soundfield mic project feasability.

Over the weekend I made up a sounfield array using Panasonic capsules.
I did it not with the idea of producing something to rival the real thing but to look at the mecanical side of it. At first I thought that all the angles were 60deg so I machined up some mounting plates and they did not fit together well.
An hour of digging in my memory of maths from school and 2mins with my sons scientific calculator and I had another design which fits perfectly together. The parts can be easily machined from say Plexiglas ( Perspex) if you have the machinary, which I do.
I do not have any info on how big the capsules are in the commercial product but I guess they are not large (15-20mm).
QUESTIONS / IDEAS

Does anyone know of a source for capsules of this size, my design would adapt to pretty well anything that was round.
If I do the mechanical side can I get someone interested in producing a PCB design with Gerber for Dave Malhams encoder/decoder. To find this go to

"homepage.ntlworld.com/henry01/ cheap_soundfield/cheap_soundfield.htm - 16k"


or Google "soundfield" and you will find it.

Steve

What directional charakteristic are your panasonic capsules? Most of the electrect types are omni.

I think the four capsules in Soundfield are cardioids - arent they?

Jakob E.

Thats right, I used uni's. It just would not work with omni's. I really want to do this project and where most would shy away from the making of the array, to me the hard part is finding suitable capsules and of course the electronics. But the matrix to produce the "B" format is dead simple and there are now plug-in's for Pro Tools and Nuendo so you dont have to build a decoder. However I would like to make an entirely portable one for mini-disc use, but I am no good with PCB design.
Steve

> I think the four capsules in Soundfield are cardioids - arent they?

I have an old paper showing all four back-chambers connected together: they are not four mikes but one 4-way mike.

If we are studying this, it may be a wee bit easier to drop the vertical pattern control and just use three elements. That would give the most useful features with (I think) considerable reduction in matrix complexity and initial confusion.

interesting idea. I can machine parts at my job if we need metal/plastic items.

now for my ignorance.. I just don't know exactly why you would use a soundfield mic, can someone clarify this for me?

PRR,

From what I have managed to learn so far the idea is to "null out" the rear part of the capsules pick up pattern to permit more accurate pinpointing. This is now done electronically but in earlier efforts may have tried to lose the rear pick up by cancellation/absorption . The RCA ribbon cardioid turned a figure of eight into a cardioid in this way. Most of the arrays from the Calrec onwards have used a tetrahedryl arrangement so I thought I would give that a try first. However if three does it OK.
Steve

I made a B-Format soundfield recording last night with my brass band. -Ther rehearsal room has a slightly buzzy fluorescent overhead, so vertical Z-axis tomfoolery can be played with, as well as left-right rotation.

Interestingly enough, there was a draught from an Air conditioning duct right in the area where I wanted to have the mic, so I was forced to fashion a crude "shaggy" windshield... worked well enough to make the recording.

I'm going to try and edit down a 20-30 second B-format set of wav files, for the W, X, Y and Z signals. they will probablyt be about 5 meg each.

If anyone wants to download VVmic (visual virtual microphone) which I know works for PC, don't know if there's a mac equivalent) then they can play around with generating 5.1, wide stereo, M/S outputs etc.

To get the angles correct, think of a pyramid. The pyramids in Egypt have four equilateral sides and one square side (the base) -Lose one of the four sides and make the base an equilateral triangle also. Now you have the four angles correct. You can do it by drawing four equilateral triangles on a sheet of card or paper, then cutting them and gluing them into the geometric shape. then, if you mark a circle on each face, then trim off the four major corners, you have a decent approximation of the soundfield capsule array.

PRR, the chambers don't appear to be joined in any way. I can take close-up shots if anyone is interested...

Keith

Hi Svart,

If you google "Soundfield" you will find the makers website. Rumaging there will tell all. It's a very vesatile mic. Also there is a picture of the array.

Steve

[quote author="PRR"]If we are studying this, it may be a wee bit easier to drop the vertical pattern control and just use three elements. That would give the most useful features with (I think) considerable reduction in matrix complexity and initial confusion.[/quote]
Since there are software decoders available for free if you can generate a true B-format encoder, the added complexity of Z-axis decoding isn't so much of an issue perhaps...

I'm not that clear on the mathmatics as to how it would change if you only made a three-element (meaning three-capsule) array... someone else would have to crunch the numbers, my brain is too numbed by the idea!

With the current 4-capsule array:
W (omni) is all four added together, then the total attenuated by 3dB.
X axis is Left Front - Right Back + Right Front - Left Back
Y axis is Left Front - Right Back - Right Front + Left Back
Z axis is Left Front + Right Back - Right Front - Left Back

By the time you've built a matrix for three outputs (three summing amps) it seems like it's anly a small step to build a fourth?

Keith

For those who are on PC's, here's the link to the VVmic application:

http://mcgriffy.com/audio/ambisonic/vvmic

Even if you're not planning to ever dabble with soundfield mics, this is a great way to see what happens when you play with cardiods and omnis etc... mic pattern interaction etc.

Download it, install it, it's tiny and runs easily.

Keith