• 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!

EM272 XLR Circuitry for Dummy Head Microphones

StephanF

Member
Joined
Aug 18, 2024
Messages
7
Likes
5
Hi All,

I am a new member and have found awesome discussions on dummy head microphones and HRTFs here on this forum. I have built a dummy head which I want to use for live video broadcasts of a local orchestra here in the Reno/Tahoe area. I had built one almost 50 years ago, when I was still in high school in Germany. I posted it to member Andysu:


I bought two matched EM272 microphones and tested the dummy head against two dynamic mics, and as I have read here on the forum, there is a emphasis in the 3-4 kHz range.

Dummy Head Test.jpg


Response Dummy Head vs. XML840.jpg

Above: Frequency response of the dummy head with EM272 relative to a pair of XML840s. I used a white noise recording played back through my 'calibrated' Denon and some awesome 'antique' Celestion SL6s with copper dome tweeters that I bought over the Internet from England. Later I got another pair with aluminum dome tweeters for surround sound. Back in Germany I had a pair of Celestion UL6s with a Monacor subwoofer and the sound was awesome. When I moved to the US in 1985, I left them with my high school buddy. The peak in the above diagram at 4 kHz is about 20 dB (!) above the XML840s. One has to consider that the level of the XLM840 (black) is about 8 dB higher than the dummy head (green).

So I looked at the surface mount circuit that is inside the XLR connectors of the EM272s and although I have good experience in electronics, I can't quite make sense of the circuit. One would assume that it would come with an FET circuit to better match the impedance for the XLR inputs of an external recording device. There is an IC with a stamp '5Ft' on it which I tried to find looking through various databases but with no luck. The only thing that would make sense is that it is a sort of voltage regulator, since the electret mics need to biased (correct?).

EM272PCB_Top.jpg


EM272PCB_Bottom.jpg


EM272Circuit.jpg

I measured across the diode and the reverse voltage starts at around 0.5 V and slowly, very slowly (over several minutes) goes up to 3.7 V or so, then at some point it is back to 0.5 V! But the sound is not affected by this. Very odd. (I also measured the forward voltage drop across the diode with my Fluke meter when the circuit was not powered, it was around 0.6V).

I am not sure if anybody can help here. My plan is to make a new surface mount PCB of the same size with an opamp that has a notch filter to eliminate the emphasis around the 3-4 kHz range. I would unsolder the original components from the Clippy PCB and reuse them in the new one. Alternatively, maybe as a first step, I would need to build a box with an external power supply that would take the EM272 signal and add the opamp with the notch filter, equipped with XLR jacks.

Regards, Stephan
 
Yes, that chip generates the bias voltage. The "breathing" when you measure the voltage is likely because of the finite impedance of your meter discharging C1. I don't know your particular capsule, but lots of them have a FET source follower built in.
 
Or some kind of common drain amplifier, like this one.

According to the datasheet, the raw capsule's FR deviation is within 2-3 dB. I would chalk the substantial deviation seen there up to HRTF.

So what does one typically do to process binaural recordings? The folks at MiniDSP seem to take the approach of EQing the peak out in post:
Binaural recording 101

Here is an appnote by HEAD Acoustics discussing the subject of EQ of binaural recordings that seems well worth going through:
 
Dummy heads are fun, many many years ago I had the JVC version, and it made quite good recordings. Of course there is a reason the Neumann goes for 9k though.

In my experience the dummy binaural only works well when listened back to on headphones, for classical orchestra I'd stick to Blumlein, or Decca Tree.

If you can record two tracks on the side for the dummy I'd do it. Lots of info out there on how-to with affordable microphones. Neat to see that people are still into the binaural stuff.
 

Attachments

  • JVC_HM200.JPG
    JVC_HM200.JPG
    46.2 KB · Views: 29
Or some kind of common drain amplifier, like this one.

According to the datasheet, the raw capsule's FR deviation is within 2-3 dB. I would chalk the substantial deviation seen there up to HRTF.

So what does one typically do to process binaural recordings? The folks at MiniDSP seem to take the approach of EQing the peak out in post:
Binaural recording 101

Here is an appnote by HEAD Acoustics discussing the subject of EQ of binaural recordings that seems well worth going through:
Thanks for the links!

