Build A Tetrahedral Ambisonic Microphone

[Dan Hemingson’s] been refining a design for building a tetrahedral ambisonic recording system. This is a set of four microphones used to record audio that can later be mixed down for a three-dimensional listening experience. His goal is an easy and inexpensive build while maintaining the highest fidelity standards possible. Lucky for us he’s made a set of extremely detailed build instructions you can use to make your own. In addition to the mounting bracket seen above he has also developed a pre-amp module that connects to the four mics; it’s part of the build instructions with schematic and board layout files available as well.

[Thanks Isaac]

27 thoughts on “Build A Tetrahedral Ambisonic Microphone

  1. Nifty! Although i don’t really understand it. Is he going to put speakers in a similar arrangement? And if so, how does is the recreation affected by the (new) room and position of the listener. I’d imagine this is something for a variable listening position though. .. but what and how and why and what!

    Still, good job and interesting concept! makes you think huh

  2. I could think of a few other uses for this.

    Namely capturing impulse responses to load up and simulate the acoustic characteristics of an environment. This is usually done in stereo, but with some work could probably fully replecate an area!

  3. @draeath like in convolution reverb you mean? that would be an interesting application indeed. Im just not that sure about “3d’ness” of sound and our ability to hear it.

    I think we mostly experience 3D sound (i dont like this phrase either) when we move, or the source moves. And if i understand correctly this bundle of mics would improve the positioning of that sound.

  4. @medix
    How do you measure wind speed with ultrasound?

    Somehow I’m not convinced that a $2.60 mic module is going to give you quality sound reproduction.

    There is a reason people spend hundreds to thousands on a single microphone.

  5. “Namely capturing impulse responses to load up and simulate the acoustic characteristics of an environment. This is usually done in stereo, but with some work could probably fully replecate an area!”

    Been there. Done that. Fons Adriaensen includes some ambisonic IRs with his JConv software (and I think some more on his website), if you want to know what it’s like.

    Another application is to convolve the ambi recording with HRTFs to get binaural sound minus the cumbersome head-sized microphone.
    This method does have it’s cons though, as it does not produce synaesthesia like actual dummy heads, and the inexact ITD causes some comb filtering.

    “Somehow I’m not convinced that a $2.60 mic module is going to give you quality sound reproduction.”

    You’re in for a surprise – all that is needed is a person with the right knowledge. What is usually compromised with these cheap mics is the signal to noise ratio, these capsules are noisy on the low side and distort rather soon on the loud side, so you have to mind your SPL.
    Even then, “hundreds of dollars” can get you a perfect microphone, the people who pay “thousands” are just paying for placebo, brand name, or just because they think using the same mic as [blindly praised old artist] will magically improve their recordings.

  6. Regarding the phase, I don’t know how big is this thing (too lazy to read so many PDFs – was dividing it so much really necessary?), but if you keep the total diameter of the capsule bundle under 2cm there shouldn’t be much problem.

  7. @Whatnot: Well, as I was saying, I’m too lazy to really check this out, but as far as I can see from the article pic, all the metal is behind the unidirectional (key word: unidirectional) capsules, so reflections off the metal should not have a noticeable effect. Plus the diameter looks really small, if anything it only affects very high frequencies.

    Handling noise, on the other hand…. (if you pardon the pun, of course).

  8. @M4CGYV3R: I’ve never really looked into it. I didn’t even know it was possible until a few years back when I was visiting a research lab at NREL where they were studying air flow dynamics along mining walls. They were using ultrasonic anemometers to get air flow and direction (in 3D). I’ve been vaguely interested ever since, just never really had the time to check it out in depth. Pretty cool stuff regardless.. ;)

  9. @Bosse:
    With 6 of them you don’t need software to decode as 3D sound, and it’s only about $5 more.

    No combination of microphones will truly mix properly to create a new single microphone profile.

    The mathematical combination of one waveform with another will not provide the same exact sound as a single waveform captured from the profile you are trying to simulate.

    The ‘virtual microphone’ idea is OK if you’re not too picky about how exact the profile is.

    @medix
    That sounds cool. I didn’t realize that moving air had enough sonic reflection or distortion to use ultrasound to visualize it.

    1. A cardioid or supercardioid pattern already comme from a mix between a omni and a figure 8 pattern.
      Ambisonic mic are capturing all information needed to reproduce all mic pattern, pointing into any direction (3D), can be format for al kind of stereo, surround, 5.1, quad, 7.1, or what ever you need. In fact, it is the way to go if you what a 3D images, since all other technics will introduce combfilter problem…exept for double MS witch is exactly the same thing minus the possibility of adding X axes (height).

      In thermes of spacialisation, there is nothing you cannot do with it.
      The only limitation will be in is spectrum, so if you don<t like the sound of the capsule (is color, there is nothing you can do about that.

  10. @M4CGYV3R — This is how you build an ultrasonic anemometer:

    http://thiesclima.com/usanemo_e.htm

    A wind velocity component in the direction of the propagation of the sound supports the speed of propagation, thus leading to an increase in the speed. A wind velocity component opposite to the direction of propagation, on the contrary, leads to a reduction of the speed of propagation.
    The speed of propagation resulting from the superposition leads to different propagation times of the sound at different wind velocities and directions over a fixed measurement path.

    I’d personally build one with an array of 6 ultrasonic units, though (N/S, E/W, U/D)

  11. vibrations @IR

    this sort of grounding approach also minimises noise..

    just a thought, but a “ghetto” way to make this is to harvest microphones from old mobile ‘phones and now-extinct analogue cordlesses.

Leave a Reply to M4CGYV3RCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.