[GreenKeys] Re: teletype machine noise levels

[Craig Sawyers]

> Thanks for the input.  PSI and inches-mercury, I am familiar with --
> pascals and bars not so much.  Any idea how an air-pressure level is
> related to a sound-pressure-level?  Seems like it should be a
> rate-of-change thing (eg: a fixed high air pressure is not making
> a sound).

Well, the Pascal (Pa) is the SI unit of pressure, equal to 1 Newton per
square metre.  Atmospheric pressure is 10^5 Pa.  1 Bar is very close to 1
atmosphere (multiply atmospheres by 1.01295 to get to bars), so for example
1 millibar is very close to 1/1000 of an atmosphere in whatever units you
are using (so it is equal to 14/1000 lb/in^2 or 31/1000 inches of mercury
etc).

Sound pressure level is defined as being:

SPL = 20 * log (p/pref)

Where pref is an accepted but arbitrary choice of 2E-5Pa (in other words
2E-10 of an atomosphere!).  The ear is a pressure-sensing device, not power
sensing.  Now not surprisingly, the ear does not have a level frequency
response.  As an example, the threshold of hearing at 3kHz is around and SPL
of -10dB, whereas at 20Hz it is +80dB.

Also, the shape of the ear's frequency response is level dependent, becoming
flatter as the SPL goes up.  To overcome the difference between percieved
sound and measured sound, the "phon" is defined as the SPL at 1kHz - and an
observer makes a decision about how loud a particular real sound is to a
reference level at a particular phon level.

So defining the curve by which an instrument measures *percieved* sound
level involves some compromises.  Particular curves have been defined that
apply well enough in certain circumstances.  The A weighting filter
approximates the response of the ear at the 40 phon level - in other words
rather quiet.  The B filter is hardly ever used, and is at some intermediate
level.  The C filter is for high level work, around 80 phons.

When measured in this way, the meter reading is written dBA or dBC to make
it clear what the measurement means.

Cheers

Craig