A while back I had the opportunity to measure 24 WM-61s for a friend who wa=
s constructing an array. They pretty much met their specs. It's worth not=
ing that when he modified the capsules to be source followers ('Linkwitz mo=
d') that they were more closely matched. As common source (no modification=
) the mean sensitivity was -35.77 dBV/Pa with a Standard Deviation of 1.36 =
dB and as source followers they were -39.51 dBV/Pa with a standard deviatio=
n of 1.06 dBV. The mean self noise was 29.5 dB SPL (A weighted) with a sta=
ndard deviation of 1.8 dB (smaller data set; I wasn't able to find the time=
to measure the noise on all of the capsules.). Pretty much all of the cap=
sules were better than the data sheet limit of 32 dBA self noise.
Is matching the capsules necessary? It's known the the limit of the angula=
r resolution of human hearing is about 1 degree in the front, and that abou=
t 1 dB of amplitude difference between the two channels of a 60 degree ster=
eo pair produces about a 1.7 degree change in the image location. If you d=
esired to keep the image straight in front of you then you would want to ma=
tch the capsules to within better than 1 dB. Of course this is the sort of=
thing that can easily be corrected in post production. If you know that t=
he source that you recorded was directly in front of the recording setup th=
en you can simply adjust the relative levels of the channels so that the im=
age is in front of you when you sit exactly between your two speakers. Of =
course, that assumes that your speakers are that well matched...
If you do want to match your capsules you could do that pretty easily by pl=
aying a source of constant level into a single speaker and placing the caps=
ules very near to the speaker and then recording the output from the capsul=
es taken two at a time into your recorder or PC. Then compare the levels i=
n your audio editing program. You won't know what the absolute sensitivity=
is but you can pick to that are close to each other. If you do this match=
ing it's important to have the capsules be as close to each other as is pra=
ctically possible.
Some people prefer to use omnidirectional microphones like the WM-61 for st=
ereo recording but many people prefer cardioids or other directional microp=
hones. The reason for this is that spaced omni setups don't produce much l=
evel difference between the capsules at low frequencies. Omnis do have the=
advantage that they generally have very extended low frequency response . =
Cardioids don't have extended low frequency response but they can be used =
in coincident or non-coincident miking in such a way that they produce good=
low-frequency localtization cues. But essentially all cardioids roll off =
below about 150 Hz or so when used at a distance from the source. Nearly a=
ll published LF responses for cardioids are measured at a distance of 30 cm=
or so.
My opinion is that most of what is written in various discussion forums abo=
ut stereo miking is rubbish. But there are some very good papers and a boo=
k, "Microphone Arrays for Stereo and Multichannel Sound Recording" written =
by Michael Williams
Book: http://www.mmad.info/BSH/audiobookshop.htm
http://www.microphone-data.com/pdfs/Stereo%20zoom.pdf
Eric
|