4. Re: Summary: [Nature Recordists] Re: EIN values and bits and peaks

["Eric Benjamin" ericbenjamin2]

You're right, I did mean a sample rate of 48 kHz and therefore a bandwidth =
of 24
kHz.

The original definition of A weighting was from an Acoustical Society stand=
ard
from the 1930s in which they specified values in a table that only went up =
to 20
kHz.  But ANSI S1.42, Design Response of Weighting Networks for Acoustical=

Measurement, defines A-weighting differently, by poles and zeros, so that o=
ne
can calculate a value for the weighting at any frequency.  As you say, sinc=
e the
weighting is rolling off strongly at high frequencies it doesn't make a gre=
at
deal of difference.  If the bandwidth were 48 kHz, then the A-weighted leve=
l for
pink noise is reduced by 4.23 dB instead of 3.70 dB.