In a message dated 1/23/2004 12:00:59 PM Pacific Standard Time,
owner-ax-digest@autox.team.net writes:
> Rocky,
> I have done sound measurements for environmental imact studies, and feel the
> need to clarify what you have stated here to show why many misjudge noise
> measurement readings, and their effect on the surroundings.
> First the scale that is used for measurement, dbA is not linear. If you
> measure a steady sound source, say an idling engine, then measure the sound
> produced by TWO equally loud idling engines side by side, the reading would
> only increase by ~ 4 dbA. The dbA scale is also weighted to try to represent
> the acoustic sensitivity of the human ear at 1K Hz, about midrange of the
> spoken voice. However it has been proven that low frequencies are what make
> noise annoying (think about the thump thump thump of the overamplifled auto
> sound enthusiast as he passes your house late at nite), as well as the
> "raggedness" of the frequencies being generated, i.e.the difference between
> music and noise. The dbA scale does not accurately measure the low end of
> the frequency range as the volume increases to the point of being wildly
> inacurrate in scaling loud noise perception, so quoting the dbA readings
> does not reflect how much you are actually annoying the neighbors. Your wife
> dropping a glass on the floor in the kitchen is not a loud noise compared to
> the game you're watching on TV in the living room if measured on the dbA
> scale, but it probably raised you 3' out of the chair!
> Also to put your measurements into better perspective, compare them to
> commonly experienced sound levels: a typical building fire alarm = 65 dba,
> car passing @ 65 mph 25' away 75 dbA, hearing protection required in the
> workplace 85 dbA, a loud rock concert on the dance floor 100dbA, jet plane
> 100' away = 130dbA.
>
dbA weigting is very midrange biased. The ides is that the human ear has a
threashold of hearing that is drastically different based on frequency. The
weighting curves were developed to help measure how loud something would sound
to
the human ear. The basic "Radio Shack" level meter gives a choice of A or C
weighting and you will see a large different in level between the two when
engine rumble is dominant. Tire squeal will show pretty close on either scale.
The other switch on the meter is the dampening. slow/fast This is not
accurate at all as all it does is slow down the meter movement. But on the slow
reading, the meter will just plain miss fast peaks in sound level like from a
backfire. If the noise is steady, the readings will match, but it takes longer
to
get there so short bursts get missed. On the flip side, the fast moving needle
can overshoot and show higher than the sound actually reached, soi that is
not always fair either.
Some professional meters also add a flat non weighted scale and possibly B
weighting as well as other response speeds. Some PC based systems can read
average and peak levels together so you can get a better idea of what the sound
is
doing over time.
OSHA sound level tests for hearing damage and protection requirements are
done at dbA weighting for a few reasons. Over many years of research it has
been
shown that hearing damage is also most likely in the most sensitive mid range
area where A weighting reads already, and many machine shops have ALOT of very
low frequency noise which has virtually no damaging effect on ears at all
since it is well below where the ear drum can even respond. Pounding bass may
be
anoying, but it really does not damage hearing until well over 120 db below 60
hz. But 90 db at 1200 hz can cause damage. Too bad many "BOOM" cars actually
boom at over 100 hz and can cause serious damage to ears, and add in harmonic
distortion and the threat to ears gets even worse, but that is a story for
another paper. In any case, using OSHA scales may protect our ears, but
probably
won't stop complaints.
I have always thought using A weighting for checking car noise was wrong. A
Viper or CP car could show quite low, but a Honda or Mazda rotary might peg the
meter even though they don't sound much louder, just higher pitched. And of
course the poor stock car with street tires just had some squeal and got kicked
out of the event.
The weighting curves were not really created to filter the noise, but to make
sound level measuring reflect what the human ear percieves, and for this
purpose we are using the weighting wrong. Take a look at this web page.
http://www.dataphysics.com/support/library/downloads/articles/DP-Aweight.pdf
The A curve is meant for low level sounds, and C curve is for high level
sounds. There is also a B curve between them. Since sound systems ( I set up
movie
theatres ) are set up at 85 db, we use the C weighting. This is fairly
accurate, but can still be off if the response is not totally flat. It is
interesting that this web page compares the percived loudness of snow blowers.
This
probably fits our use much more closely, but of course the equipment to do it
right is much more expensive. Since we are interested in 90 db levels, C
weighting
is much more appropriate, but.....If we do switch to using C weighted, the DB
limits should be scaled up a bit as the wider range measurement will read
higher. Measuring flat steady pink noise, I get 85 dbC or just 81 dbA.
I hope this helps shed some light on a tricky situation. I have been working
on movie theater sound systems for 20 years and I now work for Dolby Labs. I
know sound level pretty well, but there is still the big question about what
makes people complain. It does not have to be loud in the complainer just does
not want you there to begin with.
Gary Meissner
|