David, you wrote,
> Suspect any noise figure that
> doesn't quote "468" weighting which is the weighting that the ear
> hears at low levels and includes a boost of about +12 dB. "A" weighing
> is for high level sounds.
That last sentence makes me jump on you! But your point about the ITU-R 468=
standard weighting curve is well taken. I imagine the reason that it is so=
little used in specifications, and unavailable in common measurement plug-=
ins, is that it gives higher noise numbers than A-weighting. Marketing valu=
es corrupt technical quality. You're lucky to live in Great Britain where t=
he BBC's use of IRU-R 468 is keeping it alive.
I note that Sennheiser quotes self-noise with both curves. The MKH 8020 (om=
ni) is specified at 10 dBA and 21 dB (unnamed, I presume 468).
The ITR-R 468 method not only has a 12 dB boost in the frequencies where th=
e ear is most sensitive at low levels, it has a complicated "quasi-peak" av=
eraging system devised to better match the way we perceive the time-structu=
re of low-level sounds--the Fletcher-Munson curve was based on the percepti=
on of tones, not noise.
There's a good Wikipedia article: http://en.wikipedia.org/wiki/ITU-R_468_no=
ise_weighting
Now about A-weighting, it's not for high level sounds!
A-weighting was devised for weighting the measurement of low-level sounds t=
o be more like their audibility, mainly by severely rolling off the low fre=
quencies to match the Fletcher-Munson 40-phon curve. It does that job prett=
y well.
The reason A-weighting is now commonly used for measuring -all- levels of e=
nvironmental sounds is simply practical--it's hard enough to get a non-tech=
nical person (think, policeman responding to a noise complaint) to produce =
one reading; switching between A & C and recording both numbers has proved =
to be impractical. At high levels A-weighted readings are seriously mislead=
ing, but they are commonly used nonetheless, even in scientific work.
-Dan
|