Looking for in-ear measurement microphones for customized HRTF

[Fowly]

Hi everyone,

I'm looking for in-ear measurement microphones, similar to what Sony is using in this video :

Just like on the video, I'm planning on recording true stereo IRs for speaker simulation on headphones, as well as doing customized headphones calibration. I have trouble finding microphones that go deep enough in the ear canal, I want the measurements to be as correct as possible.

Any ideas ?

 

[NTK]

David Griesinger showed in this video how he constructed his.

 

[Fowly]

Thanks, so this is similar to the probes used in "real ear measurements" in the medical field right ? I'm not familiar with those devices, but is it suitable for my use case, which is both measuring and impulse response recording ?

 

[NTK]

Here is Dr Griesinger's presentation of his effort. I am pretty sure you can use them to measure your HRTF. Otherwise, I can't see of a reason for Dr Griesinger to go through the trouble. Sorry, I don't understand what you mean by impulse response recording.

http://www.davidgriesinger.com/Binaural_hearing_and_headphones.ppt

 

[NTK]

Genelec also have something they called Aural ID. But their pricing is a bit hard to stomach.

Aural ID - Genelec.com

The ultimate in tailor-made headphone monitoring. Create your profile, download the plug-in, and enjoy a headphone listening experience you can finally trust.

www.genelec.com

 

[Fowly]

Here is Dr Griesinger's presentation of his effort. I am pretty sure you can use them to measure your HRTF. Otherwise, I can't see of a reason for Dr Griesinger to go through the trouble. Sorry, I don't understand what you mean by impulse response recording.

http://www.davidgriesinger.com/Binaural_hearing_and_headphones.ppt

Very interesting, thanks for the share. I'm wondering how much inaccuracies does the external placement of the capsule creates. It's probably in the lower frequencies, so maybe this system is accurate after all.

About the impulse response recording, I want to use those microphones to capture the impulse responses of speakers in a mixing/mastering room, but at the eardrums. So in theory, applying those IRs using a convolver should perfectly recreate the sound of speakers on a pair of headphones (provided that they have been personally calibrated). It will allow for mixing/mastering on the go.

Genelec also have something they called Aural ID. But their pricing is a bit hard to stomach.

Aural ID - Genelec.com

The ultimate in tailor-made headphone monitoring. Create your profile, download the plug-in, and enjoy a headphone listening experience you can finally trust.

www.genelec.com

Yeah I saw that. With all the respect that I have for Genelec, I don't believe this system is any good. It relies on the typical "average frequency response" of headphones to calibrate them.

 

[SDC]

I use Microphone Madness, Sound professional, and Samson 3mm lavalier microphone. With Final audio type E small eartip.

MM has the best left right matching but brittle. And the cable is bit thick for good fixture.

SP subminiature mic is cheap and usable but bit bigger than 3mm.

Samson 3mm has left right matching problem(at least mine has). But capsule is small and cable is flexible so good if you can fix the left right problem some how...

 

[Fowly]

View attachment 307351

I use Microphone Madness, Sound professional, and Samson 3mm lavalier microphone. With Final audio type E small eartip.

MM has the best left right matching but brittle. And the cable is bit thick for good fixture.

SP subminiature mic is cheap and usable but bit bigger than 3mm.

Samson 3mm has left right matching problem(at least mine has). But capsule is small and cable is flexible so good if you can fix the left right problem some how...

How are you calibrating the frequency response of the microphones themselves ? Do they come with their own measurements ?

 

[SDC]

How are you calibrating the frequency response of the microphones themselves ? Do they come with their own measurements ?

I use implucifier.
It produces BRIR file by comparing headphone response and room-speaker response so there is no need for calibration.

 

[SDC]

I use implucifier.
It produces BRIR file by comparing headphone response and room-speaker response so there is no need for calibration.

And about calibration.
Members in Korea compared it with Earthwork microphone and it is like below

Edited.

 

[Fowly]

I use implucifier.
It produces BRIR file by comparing headphone response and room-speaker response so there is no need for calibration.

I don't think I understand the setup completely, I never used Impulcifer. Is it comparing the room-response measured by a omnidirectional mic (like the Earthworks) to the measurements made by the binaural mics ? That would make sense to me, but you said that it compares the room-speaker response to the "headphone response".

 

[SDC]

I don't think I understand the setup completely, I never used Impulcifer. Is it comparing the room-response measured by a omnidirectional mic (like the Earthworks) to the measurements made by the binaural mics ? That would make sense to me, but you said that it compares the room-speaker response to the "headphone response".

Brir file is created by following steps

1. Wear the ear mic

2. Measure the room with ear mic

3. Measure the headphone with ear mic

4. Run the program.


Result from 3, headphone responce is created
Result from 2, room responce is created

Program than EQ the headphone to target from 3.

Reverb filter is created from 2 and applied to target.

By this room-speaker-hrtf interaction is fully recorded and applied to headphone.

All steps are done with ear mic and no omnidirectional room measurements.

 

[Fowly]

Brir file is created by following steps

1. Wear the ear mic

2. Measure the room with ear mic

3. Measure the headphone with ear mic

4. Run the program.


Result from 3, headphone responce is created
Result from 2, room responce is created

Program than EQ the headphone to target from 3.

Reverb filter is created from 2 and applied to target.

By this room-speaker-hrtf interaction is fully recorded and applied to headphone.

All steps are done with ear mic and no omnidirectional room measurements.

