• 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!

Ideas for Building a New Flat Plate Measurement Rig

maverickronin

Major Contributor
Forum Donor
Joined
Jul 19, 2018
Messages
2,527
Likes
3,308
Location
Midwest, USA
I'm looking to build a new headphone measurement rig for purposes of modding and EQing. My old/existing one is a Dayton EMM-6 mounted flush through a damped aluminum plate because that was the easiest to build at the time. Since it uses a full size measurement mic it's way too wide to actually make it "wear" a pair of headphones like a human head. Consequently, it was only easily usable with the Stax Lambdas I built it for since their cups easily detach from the headband. They can just be laid flat, and fastened in place on the plate.

I'd like to build a new one which is more convenient to use. I've been mulling over a few ideas and was hoping for some useful advice here.

Mod a MiniDSP EARS

(I'm skeptical of the idea of building an artificial pinna, then subtracting it from the measurement afterwards, plus the EARS' compensation isn't supposed to be very good anyway so I'd want to mod it into a flat plate.)

This might be the easiest and quickest depending on its construction. From pictures it looks like you can remove the artificial pinna with a few screws. Then it might just need a new flat plate flush with the mic capsules. The EARS also comes with calibration files for the raw capsules, which I haven't seen anyone complain about.

Unfortunately, my google-fu isn't good enough to find anyone who's tried this, or even a teardown which would show me what's beneath the white plastic pinna layer.

Mount a Standard Measurement Mic at an Angle

I found this rig on Homebrew Headphones but I'm not sure what kind of artifacts the recess and angle might introduce.

nVvhyKB.jpg


Mount a Shorter Measurement Mic Perpendicularly

The problem is I can't find one. All the inexpensive ones (Dayton EMM series, MiniDSP UMIK) are just closes of the same design which is too long. Browsing Sweetwater and B&H it seems the more expensive pro brands also use the same basic form factor.

Does such a thing even exist?

Mount a Raw Mic Capsule in a Flat Plate

This wouldn't be to difficult for me to build and I already have an interface with phantom power for my EMM-6. I don't really know much about picking out a capsule though. Can you even get raw capsules with calibration files?
 

Soniclife

Major Contributor
Forum Donor
Joined
Apr 13, 2017
Messages
4,500
Likes
5,417
Location
UK
I'd be interested in a cheap solution to this.
I hacked something together with blutak, a CD and my UMK1 the other week, it sort of worked but wasn't easy to get repeatable measurements.
Hopefully @solderdude will comment.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
I built my rig as described here.

This is based on a WM61A electret capsule. Alas no longer available.
There are Chinese 'copies' claiming they are genuine but they are not.
Of course there are plenty of other mic capsules that are cheap and can be used.

The downside is that they become non-linear above 100dB SPL or so and have rather high 2nd harmonics. Most electret do b.t.w.

The next step is calibration which is the tricky part I wrestled with for quite a few weeks.
My rig as it is now thus is based on the WM61A which may very well differ from other mics.
As it picks up hum I bought an iMM-6 from Dayton and made a calibration file based on my old calibration (done in a hardware mic amp).
The calibration file from the iMM was not correct for this rig. May well be fine as mic in free air.
Unfortunately, after a few months bass response started dropping suddenly and not compensationable any more.
Out the window with that mic. Not recommended.

How to calibrate is the big question of course. For anything below 1kHz this is easily solved. There is one headphone that is pretty constant in quality and has been measured by just about anyone that owns such a thing. Fortunately, probably due to its small diameter driver, it has hardly any pinna activation (checked this myself as well). When you haven't guessed yet... HD650
So ... slap it on the tes rig. Do a sweep. Compare it to other measurements (preferably without Harman boost) and make a calibration file so that it matches the plots that are made using a high quality rig. Below 1kHz Tyll's plots can also be used.
Then comes the task of calibrating the upper frequencies. Everything above 6kHz is kind of suspect.

