naturerecordists
[Top] [All Lists]

1. Re: Surround Sound / Wildlife Recordings

Subject: 1. Re: Surround Sound / Wildlife Recordings
From: "Eric Benjamin" ericbenjamin2
Date: Fri Aug 27, 2010 8:40 am ((PDT))
>> [Four 4060s aligned in a tetrahedral pattern will provide the basis for=

>>ambisonic surround  ~Bernie]

and Marcus Buick wrote:
> my understanding of ambisonics <>
> the four capsules in the tetrahedral array are not only aligned symmetric=
ally,

> but electronically as well

It is possible to make an Ambisonic recording using omni microphones.  At s=
ome
risk of exceeding the technology threshold, I'll elaborate.

The original Soundfield microphone uses four nominally cardioid capsules
assembled in a tetrahedral pattern.  The omni output is created by summing =
all
four capsules together, and the three figure-of-eight outputs are realized =
by
subtracting the capsule signals in pairs.  For instance, subtracting the tw=
o
back capsules from the two front capsules gives a forward facing figure of=

eight.  The electronics does the subtraction, and it also provides compensa=
tion
(equalization) for changes in response due to the size of the array.  The r=
esult
is something that performs pretty well over the entire audio range, althoug=
h
both ends of the spectrum are problematic.

It's easy to see that if you have a group of omni microphones that you coul=
d get
a good omni response by simply summing them.  What is not so obvious is tha=
t you
can get a figure-of-eight pattern by subtracting them.  But they need to be=

close enough together that there aren't multiple wavelengths of sound separ=
ating
pairs of capsules.  There is another problem. The Fig8s that are derived in=
 this
way roll off at 6 dB/octave below the frequency at which the capsules are
separated by a wavelength.  So you end up with a frequency response that lo=
oks
like a haystack; there's no flat region in it.  This can be fixed by
equalization.  But it takes a lot of EQ.  If, for instance, the capsules we=
re
separated by a single wavelength at 10 kHz, and it was desired to have a
response that goes down to 100 Hz, then the system would need to have 40 dB=
 of
bass boost.  All of that bass boost increases the noise, although mostly in=
 the
bass.

That sounds pretty grim.  But omni microphones are almost always better in =
SNR
than figure eight or cardioid microphones.  Also, they have essentially
unlimited bass response as compared to figure-eight or cardioid microphones=
 made
in the ordinary way.  Those all roll off below about 150 Hz to 200 Hz, desp=
ite
what the manufacturer's spec sheets may say.

So it ends up being kind of a wash.  If you use omni microphones to achieve=
 a
figure of eight response then you end up with a Fig8 that rolls off below, =
for
example, 100 Hz and above 10 kHz.  But regular figure of eights do that too=
.
And the omni SNR is made worse by the processing, but then regular figure o=
f
eight and cardioid microphones have worse SNR too.

It turns out to be a good idea to put the microphone capsules on the surfac=
e of
a rigid sphere.  Imagine something about the size of a baseball, although y=
ou
would probably want to use something easier to work with than a baseball.  =
There
is a pretty large technical literature on this subject.   I'd say that ther=
e
have been several dozen technical papers describing various aspects of it i=
n the
last five years.  Most of them are full of things like Bessel functions and=

Hankel functions that really aren't friendly to the typical nature recordis=
t.
But you don't need to know that stuff to actually build and use a microphon=
e
array.

> I own eight 4060=E2=80=99s, and would gladly reconfigure four of them for=
 ambisonic
use

If this is something that you really would like to pursue then I would glad=
ly
help you.  But we would probably want to take the discussion off list.  You=

would need to be comfortable with the process of drilling holes in somethin=
g to
make the array, and we would need to find a process that allows you to do t=
he EQ
in a way that is easy for you.

Eric



<Prev in Thread] Current Thread [Next in Thread>
Admin

The University of NSW School of Computer and Engineering takes no responsibility for the contents of this archive. It is purely a compilation of material sent by many people to the naturerecordists mailing list. It has not been checked for accuracy nor its content verified in any way. If you wish to get material removed from the archive or have other queries about the archive e-mail Andrew Taylor at this address: andrewt@cse.unsw.EDU.AU