To further complicate this matter of whether 2dB, 3dB or 5dB should be adde=
d to
the A-weighting to get a better approximation of noise as we hear it...
I found four microphone manufacturers whose spec sheets include self-noise =
in
both dB A-weighted and ITU-R 468 (usually still called CCIR 468, which is a=
bit
old-fashioned). Everything I have read indicates that the ITU-R 468 curve i=
s
more revealing of the noise, since it does not completely remove the low
frequencies and has a higher peak where we are more sensitive (at 6.3 kHz).
The following compares these two figures, units omitted for online clarity:
Sennheiser MKH 8090: 13 / 23
Neumann KM 183: 13 / 24
Schoeps CMIT 5 U: 14 / 24
DPA 4006A: 15 / 27
A pattern is emerging. The ITU-R 468 figures are 10-12 dB greater than the=
A-weighted measurements. This indicates to me that we should be adding 10dB=
minimum to noise figures given using the A curve.
-- Robin Parmar
P.S. Merry Christmas! Enjoy the many sounds of the holidays!
|