Klas Strandberg wrote:
> Two omnis near one-another and connected out of
> phase will only cancel one-another,
Klas,
At some risk of being chided by the rest of the
membership for getting too far off track, the reason
two omni microphones placed back-to-back give a dipole
(a figure-of-eight) has to do with the fact that there
is a slight spacing between them. That is, they are
never _exactly_ at the same place. The MX183
microphones are a good example. Two of them back to
back have a nominal spacing of 44 mm. One can make
the approximation that they are coincident at very low
frequencies. So, assuming that they are of the same
sensitivity, or the sensitivies have been adjusted to
be the same, then sound arriving at the two
microphones will be (nearly) cancelled when the two
microphone signals are subtracted. But in fact, there
is a phase difference between the two microphone
signals, and the phase difference becomes larger as
frequency increases.
If you imagine a sound arriving from the side of the
two microphones, the microphone signals will be
identical, and the subtraction process will give zero
output. For a sound arriving on the axis of the two
microphones, there will be a delay for our MX183 array
of .044/343 =3D 0.128 milliseconds between the first and
second arrival. This amounts to 46.2 degrees at 1
kHz. So the output from the array is 10.24 dB less
than the signal from either capsule alone.
So one can make a figure-of-eight microphone this way,
and it will have quite good polar patterns, but after
being equalized to have flat response it probably
won't have a very good signal to noise ratio.
There are some problems with the method of subtracting
two cardioids, too. One problem is that real
cardioids aren't comprised of a perfect
figure-of-eight co-located with a perfect omni. And
in this case, where you want the subtraction to be
perfect, it's not, because of the same problems having
to do with the spacing between the capsules.
Eric Benjamin
|