The method I used started out from the Connover site which has data on the WM61A as a basis.
This seemed to give results that 'sounded' the same as what the plots showed.
I 'confirmed' this with 'flat' speakers in the backyard. Flat at 1m (well with its wiggles of course) using a monitor I borrowed.
Then played music on it (sitting on 1m distance to my spread apart with an angle) and putting on and off the HD650, also used white noise to check.
Playing with an equalizer in the headphone path I tried to match the 'tone' of mids to highs and it appeared as though a small correction was needed.
With that (small) correction applied measurements seem to correlate with most headphones (some do so less because of lack of a pinna)
Then I applied 'bass correction'.

Not based on preference of individuals but based on the fact that I listen at 70dB average and recordings are usually mixed at around 80dB average. So my 'choice' of compensation has a 'natural' form, that of the difference between 70 and 80Phon.
That works for me and when a headphone measures 'flat' it sounds very realistic.

Of course everyone is free to use their own 'standards' or theories.
The rig was made for my own measurements/curiosity. Not to set a standard.
For the few Euros it costs me the results are close to 'real' HATS.
Of course they differ from HATS with Pinna but that's another story that has to do with how the correction curve is obtained. It is an 'average' and smooth compensation where in reality the needed compensation (compared to 'audible flat') should differ between headphones.
It's why you see massive peaks and dips with HATS above 6kHz where, when one listens to sweeps or noise bands, these massive 'errors' do not seem to exist. For that reason I would not EQ on HATS results above 6kHz but rather look at pinna activation effects and do the 'peak' filtering by FP measurements. My initials happen to be F.P. as well so that is a sign from above :oops:

Anyway success with the build. Calibration is the hard part because of the references involved.

Wrote something here as well about calibration.

More reading material and discussion here.
One should realize that the plots there are differences of raw measurements.
These should have their own correction done before you judge them.
 
OP
maverickronin

maverickronin

Major Contributor
Forum Donor
Joined
Jul 19, 2018
Messages
2,527
Likes
3,308
Location
Midwest, USA
The Primo capsules might work well

Those look nice. I'll check around for local suppliers.

How to calibrate is the big question of course. For anything below 1kHz this is easily solved. There is one headphone that is pretty constant in quality and has been measured by just about anyone that owns such a thing. Fortunately, probably due to its small diameter driver, it has hardly any pinna activation (checked this myself as well). When you haven't guessed yet... HD650
So ... slap it on the tes rig. Do a sweep. Compare it to other measurements (preferably without Harman boost) and make a calibration file so that it matches the plots that are made using a high quality rig. Below 1kHz Tyll's plots can also be used.
Then comes the task of calibrating the upper frequencies. Everything above 6kHz is kind of suspect.

I've got an HD650 and that does make great "sanity" check for a system, though the last time I did this with just a Dayton mic pushed through an aluminum plate (which was nearly 8 years based from the dates on my old Arta impulse response files...) I found that I really didn't need much of a compensation so I was hoping to find some calibrated capsules and skip a step.

My google-fu isn't turning up much on that front though. I could conceivably calibrate against my Dayton EMM-6 but it's kinda old now so that might not turn out so well.
 

pwjazz

Addicted to Fun and Learning
Forum Donor
Joined
Sep 21, 2018
Messages
507
Likes
746
You could try something like crinacle's approach:

https://crinacle.com/graphs/headphones/

I personally use an in ear mic capsule mounted in some Symbio W hybrid tips and take measurements on my own head using my natural outer ear. I reference everything to my HD58X's measurements as a baseline. You can read more about that here:

https://forum.headphones.com/t/sound-professionals-sp-tfb-2-binaural-microphone/4759/7?u=pwjazz

EDIT - The tricky part with my in ear approach is achieving a consistent insertion depth. Well fitting, slightly large tips help with that.
 
OP
maverickronin

maverickronin

Major Contributor
Forum Donor
Joined
Jul 19, 2018
Messages
2,527
Likes
3,308
Location
Midwest, USA

pwjazz

Addicted to Fun and Learning
Forum Donor
Joined
Sep 21, 2018
Messages
507
Likes
746
How do CSDs turn out with that setup?

Good question, I haven't really paid much attention to them as I focus mostly on frequency response. When I get a chance I'll take a look.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
Would have to try it but a wand-microphone in a plate may not give the desired results.
Fun to play with perhaps.
My thoughts are in post #5
 
Top Bottom