Mike Young wrote:
"Henry Kolesnik" wrote in message
m...
I'm not certain to which group I should post this question, so I'm shot
gunning hoping to find it. If anyone can point me, please do.
However....
Has anyone used the HP 3582A audio spectrum analyzer to plot the freq
response of audio devices like mikes, speakers, etc. Knowing the mike
response I'd like to see the spectrum of plucked guitar strings and their
harmonics and see how different manufacturers sound varies in a plot. I
got a 3582A at a swap meet, its complex and probably overkill but that's
what I have and I am anxious to try it out and would appreciate any tips.
Thanks

Another option, FWIW, is to use the PC to capture and analyze. The audible
spectrum would seem to be quite unchallenging for common desktops.
Compression schemes such as MP3 and WMA, after all, operate on the frequency
domain. In fact, Windows Media Player delivers the FFT stream to
visualization modules. There's likely free software somewhere that can do
what you're looking to do. OTOH, you then have the soundcard's
non-linearities to deal with, which is likely not on par with lab or even
studio grade equipment.

Actually, pretty much any modern PC audio card will have plenty low
enough distortion for the application, and quite likely better than the
HP3582. The 3582, is, after all, quite an old instrument. The advent
of digital audio changed the whole playing field for audio range ADCs
and DACs, as well as support parts like op amps (though codecs as used
in most inexpensive PC audio systems don't need additional preams on
their inputs anyway). Frequency response at the low end can be an
issue; there are many vibrations that are too low in frequency for a
sound card to deal with. Think "seismic" and you'll understand. The
AC coupling in the HP E1433 is programmable to very low frequencies, as
it's actually a feedback control system run through the DSP following
the ADC, but most AC coupling will have a 3dB corner frequency around
1Hz, or higher if it's really just for audio. Sound cards may well use
codecs whose ADCs have digital AC coupling, as well, and the card may
not provide a way to turn that off.

Another package that can be used to do FFTs on captured audio, and to
otherwise manipulate the audio file, is the free Scilab package. You
can play to your heart's content with math functions on the audio (or
on pretty much anything else you can represent numerically), but of
course you can get yourself into trouble if you don't understand what
you're doing. And along those lines, I'd strongly caution you to not
read too much into what Scott wrote, "Yes, lots of people have used HP
FFT analyzers for all kinds of things.
There isn't a lot of trickery... they pretty much work." The
_analyzers_ may work, but if you don't understand what they do, you may
make very bad measurements with them, and you may badly misinterpret
the results you see. There are, I believe, HP/Agilent ap notes on
making _good_ measurements and properly interpreting the results.
Well, I KNOW there are, though finding them on the web may be a bit of
a challenge.

Cheers,
Tom

"K7ITM" wrote in message
ups.com...
Mike Young wrote:
"Henry Kolesnik" wrote in message
m...
I'm not certain to which group I should post this question, so I'm shot
gunning hoping to find it. If anyone can point me, please do.
However....
Has anyone used the HP 3582A audio spectrum analyzer to plot the freq
response of audio devices like mikes, speakers, etc. Knowing the mike
response I'd like to see the spectrum of plucked guitar strings and
their
harmonics and see how different manufacturers sound varies in a plot.
I
got a 3582A at a swap meet, its complex and probably overkill but
that's
what I have and I am anxious to try it out and would appreciate any
tips.
Thanks

Another option, FWIW, is to use the PC to capture and analyze. The
audible
spectrum would seem to be quite unchallenging for common desktops.
Compression schemes such as MP3 and WMA, after all, operate on the
frequency
domain. In fact, Windows Media Player delivers the FFT stream to
visualization modules. There's likely free software somewhere that can do
what you're looking to do. OTOH, you then have the soundcard's
non-linearities to deal with, which is likely not on par with lab or even
studio grade equipment.

Actually, pretty much any modern PC audio card will have plenty low
enough distortion for the application, and quite likely better than the
HP3582. The 3582, is, after all, quite an old instrument. The advent
of digital audio changed the whole playing field for audio range ADCs
and DACs, as well as support parts like op amps (though codecs as used
in most inexpensive PC audio systems don't need additional preams on
their inputs anyway). Frequency response at the low end can be an
issue; there are many vibrations that are too low in frequency for a
sound card to deal with. Think "seismic" and you'll understand. The
AC coupling in the HP E1433 is programmable to very low frequencies, as
it's actually a feedback control system run through the DSP following
the ADC, but most AC coupling will have a 3dB corner frequency around
1Hz, or higher if it's really just for audio. Sound cards may well use
codecs whose ADCs have digital AC coupling, as well, and the card may
not provide a way to turn that off.

Another package that can be used to do FFTs on captured audio, and to
otherwise manipulate the audio file, is the free Scilab package. You
can play to your heart's content with math functions on the audio (or
on pretty much anything else you can represent numerically), but of
course you can get yourself into trouble if you don't understand what
you're doing. And along those lines, I'd strongly caution you to not
read too much into what Scott wrote, "Yes, lots of people have used HP
FFT analyzers for all kinds of things.
There isn't a lot of trickery... they pretty much work." The
_analyzers_ may work, but if you don't understand what they do, you may
make very bad measurements with them, and you may badly misinterpret
the results you see. There are, I believe, HP/Agilent ap notes on
making _good_ measurements and properly interpreting the results.
Well, I KNOW there are, though finding them on the web may be a bit of
a challenge.

Cheers,
Tom

Hi- there is a HP newsgroup on Yahoo, try that. Jim.