• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

DIY Headphone Measurements - Ear Simulation?

Dreyfus

Active Member
Joined
Feb 7, 2019
Messages
247
Likes
296
Location
Germany
Hello everyone,

I've been planning to build my own dummy head for quite a long time now and am gradually comming closer to actually building the construct. The idea is to build a binaural microphone I can use for over ear headphone measurements (no IEMs) as much as for field recording.

The base will consist of a plastic mannequin head that is filled with compressed absorbing foam. As for the mics I will use some Primo EM172 electret capsules which are quite cheap and come with a great signal to noise and sensitivity. Linearity is decent for the price, as well. Maybe not the end of the road, but probably good enough for a hobby rig.

And then the heart of the matter: ear simulation.
I have finally found a pair of silicon ears which is somewhat average in size, realistic in shape and not too stiff. This should allow me to capture some generic pinna inteferences and directional bands. So far so good.
But what about the ear canal? Using pinna only, recording the sound somewhere in the concha cave, there will be no ear canal resonances which shape the timbre and are naturally processed by our brain trying to accomplish tonal balance.

There is a a lot of buzz around the net whether it is practically beneficial or not to model the ear canal.
Some say you can only get meaningful results when working with a state-of-the-art coupler by B&K or GRAS. Some say you can just use some generic 711 couplers from Taobao. Some say you just have to attach a plastic tube (18 to 25 mm long) to the ear to cover the fundamental resonances. And some say you don't need to worry about that at all because the canal is anatomically random anyway, shifting the resonant frequency in a unpredictable fashion.

After all, there is probably a piece of truth to all of them. But I wonder which of those options would fit best to my project.
To make it clear at this point: of course I can't reinvent the wheel or fit IEC standards and such with this rig. And I won't be so naive to expect any reliable correlation between the measured frequency response - depending on the impedance of the capsules, the flange and the coupling, the (here: random) shape of pinna and ear canal ecetera - and my personal perception. But I want try to account for the main requirements and difficulties for over ear measurements and expect a relative basis of comparison to judge overall tonal balance, changes with pads or varying case damping.

I know that there is no perfect standard, that even the recognized flagship couplers have their flaws and that there is also a lot of technical politics involved. And the more I read, the more I'm worrying about the usefulness of this rig. But I still do have hope that there is a practical solution that is good enough for the purpose and allows somewhat meaningful and repeatable results.

What are your thoughts?


Generic 711 coupler from Taobao? Plastic tube? Or just going completely without the ear canal?

I would welcome your response :)

Best wishes,
P48
 

PaulD

Senior Member
Joined
Jul 22, 2018
Messages
453
Likes
1,340
Location
Other
P48, you might be better off asking this on the Micbuilders group, it used to be on Yahoo. There are experienced mic designers and builders on it. https://groups.io/g/MicBuilders
 
Last edited:
OP
Dreyfus

Dreyfus

Active Member
Joined
Feb 7, 2019
Messages
247
Likes
296
Location
Germany
Thanks for your replies!

Custom canal:
...
Definitely not DIY.
I already considered to get my ear canals casted by an audiologist. But then I would probably have to create the custom molds and cast negatives made of silicone myself ... which is in an art of its own. On the other hand, having them made in the audiologist's lab might be too expensive.
The next question would be: How to join the silicon pinna (which is even harder to custom mold) and the ear canal together?

There would be a lot of things to master.

P48, you might be better off asking this on the Micbuilders group, it used to be on Yahoo. There are experienced mic designers and builders on it. https://groups.io/g/MicBuilders
Thanks for the tip!
I'm actually a user of that group. And I don't whether there are that much users who have experiences with ear simulators and couplers. But I might give it a try.
 
Last edited:
OP
Dreyfus

Dreyfus

Active Member
Joined
Feb 7, 2019
Messages
247
Likes
296
Location
Germany
There are quite inexpensive DIY ear mold kits out there, though I don't think they go as deep as a professional impression.
Is it for casting negatives or positives?

I do need the negative, the "outter shell" of course.
 
