100 "Fritz" dummy head mic as well as the HP50 and also an Ether C on its way, so if you like I can probably do some measurements of these two headphones using RoomEQWizard. Since the KU100 is a dummy head mic that Neumann specifically advertises as diffuse-field equalized, perhaps such measurements may be of interest?

Tyll Hertsens's picture
Hmm...HeadRoom has a KU100 as well. I could probably do some measurements with it...why not, eh? I'll look into it.
zobel's picture

Tyll, will you keep the same SPL at a certain frequency, or with white or pink noise, for every test of every headphone on each head?
Don't you think that would be far more relevant that applying the same voltage signal to them?

If you do a complete set of measurements that you normally do, on each head with all the cans you are testing, wouldn't it make sense to keep the same power level that produced equal SPLs, for all tests on each headphone, on each head?

Do you think pink noise would be sufficient to balance up the outputs?

Tyll Hertsens's picture
Yes, absolutely. I will test all the heads at the same levels. Will probably just use a 500Hz test tone.
ultrabike's picture

for it :)

Onix's picture

Till, are measuring heads similar to those binaural microphones shaped like heads in regard of materials?

tinyaudio's picture

Quite similar, but much more detailed.

binaural microphones like the KU100 (or older versions of the Neumann dummy head, like the KU80 and KU81) usually have "hard" ears, and the exact position of the microphone inside the ear varies.
This means that the ears will not change their shape when a headphone sits on them, and for more shallow headphones this means that there will be no seal since the hard ear will not allow the headphone to come close enough to the head to form a seal.
The exact position of the microphones in the ear canal also influence the quarter-wavelength-resonator that is formed by the ear canal. Not very important for microphones, but VERY important when you want to measure a source that is very close to the ear (such as headphones).
Another thing that is often overlooked in binaural microphones is acoustic impedance of the ear.
There's a lot to consider for that, and it's the reason why industry standards such as the 711-coupler are mechanically very complex, because they have to simulate the exact compliance of the eardrum, to exactly replicate the load that the ear poses for the headphone.
For binaural microphones most of that is not relevant since the sound source is mostly quite far away from the dummy head (at least >1m).

So in short:
Measurement dummy heads are very much like binaural microphones, but MUCH more intricate and detailed.
Nevertheless, measurement heads would work perfectly as binaural microphones, but they're also much more expensive :)

ivanhagberg's picture

Hi Tyll!

Would it not be interesting for you as well as for the rest of us, now when you are involved in setting up a measurement system - that you meanwhile do some purely SUBJECTIVE reviews of some well reputed new head phones? This way the reviews will not come to a halt during this measurements setup and the measurements could be added once the setup is finished and tested.

I prefer you subjective perspectives and comparisons of the headphones you are reviewing. And afterwards the correlation between subjective and objective would be VERY interesting to find out - both for your self and for us. What do you think?


Tyll Hertsens's picture
I've been trying to do a headphone every week...though it goes longer sometimes. I'll continue that schedule during the head measurement process. Expect reviews on Fridays; measurement updates on Tuesdays.
logscool's picture

It seems to me there are not really sufficient data points to create a meaningful waterfall plot from your impulse measurements. Is that all of the data that you capture for that test, does the data get downsampled or something when it goes into the spreadsheet or do you just set the sampling rate of the AP to get 512 samples from each test over the desired length? Maybe I am missing something.

I'll definitely be interested in seeing the data, I think the key here isn't so much whether the DF compensation is consistent or the Harman Target can be applied successfully. What is important is that the variations are consistent and can be interpreted linearly so that you could calibrate two systems to each other, apply a compensation to one of them and get them to display the same measurement with a different headphone.

ADU's picture

5 headphones is a small number to draw conclusions from. And the HD800 doesn't have the most consistent response in the treble for a control headphone,... though it is one that most audiophiles are pretty familiar with.

logscool's picture

I think it is the sort of sharp peaks that we see in the HD800 that will make it a good test candidate and something that will be more likely to show the non-linearity in relationship between couplers. Plus it is a headphone that is well regarded and has been proven to measure incredibly consistently from one headphone to another.

idesign's picture

Hi Tyll, Do you have plans to do any new amplifier/DAC measurements and tests? There are a lot popular amps on the market that have not been covered on Innerfiedleity ranging from the HDVD800, Hugo TT, Zana Deux Super, Black Widow Mk2, DNA Stratus, Gungnir+Mjolnir 2, Yggdrasil+Ragnarok etc. Would like to see more amp/DAC measurements and reviews on the site. The Wall of Fame has a lot of older amps.

bobusn's picture

Tyll, keep up the great work!

Matzio's picture

Hi Tyll,

Great idea. Thanks a lot for your hard work.

How about adding the NightHawks?

AudioQuest used a G.R.A.S. HATS, too, the 45BB KEMAR. And they published some (simplified) measurements.

Would be a good starting point to compare. And to prove AudioQuest's key assumption about the influence of incomparable transfer functions of HATS, the general validity of ISO-standard DF/FF equalization curves and the uncertainty of headphone measurements above 8 kHz.