So the capsule has it FET preamp, as Palmer pointed out. I am currently recording the voltage across the diode, there is a slow rise over 5 minutes to around 3.5V and the input impedance of my Fluke 287 will drain the supplied current a bit. I don't see the sudden voltage drops that I thought happened - I must have shortened out the voltage across the diode somehow.

I am considering buying an XLR preamp from Amazon and adding a filter to attenuate the 20 dB peak at around 4 kHz. I have simulated this in MultiSim and may publish my circuit here once I perfected it. My plan is to build a custom PCB with a potentiometer that allows me to adjust the 20 dB attenuation.

I studied physics at the RWTH-Aachen in Germany (graduated in 1984). There was a demonstration of a dummy head recording of a choir played back on a pair of loudspeakers. It was amazing, you could pick out each voice spatially! This was done at the Department of Electrical Engineering and sponsored by Mercedes Benz, they wanted a measuring device to objectively quantify road noise in their cars while been driven. The spin-off resulted in the startup of HEAD Acoustics! I had contacted them in the past. And yes, the choir recording was equalized for the reproduction on speakers.
 
Yes, that chip generates the bias voltage. The "breathing" when you measure the voltage is likely because of the finite impedance of your meter discharging C1. I don't know your particular capsule, but lots of them have a FET source follower built in.
So the capsule has it FET preamp, as you pointed out. I am currently recording the voltage across the diode, there is a slow rise over 5 minutes to around 3.5V and the input impedance (10 MOhm) of my Fluke 287 will drain the supplied current a bit. I don't see the sudden voltage drops that I thought happened - I must have shortened out the voltage across the diode somehow.
 
Here is the recording of the voltage measured across the diode:

1730828997894.png

The charging time constant comes to 118 seconds, or close to a 10 MOhm charging resistor with the 10 uF capacitor. Since my meter has a 10 MOhm input impedance, this of course is a bit off. So the chip in the XLR connector PCB is equivalent to a 3.5 V voltage source and a charging resistor of about 10 MOhm.
 
Dummy heads are fun, many many years ago I had the JVC version, and it made quite good recordings. Of course there is a reason the Neumann goes for 9k though.

In my experience the dummy binaural only works well when listened back to on headphones, for classical orchestra I'd stick to Blumlein, or Decca Tree.

If you can record two tracks on the side for the dummy I'd do it. Lots of info out there on how-to with affordable microphones. Neat to see that people are still into the binaural stuff.
My idea is to provide two audio channels for the live broadcast - but I don't know if that is even (technical) possible. One stereo channel for playback on speakers, but many people watch it on their cell phones with either headphones or earbuds, so these people will have a much more vivid audio experience listening to a dummy head stereo channel. I have recorded that orchestra in the past for future CD releases but the music director hasn't had much luck with selling CDs in the past 10 years or so. I have 8 good mics but had really bad luck with mixing due to comb filtering. The mixes sounded 'hollow' to me, so best to just use two mics close together, angled at 120 degrees. My mics are all cardioid condensers, I suspect that I get a lot of wall and floor reflections besides neighboring mics picking up the same sound source with different delays. I am 'just' a volunteer, I still have a lot to learn about sound engineering...

I always loved the TELARC recordings, they only used three omnis spaced far apart!! I wish I could somehow find out how they mixed it. Here is a picture from a vinyl cover, on the right one can see how they placed the mics:

TELARC Microphone Placement.jpg


In my case, I have to place at least one extra mic for soloist, voice or violin, so their sound then mixes with whatever the mics pick up from the orchestra, causing comb filtering.
 
I recorded the Messiah performance of the Tahoe Philharmonic Orchestra & Chorus with my dummy head. Using Audition, I applied a filter function:

1734375589104.png


This simple filter balanced the recording quite well.

IMG_5607.jpg



IMG_5614.jpg


Here are a couple of audio samples from the concert, the first one has voices coming from behind, one from the balcony and three voices from the far back:


And the ending:


Enjoy!
 
Back
Top Bottom