"Jerry G." wrote in message ...
"Hank2" wrote in message
. com...
I have recently bought a Radio Shack Digital Sound Level Meter. It has a
facility for taking "average" readings over periods up to 200 seconds. I
tried an experiment where I performed an "average" over a set period during
which, for half of the time I turned the noise source off. Instead of
indicating and average equal to 3db below the source power, a drop of around
8db showed up. Obviously it was not averaging power.

Other tests where I varied the sound source power during the averaging
period, showed that the meter was not indicating average db readings either.

If the "average" indicated is not of power, nor of dbs, just what is it?

An explanation would be much appreciated.
The calibration of the meter is referenced in a calibrated standard test
chamber using calibrated standard referencing equipment. Using it in a
normal room or stage setting will be very different. I will give you an
equivalent, according to the effects of the acoustics of the environment.
The sound source that you are using is bouncing around off of objects in the
area, and is not necessarily accurate in itself to begin with.

If you read the instruction booklet that came with the meter, you will see
that it is fairly complex in many aspects, yet it can be simple to use as a
fixed reference for a particular circumstance. As for its response, it
should have the options to follow a specific weighted curves.

Mr. Greenburg, do you EVER bother to actually read the posts you are
responding to? The poster had a very specific question about a specific
function that is NOT behaving the way expected. Your answer had pretty
much nothing to do with his question.

As to the question itself, it's not entirely clear, as I have not
seen the specific unit, what they mean by a 200 second averaging
period. If it means the time constant of the averager is 200 seconds,
then it is entirely possible for the behavior you have seen, since
the averager is not a fixed time window precisely 200 seconds that
excludes information outside that range. Rather, the averager is
"leaky" with a decay time of 200 seconds, as one example.

When you first turned your noise source on, the value in the averager
was 0 and, at the end of the first 100 seconds, you'll have the average
over that time, which all things being equal, should be half that of
the short-term average value during that 100 seconds. Now, you turn it
off for the next 100 seconds, and you'll be averaging that first value
with the remainder of the time, where the value is 0.

Given that, as a very rough guess, I would expect, then that the value
at the end of the first 100 seconds to somewhat less than the full value,
maybe like - 6 dB, and the value at the end of 200 seconds to be less, like
maybe -12 dB.

If you wanted to closer to what you expect, run a constant level noise
source for quite a long time, maybe 15 minutes, then shut it off, and
examine it 100 seconds later. THat may give you something along the
lines you expect.