Sune T. B. Nielsen wrote:
Under technical specifications in manuals for e.g. microphones and
headphones there is often some measurement called frequency response.
This is usually meaningless.
I guess the meaning is to picture how high and low frequencies the device
manages. But would this not vary with the amplitude/volume? Or A) is it
reflecting the average of different volumes B) is there a specific volume
that is used as standard?
Right. A real frequency response specification has a tolerance to it.
Like, "20 Hz - 20 KHz +/- 3dB". If the tolerance is not there, the
specification is totally meaningless and is a sign that the manufacturer
is not taking you seriously.
How are such measurements done? I would like to measure five different
ghettoblasters/boomboxes in order give a somewhat objective idea of what
they are capable of and how 'good' they represent different (areas of)
frequencies.
In the case of the numbers on the datasheet, they are normally made up
by the marketing department.
If you want to make actual measurements, the cheapest way to do it is
probably with a signal generator and an SPL meter. The cheap Radio Shack
meter will do the job. Set the meter for C weighting, then put the
boom box up on a ladder in the middle of a big meadow with nothing around
it. (You really want an anechoic chamber, but if you don't have access to
a chamber, you can do measurements outdoors to get reasonable ballpark
numbers). Put signal in, measure the output level. Sweep back and forth
and you'll probably find a couple frequencies where there are big peaks
and dips. Plot it out on log/log paper. (You can then subtract the
C-weighted response of the meter, but since the meter response will only
be non-flat at low frequencies and your low frequency measurements won't
be much good without a real chamber, it's probably not worth the trouble).
There are some fancy impulse measurements that you can do, which will give
you better accuracy without a chamber, but do a swept-sine by hand a few
times first to get a sense of how it works.
What equipment is used and what sounds?
Either a swept sine is used, with equipment that automatically plots the
response, or an impulse is sent through the thing and recorded, and the
frequency response derived mathematically from the impulse response. The
impulse response trick (MLSSA is typical software) allows you to deal more
easily with room effects but requires some compute power to do.
You will find the general measurements on a boom box vary widely with
direction, too. On-axis the response will be different than 45' off-axis.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
|