Last edited:

NTomokawa

Addicted to Fun and Learning
Forum Donor
Joined
Jan 14, 2019
Messages
779
Likes
1,334
Location
Canada
This is one of the many that I've seen. The price usually doesn't go above that.
 

Wombat

Master Contributor
Joined
Nov 5, 2017
Messages
6,722
Likes
6,459
Location
Australia
OP
Dreyfus

Dreyfus

Active Member
Joined
Feb 7, 2019
Messages
247
Likes
296
Location
Germany
Yes.
And I'm also aware of its issues. :rolleyes:

10237970.png

https://www.head-fi.org/threads/audio-measurements-on-a-headfi-budget.893084/

Even when you accept that the system is nowhere near an IEC standard due to the lack of an ear canal, there would still be the stiff pinna (hard silicone which does not bend with pressure) as a major flaw. I also dislike the fact that they offer no mic specs.

Thanks for the mention, though!
 

Earfonia

Active Member
Joined
Feb 21, 2019
Messages
278
Likes
398
Location
Singapore
But what about the ear canal? Using pinna only, recording the sound somewhere in the concha cave, there will be no ear canal resonances which shape the timbre and are naturally processed by our brain trying to accomplish tonal balance.

There is a a lot of buzz around the net whether it is practically beneficial or not to model the ear canal.
Some say you can only get meaningful results when working with a state-of-the-art coupler by B&K or GRAS. Some say you can just use some generic 711 couplers from Taobao. Some say you just have to attach a plastic tube (18 to 25 mm long) to the ear to cover the fundamental resonances. And some say you don't need to worry about that at all because the canal is anatomically random anyway, shifting the resonant frequency in a unpredictable fashion.

After all, there is probably a piece of truth to all of them. But I wonder which of those options would fit best to my project.

What are your thoughts?
P48

My opinion, NO Ear Canal for Binaural recording.
Reason, when ear canal resonance included in the recording, when played back using either In-Ear Monitor or headphones another ear canal resonance will be developed, therefore the ear canal resonance from the dummy head and the ear canal from listener human ear will mixed causing unnatural tonality.

I think for binaural recording, the best position for the microphone should be around the position of the In-Ear monitor nozzle that is inserted into the ear canal. So just at the entrance of the ear canal. Mic position = Earphone nozzle position. Or Mic position = Headphone driver position. Depending on the means to playback the recording. My2cents.
 
OP
Dreyfus

Dreyfus

Active Member
Joined
Feb 7, 2019
Messages
247
Likes
296
Location
Germany
Hey Earfonia,

the issue of impedance doubling when playing back binaural recordings can be minimized by equalization. You could record a reference signal - let's say pink noise emitted by a flat speaker at 0° azimuth - then play it back with a pair of headphones and tune an individual correction curve by equal loudness matching.


When done right, this procedure can restore the correct timbre and spatial cues.

However, there will always be some errors when there is a certain physical mismatch between your head and the dummy head, of course. You might get the correct frontal localization following the procedure above. But as soon as the recorded sound source starts moving, the resonances might differ from your own HRTF and therefore produce false coloration, leading to confusion and in-head-localization.
This is the situation for pretty much all binaural recordings made with dummy heads. And you are probably right recommending to skip the ear canal resonances for such recordings because there might be less physical inconsistencies.

The only way to overcome or at least substantially minimize such issues would be to create a realistic replica of your own head and tune it towards your own perception of sound. And in fact, I have already tried to mold a pair of silicon ears to start with. I got some impressions of my own pinnas with roughly three-fourths of the canals made by a local audiologist:

20200310_171432.jpg

But as it turned out, the compound used was way too viscous which lead to some inaccuracies and spots with trapped air around the pinna. I tried to cast it with fine alabaster and rework the shape. But that didn't go so well ... back to the drawing board. :facepalm:

Maybe I will try to repeat the process with some more fluid alginate some day. Until then I will concentrate on tuning the system for headphone measurements with a "generic" pinna I got from POP Microphones:

ears_02.png
ears_03.png

As for middle ear simulation I could not resist and pulled the trigger on a pair of 60318-4 couplers in January.
I designed an adapter which will attach the coupler to the silicon ear:

lfCGRWn.png
r3jW1Kg.png

There is a canal extension of roughly 7 mm between the coupler and the ear canal entrance point of the silicon ear. The canal extension ends in a conical shape (see second picture) to fit the end of the silicone concha.
I hope that the canal extension won't introduce any nasty resonances. Otherwise I have to shorten the gap between the coupler and the ear canal entrance.

