[SpHEAR-devel] SpHEAR - MGW ambisonic microphone help, please!

[Fernando Lopez-Lezcano]

On 06/28/2018 02:39 PM, mwildo at tiscali.co.uk wrote:
...
> I have recently started to develop my own version of a 4 microphone ambisonic array. My 3D-printable tetrahedral design
> is now well advanced.

Cool! Send us pictures :-)

> I understand that cardioid capsules are rear-vented. Is it important that the containing ‘sphere’ of the capsule mount
> is as ‘open’ as possible so as not to upset the designed performance of the capsules, or should the mount be ‘as
> ‘closed’ as possible?

A tetrahedral design relies on the capsules being cardioid, the polar 
patterns of the capsules overlap in such a way that the first order 
Ambisonics components are extracted by simple additions and 
substractions (minus the high frequency effects, of course).

So it is crucial that the back of the capsules be as open as possible, 
otherwise they would not behave as cardioids and the shape of the 
recovered spherical harmonics would be distorted.

A secondary effect which I'm looking at in more detail is that the 
capsules enclose a "cavity" that acts as a resonator and also affects 
the response of the microphone. That is more evident in the 8 capsule 
design as the resonant frequency is lower.

> At present, the area of calibration seems to me something of a ‘dark’ art. Would it be possible for someone to write a
> ‘read me’ which explains the calibration process simply from ‘the ground up’? Maybe there is a definitive document out
> there somewhere.

The git repository for the SpHEAR project has some information:
https://cm-gitlab.stanford.edu/ambisonics/SpHEAR/
(the initial page is, of course, outdated)

The current calibration procedure (in terms of software) is somewhat 
described here:
https://cm-gitlab.stanford.edu/ambisonics/SpHEAR/blob/master/calibration/src/sphear_design_example.m
(but needs some commits right now)

See this also (old article on the project):
https://ccrma.stanford.edu/~nando/publications/sphear.pdf
(but the calibration procedure described there has been changed)

This all relies on having good impulse response measurements in the 
first place...
-- Fernando