I think to measure EIN you need a machine that will provide a signal of
known voltage, and measure the output accurately. Set the machine to provid=
e
the recorder a certain mic voltage, say 1 mV, and adjust the recorder to
give a certain output, say -20 dB. Then short the mic, and measure what you=
get out of the recorder. It's not as direct measurement, though, as you mus=
t
measure what is recorded, not what comes out of the headphone jack (maybe
that would work for a close measurement, though).
The 671 calculation by Raimond is done this way:
Start with the signal-to-noise ratio or 65 dB. That means the noise is 65 d=
B
below the signal, whatever that is. In the specs, it reports the "MIC level=
"
is 1.2 mV. To find this in dBu, signal =3D 20 log(1.2/775). 775 mV is the r=
ms
reference value for the dBu. Signal =3D -56 dBu, so the noise is 65 dB belo=
w
that, -121 dBu.
With the exception of SD and one other company I can't recall at the moment=
,
all EIN's are calculated with these sort of manipulations. I think they are=
all slightly suspect, as the two specs could be measured under different
conditions, making them not exactly "combinable." Max signal values are
calculated the same sort of way.
Bruce Wilson KF7K
http://science.uvsc.edu/wilson
|