Alright, it now makes sense to me. Thanks ! How does it sound ? I'm not convinced that an ear canal blocking microphone can produce convincing results for this task, as they give a completely different HRTF compared to the sound received at the eardrum (ear canal blocking is ECEbl in the figure below) :

But maybe I'm missing something ?

 

[SDC]

Alright, it now makes sense to me. Thanks ! How does it sound ? I'm not convinced that an ear canal blocking microphone can produce convincing results for this task, as they give a completely different HRTF compared to the sound received at the eardrum (ECEbl in the figure below) :

But maybe I'm that the ear canal gain is the same for a sound coming from the speakers or headphones, so it doesn't matter ?

About the tip blocking the ear canal... Yes it does change the tone.

I made brir files for about 25 members and 100 headphones in Korea and results were mixed...

Lack of mid low and excessive high was major problem...

But bacch also use tips that blocks the ear so I believe the problem is small and is fixable through little shelving eq.


For perfection Sony earmic is the answer but I have no idea how to make it and where to buy it...

Edit:

With 3mm mic you can make a very small ear mic that can reach the eardrum.
But it is dangerous so thats the reason I stopped with final typeE small tip. As it goes in deep enough in my experience...

 

[bennybbbx]

is there any progress with that ? . I also like see how the FR look with my own auricle . I not trust omnidirectional measure microphones for room measurements. this seem logic because at higher frequencies the auricle is more directive and so filter out room reflections so room modes are diffrent in ear canal. auricle also also is ~90 degree to speaker. I replace my display with a curved one so maybe it work as a diffusor because i think it sound better but no measure changes notice . so i measure with the mini dsp ear headphone measurements system https://www.soundimports.eu/de/mini...p_Msu8JuPwycMoFlknw5L7vqRuuKZVncaAgEqEALw_wcB and the HPN profile and a omnidirectional microphone.

i compare always mono music in correctet superlux HD 387 headphone which have a FR here https://www.audiosciencereview.com/forum/index.php?threads/superlux-hd387-review-budget-iem.33245/ with speaker (what diffrence i hear) and i notice that the measure with the auricle with the sound much more simular . I align the measures to 80 hz so both can more easy compare. you can see upto 250 hz all same. also the often show extrem 800 hz boost the hs5 should have happen not with the auricle. the 500 hz gap is not auricle measure which is correct. the 8 khz boost on auricle i hear too and the down go over 10 khz. but i hear that it not go down so much. maybe need a 3D scanner and build with 3D printer a auricle for omnidirectional measure microphone to plug on it

I think the correction curve for the auricle that is attach for the ear is correct, because in range of 3-6 khz the correction curve do lots changes but curve look smooth

 

[Mr. Haelscheir]

I've been using the in-ear mics supplied by https://www.earfish.eu/, though one of the mics had great bass roll-off compared to the other and my Genelec mic, and I find their effectively 7 mm diameter including the foam extremely uncomfortable. Otherwise, I've shared my results with such in https://www.head-fi.org/threads/anyone-into-crossfeed.961533/post-17950791 (post #32).

An advantage of blocked canal measurements is insulation of your eardrums from the test tones and hence the ability to achieve higher SNR measurements, whereby I have personally been able to get some quite nice measurements of headphone distortion. I currently think blocked canal measurements can capture most of the natural treble nulls, or at least certainly enough for convincing head-tracked binaural rendering in every direction. I would vie that so long as you use the same microphones and measurement method between the HRTF capture and the process of EQing the headphones toward that response (my in-ear mics had decent consistency between seatings), you can at least ensure that the FR that your calibrated headphones will be imparting on the canal entrance will effectively be the same as with the speakers, whereby any errors from differences in ear canal HRTF from 0-degree or 30-degree speakers measurements etc. versus 90-degree headphone incidence should not be too great or can be corrected with EQ.

Otherwise, for the purpose of distortion measurements, does anyone know of 3 mm or 4 mm diameter capsules with low noise and distortion, ideally not electret if possible or at least having direct support for 48 V phantom power? In https://www.audiosciencereview.com/...hones-with-motu-m2-and-rew.49384/post-1783062 (post #6), I had found tradeoffs between the VXLR Pro and VXLR+ Phantom power to plug-in-power adapters, the former having better noise but problematic bass third-order harmonic distortion probably due to an internal transformer while the latter exhibits increased midrange noise.

 

[bennybbbx]

yes 7 mm is too large and is not usefull because it is much outside ear canal and can cause damage in ear skin. the ear skin is very thin and there is very few distance to ear bone. a damage cause long healing and much pain.

The smallest i see are MEMS microphones. it have 3x4 mm in size. https://www.reichelt.de/mems-mikrof...iPMbwN1FcFYun6vQr6EnuSJx7LSK0BiMaAoAaEALw_wcB the ikmultimedia measure microphone use such as this. if the microphones you have are not good in bass doesnt matter. you can do a compare measure. In bass does only change minimal if your microphone position is 2 cm diffrent when you do compare measure with a better microphone

for mems microphones is tell it have high noise level as disabvantage in general but the ik multimedia arc microphone i have is more than 6 db lower in noise as my sonarworks measure microphone i have. the mems microphone is not very linear and need more correction curve. thats correction curve i use. with that cureve it is same as sonarworks microphone

I attach the calibration curve i made maybe somebody want use the ik multimedia mcrophone in other measure software as arc or can enhance the curve with compare to a very expensive microphone