On 08/16/2018 10:22 AM, Bo-Erik Sandholm wrote:
I agree with Len, we have not seen any technical spec of self noise level
of the MEMS (clusters?) that are used in Zylia.
Only text saying that in normal musical recording situations self noise is
not disturbing :-).
:-)
I have a personal theory that self noise of physical elements in an
ambisonic mic array is not directly additive.
The basis for my theory is that as we convert to B-format the noise from
all the physical elements are distributed over a spherical surface,
and the noise level for a virtual microphone in decoding do not have the
full sum of the added microphone noise levels.
Only coherent noise within the take up volume of the virtual microphone is
relevant in that directional microphones response.
I think the noise we are talking about is that of the difference
microphones. In an open array built with cardioids that would be for
order 2 or higher, in a rigid sphere array with omnidirectional capsules
that would be for order 1 or higher.
Those components drop in level at low frequencies at a rate of n x
6dB/octave (starting at high frequencies). For the second order
components of a microphone made of cardioids n = 1 (6dB/oct), add
6dB/oct per order increase for higher orders.
As the component drops in amplitude towards the low frequencies you need
a filter that compensates for the drop, and of course it amplifies the
sef-noise of the capsules as well. At some point you have to give up or
the noise becomes a problem (where exactly depends on the self-noise of
the capsules and what kind of materials you are recording).
In the second order microphones I'm building for the SpHEAR project I
can use the second order components down to about 400-500Hz ("unity
gain" for them is at around 9KHz). Even then the noise is objectionable
(but not necessarily "disturbing" :-) for recordings that have wide
dynamic range (my encoder uses an expander on those components to try to
minimize that effect). You can definitely hear the noise if there is a
silence in the recording. Of course it disappears if you mute the second
order components :-)
A third order microphone (made of cardioids - I don't think there is
one) would be worse, the drop would be 12dB/oct for the third order
components. So you would have to limit the low end of the frequency
response at a higher frequency.
AFAIK nobody specifies the noise specs for _those_ components in an
Ambisonics microphone. In a first order microphone made of cardioids
that is not a problem, as is the case for the first order components of
a second order microphone made of cardioids.
(as a reference, the open source octofile software released for the
OctoMic shows that their calibration has three choices for the low
frequency cut off for the second order components, 500Hz, 900Hz and
1.5KHz).
-- Fernando
_______________________________________________
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit
account or options, view archives and so on.
