Regarding the recent post about digital audio recording, I beg to differ
with the assessment given.
Just about every professional sound device being released by the music
hardware industry these days has a working sample rate of 96 kHz. I just
bought a Digi002 rack from Digidesign which is a rack-mount portable device
that attaches to any computer running protools by firewire. It has multiple
ins and outs and some very sophisticated internal hardware for D/A and A/D
conversion. Because of the Apple open spec called coreaudio, you can use it
with just about any piece of new audio software. It comes bundled with
Protools which is the recording software of choice for just about everybody
in film and music. I use it a lot out on the water on a powerbook making
ample use of their high end noise reduction unit called DINR which gives me
512 bands of mapped EQ control to notch out any or all or even no part of
water current, boat noise, or anything else that mars the recording but
which appears in a different frequency band then the whales. Sometimes I
use it in tandem with my Tascam DAT machine although that is severely
showing its age with a 48 kHz sample rate.
I'd avoid the NI card, if at all possible, and absolutely if you are not a
programmer, because it won't work with anything but their own proprietary
software which you then have to program to use. But I'm not officially a
scientist so I don't need the kind of calibration it offers. On the other
hand, I don't think digital recording software, with its bits and bytes,
actually needs that kind of calibration any more, although, of course, the
hydrophone and the specialized wide band preamp still do, although NI can't
help you there. Much of the issue is not calibration, but something much
simpler called zero gain. You can read about it in just about any book
about recording. In fact, anyone getting involved in field recording ought
to read at least one text about the basics of recording.
I also notice the advent of a brand new generation of portable professional
hard drive recorders mostly used by the film industry with effective sample
rates of 196 kHz and even higher, so you can record frequencies up to nearly
100 khz. These units range in price with some units going for about $2500.
I also believe there's a lot better hassle-free, ergonomically designed
stuff than Ishmael available for a nominal fee on the internet. However, I
just know the mac stuff and Amadeus is one of the best for the money. The
old but wise PRAAT has some phonetics graphs that might really surprise
anyone doing cetacean call analysis. For those wishing more sophisticated
graphing tools, our group has been happy using Matlab to generate wavelets
of several cetacean species. You can see some of the results in a gallery at
interspecies.com.
on 12/29/03 10:31 AM, Dave Mellinger at [EMAIL PROTECTED]
wrote:
> Yes, the mic/line-in port on most sound cards has a maximum sampling rate
> of 48 kHz, so the maximum frequency it can capture is half that or less.
> Typically you see good frequency response up to 20-22 kHz.
>
> To get higher frequencies, you need a more specialized data acquisition
> card. I've used National Instruments cards with success, specifically the
> DAQCard 6062-E, which is a PC card (PCMCIA card), and the PC6071-E, which
> is a PCI board. These should give you single-channel input at the
> frequencies you need.
>
> You may also want to try out Ishmael, which can record sound onto disk
> according to a schedule you set up, display real-time spectrograms, detect
> calls, localize calls, and a host of other tasks. It can interface to the
> above NI cards as well as some other cards. It's freely available at
> <a href="http://cetus.pmel.noaa.gov/cgi-bin/MobySoft.pl";
> rel="nofollow">http://cetus.pmel.noaa.gov/cgi-bin/MobySoft.pl</a> .
Best Regards,
Jim Nollman
Interspecies Inc.
<<a href="http://interspecies.com/";
rel="nofollow">http://interspecies.com/</a>>
|