Next step will be printing the adapters and gluing them to the back of the ears. Then I will try to fit the whole thing to a dummy head ... or design a base unit with two contact surfaces for the ear pads.

The mics will probably be some Line Audio OM1. I will see if I can remove the metal housing and fit the capsule with some isolation tape wrapped around directly into the coupler. In case that doesn't work I could still try a matched pair of the smaller Primo EM258.

To be continued. ;)

Regards
P48
 
Last edited:

Earfonia

Active Member
Joined
Feb 21, 2019
Messages
278
Likes
398
Location
Singapore
As for middle ear simulation I could not resist and pulled the trigger on a pair of 60318-4 couplers in January.
I designed an adapter which will attach the coupler to the silicon ear:

View attachment 56408
View attachment 56409

To be continued. ;)

Regards
P48

Whoa 2x 60318-4 couplers :D I wonder how much that cost ;)
I have one China plinth 60318-4 coupler for IEM measurement, and recently bought another one, only coupler without mic (~ $90) for testing other measurement mics with it.

Few years back I wanted to start similar binaural recording projects, bought 4x matched EM172 and 4x matched EM258. Tried a pair of EM258 in simple configuration, sounds great. I did once recorded a full recital for about an hour with binaural mic in my ears. I found for orchestral recording it doesn't sound very interesting, because there is not much movement of the sound source. We don't need binaural for such setup. The same event I recorded with an ORTF microphones, and it sounds subjectively better IMHO. I also have a pair of MS-TFB-2 binaural mics from Sound Professional that looks like using a pair of EM258 as well. But the project rather abandoned now due to lack of time. So I'm interested to see the progress and result of your project.

Here is EM258 in foam ear tip inserted deep into my ear canal:

20180414_Right_s.jpg20180414_Left_s.jpg
 
OP
Dreyfus

Dreyfus

Active Member
Joined
Feb 7, 2019
Messages
247
Likes
296
Location
Germany
No worries, just the cheaper ones from Taobao! I'm not mad enough to waste money for the official IEC fan gear, yet. :D

May I ask how you wire your capsules up? I prefer the Simple P48 circuit for its simplicity and practical performance. That way I can use my Babyface Pro for the recordings / measurements.

I will see what fits best to the couplers. So far I think I will purchase a pair for Line Audio OM1 which come with hand-selected capsules and draw a pretty flat response:

OM1plot[1].png

Line Audio OM1 - Specifications

That gives me some more confidence than working with bulk stock capsules like the Primo EM172 / EM258 or the PUI 5024HD.

---

By the way, here is my attempt of EM258 in ear measurement:

EM258_in_ear.png

The capsule is glued with Suguru to a foam tip. The cable has to bend a bit when inserting the piece into the ear canal. Otherwise it works very well.
In Ear Microphone Test - EM258
 
Last edited:

Earfonia

Active Member
Joined
Feb 21, 2019
Messages
278
Likes
398
Location
Singapore
So far I've been using several Zoom portable recorder with their 3.5mm mic input. I have Zoom H2n and H4n Pro. Ever used my friend H6 as well to record a concert with both ORTF mic and binaural mic at the same time. I use Zoom devices for portability since they are battery powered.

I have RME Babyface Pro as well with 2x Rode VXLR+ that I usually use for the IEC coupler, measuring IEMs. Probably one day I diy binaural mic with separate left and right cable to plug to the 2x VXLR+.

Speaking of matching the capsules, I made this coupler from heatshrink tube + Primo EM172. It was experimental coupler to measure IEMs, but I found I could get consistent result, therefore I could use that coupler to check the EM172 matching characteristic. Kind of a tool to hand match the capsules. The EM172 is tightly fitted in the coupler with the 10mm Rubber Holder from Micbooster. So it is removeable.

20191114_082357s.jpg20191114_082458_tn.jpg

The nozzle is sized to fit medium size IEM ear tips, to couple IEM driver with the EM172.

Not sure if such heat shrink tube coupler can be useful for the binaural project. Maybe to experiment with the size and length of the ear canal.

Some experimental IEM measurement with DIY heat shrink tube. Around 4-5 layers of heat shrink tube to make a sturdy and stiff coupler.
20191113_211916_tn.jpg
 

Earfonia

Active Member
Joined
Feb 21, 2019
Messages
278
Likes
398
Location
Singapore
OP
Dreyfus

Dreyfus

Active Member
Joined
Feb 7, 2019
Messages
247
Likes
296
Location
Germany
Hi Earfonia,
sorry for my late response! Was quite busy the past few days.

The Rode VXLR+ is a handy tool for adapting electret microphones to phantom power. I have build a similar adapter with the following circuitry: http://www.jp137.com/lts/XLR.PIP.pdf
Building the adapter myself allowed me to tune the load resistance. This gave me the chance to apply a higher drain-source voltage which increases the mic's sensitivity.

Using the coupler for checking the mic's FR is a clever idea! I might try that in the future.
Also interesting to see your IEM measurement equipment. I think I have seen your gear on Head-Fi before. :D


As for my headphone measurement rig, I fear that the nozzle of the Line Audio OM1 will not fit the 60318-4 coupler. Even when I compensate for the conical shape of the tip, the insertion depth will not hit the reference point for the capsule. This would introduce additional volume, hence distort the couplers impedance.

I will either have to DIY with some Primo capsules or look for other measurement mics with thinner nozzles. The Beyer MM1 would probably be the best I could get in the lower price range. They are more noisy than the Primos but have quite low tolerances and lower THD. They also come with a reputable measurement sheet. I already own a sample I once bought used from eBay for calibrating my speakers and am quite happy with it.

There is just one pitfall: A binaural setup with two mics mounted inside of a dummy head will drastically limit the length the microphone bodys. Thus, I would either have to vary the mounting angle to fit both mics right next to each other OR disassemble the units and manually shorten or carefully bend the nozzles ...

Beyer_MM1.jpg

Have a nice weekend,
P48


(1.) EDIT:
Could it be that most of the measurement mics below $100 are actually the same?

Presonus PRM1:
- Sensitivity: 14 mV/Pa
- Impedance: 200 Ohm
- DR: 106 dB
- SNR: 70 dB
- Max SPL: 132 dB

Superlux ECM-999:
- Sensitivity: 14 mV/Pa
- Impedance: 200 Ohm
- DR: 108 dB
- SNR: 70 dB
- Max SPL: 132 dB

Sonarworks XREF 20:
- Sensitivity: 14 mV/Pa
- DR: 106 dB
- SNR: 70 dB
- Max SPL: 132 dB

t.bone MM-1:
- Impedance: 200 Ohm
- SNR: 70 dB
- Max SPL: 132 dB

Same housings and pretty much the same specs. :oops:

(2.) EDIT:
According to GRAS' specifications the 60318-4 coupler shall be used with a 40AG pressure microphone (PZM). Here is its frequency response graph:

gras_40AG_response.jpg
https://www.gras.dk/products/measur...s-200-v/product/ss_export/pdf2?product_id=167

Looks like the referenced response in the spectrum below 20 kHz is most equal to the 90° configuration.
So I guess that a "consumer" microphone tuned for diffuse field measurements should be more or less sufficient. GRAS recommends a random incidence correction of -3 dB in the upper frequencies, though. See also p. 19 to p. 41 of this presentation: https://www.gras.dk/files/MiscFiles/News/GRAS Alma 2017 Micropones and couplers PWA 170102.pdf
 
Last edited:
D

Deleted member 16543

Guest
Hello everyone.
I'm interested in following this conversation as I have somewhat extensively investigated binaural microphones a few years ago.
My efforts were directed at making extra realistic binaural recordings, not headphone measurements. However, some general considerations can still be made and considered valid. Specifically, the fact that to measure something one has to make sure to not perturbate the entity they intend to measure. Which for sounds perceived by our eardrums translates into the necessity of using ear canal replicas.

By using headphones like PSB M4U or NAD HP50, which are 'voiced' with a procedure that tries to match the response at the eardrum with that of a balanced speaker in a room, or other headphones considered well balanced, it could be possible to come up with a base response curve that's representative enough (at least up to a certain frequency) of a balanced headphone set.
Other headphones could then be measured and some valuable information could be extracted, if not exactly quantitatively at least qualitatively, as to what type of sound they are likely to have in comparison to the reference headphones.
 
OP
Dreyfus

Dreyfus

Active Member
Joined
Feb 7, 2019
Messages
247
Likes
296
Location
Germany
Glad you are joining! :D

The plan is indeed to validate the response of the system by comparing it with a known standard like my AKG K371, which could be considered to be "balanced" (in case you are fine with Harman's approach). There will be some fundamental anomalies of course since my build as a whole won't meet the established standards. Different pinna, different (inner and outter) coupling, different microphone. But I'm fine with that as long as the measurements stay consistent and repeatable.

I still haven't decided for a mic, yet because there are so many things to consider ...
  • How well does the mic fit inside the coupler and dummy head?
  • If too large, could I mod it with manageable effort?
  • What about its SPL and THD capabilities?
  • Whats the SNR?
  • How flat is the response?
  • Does it refer to 0° or random incidence?
  • Does it need calibration and compensation?
  • Will it be compatibile with my interface?
  • What's the price?
... just to name a few.
Same for the dummy head ...
  • Does the size fit a large variety of headphones?
  • How well does it couple to different designs (pads)?
  • How to attach the ears? (silicone is really hard to glue)
  • How to access the inside (cut it in half, scalp the top or just drill a large hole)?
  • What about insulation and damping?
  • Or just skip the head and 3D print a coupler like the miniDSP E.A.R.S. for simplicity?
I have found solutions for most of these questions. But its hard to find the right balance, keeping cost and time effort versus practical utility in mind.


To be continued.
 
D

Deleted member 16543

Guest
You should be able to get meaningful results with a relatively simple setup.

The thing is, unless you use a realistic ear canal replica, you might get measurements that conceal or exaggerate certain differences in the measurements, depending on the frequency.

B&K and GRAS just recently upgraded their ear models to include ear canals. I don't know how much they cost or if you can even buy just the ears. The other thing would be a plastic tube. I used that method with success for recordings, but since you intend to do measurements I would say that a plastic tube is probably pushing the limits of what's scientifically correct.
If this is for your own interest, an exact replica of your ear canal would be the best thing, at least for headphone measurements, to tell you how different models will sound to you.
But then if you also want to make recordings, you are bound to produce something that works only for people with similar geometry of ear canals to your own.

As far a the mic goes, I would get one with high SPL handling (the canal response has peaks that easily exceed 15 dB, up to 20 with pinna+canal for binaural recordings). Don't worry too much about noise floor (signal will be orders of magnitude higher than that), and use a capsule about the size of the eardrum, placed where the eardrum is supposed to be. With bigger ones there might be issues with only part of the diaphragm being exposed to the soundwave, and that may or may not produce cancellation issues that could potentially conceal some frequencies.
I would use omni capsules. Flatness of frequency response is not that important, as you are only interested in differences between headphones measurements, and for binaural recordings one needs to go through a specific EQ calibration procedure anyway. Also, omnis FR is usually quite flat across the whole audio range, so I wouldn't worry too much. Primo makes some great electret capsules.
Then you need a capsule preamp. Naiant makes very reasonably priced preamps, if you are in the US or Canada.

For the head I had to look quite hard to find one with realistic dimensions (important for binaural recordings, but not for headphone measurements). I ended up using one of these https://www.roxydisplayinc.com/webpage/head-wig/head/male/eraw2.html
I cut the head and inserted my own 3D printed ears from the inside, so that they would be flush with the exterior surface (I modeled them to have that feature). I used foam to dampen internal resonances.
 
Top Bottom