From: Jeremy Minns <>
>
> Seth,
>
> I can offer some help on two of your questions.
>
>
>>First, I am a little concerned about the accepted legitimacy of
>>MD-compressed bird songs for scientific archival.
>
>
> In a 1999 posting to NEOORN Morton e Phyllis Isler (ornithologists at the=
> Smithsonian and specialists in bird vocalisations) put the matter very cl=
early:
>
> "About a year and a half ago, NEOORN had an excellent back and forth
> discussion about Minidisks, but a number of you may not have been enrolle=
d
> at that time, and Charles Duncan's quote from the Budney [curator
> of Macaulay Library of Natural Sounds] and Grotke [MLNS audio engineer]=
> paper belittling Minidiscs requires a renewed response.
>
> The key question is whether audio engineer Grotke's conclusions about
> deficiencies in the Minidisk are relevant to its application to recording=
> and using avian vocalizations. Our conclusion, after years of using
> Minidiscs, is that his conclusions are not relevant. We have also put
> Nagra, cassette, and DAT recordings directly into CANARY (the Cornell
> Bioacoustics Workstation) and compared these with the same recordings put=
> onto Minidisks and then into CANARY. We have seen no differences in the=
> displays (e.g. spectrograms) and because we measure vocal characteristics=
> from displays, we can see no difference in vocal measurements that we
> take. We have never found the slightest evidence of sound distortion
> caused by the MD."
The original paper that was so negative about minidisc is still on
Cornell's site in it's section on equipment recommendations. (as well as
other copies at other sites)
As I posted a while back, this paper, which is the primary source of
most criticism of minidisc for nature recording, is false. The only flaw
illustrated, involving the recording of 6.5khz & 7.5khz signals and then
doing a sonogram on the resulting sound from the minidisc recording is
bogus as far as showing a flaw with ATRAC. This apparent flaw is easily
reproduced with digital equipment, all you have to do is record with a
signal 10-15dB over the clipping level. What it really shows is clipping
artifacts.
I easily reproduced this sonogram using my portadisc. When you feed the
original signal in set well up into clipping, here's the before and
after sonogram:
http://frogrecordist.home.mindspring.com/naturerecordists/Clip_6.5_7.5.jpg
Note by the color code that the main signal has been amplified to cause
the clipping. The intensity of the subbands can be altered by changing
the gain within the clipping range.
And if you simply lower the gain to where no clipping occurs (to the
correct record level for digital) this is the before and after, note all
the ATRAC flaws in the 2nd half of this sonogram:
http://frogrecordist.home.mindspring.com/naturerecordists/MD_6.5_7.5.jpg
If you look closely you can find the before and after transition point
as a change in color, I did not exactly balance before and after levels.
I'm sure if everyone thinks about it a bit they will sort out why such
bands are produced by clipping. And resolve to make sure they set the
gain levels in their recordings to avoid clipping.
If we are being charitable one can assume that those folks, who so
strongly advocate analog methods even today just simply did not
understand what clipping does in the digital recorder. So they recorded
using their analog recording habits, and of course those caused
clipping. We can then say they still did not recognize it and published
their paper. The paper is clear evidence of someone who did not
understand digital recording.
The real crime is this paper is still the main reference on Cornell's
own site about minidisc. Still misleading new recordists, still used as
a reference to support those who don't like minidisc.
http://www.birds.cornell.edu/lns/recordingnature/html/recordingnature_teche=
quip2.html
I often wonder what things would have been like if they had of known
what they were doing back then and tested correctly. Correct testing of
perceptual encoded material is not near as simple as folks seem to
think. The test used in the Cornell paper is highly inappropriate for
finding out anything about how well a perceptual encoder is doing. The
choice of test is another indication of how inexpert the originators of
the paper were. But that's a separate subject.
I've spent many years doing scientific survey work using minidisc as the
recorder. The results of my work are highly admired, and form a major
block of the scientific information on frogs in Georgia. 37 CD's filled
with recordings from over 800 sites are part of the scientific archive
for Georgia. None of the scientists I work with have had any qualms
about minidisc. Except to envy it, many of them are still using cassette
due to lack of funds to buy new equipment. And these folks are experts
on the calls I record. Even the bird experts pick through my stuff for
the incidental birdcalls I recorded.
The Georgia frog CD is all minidisc recordings, except for a few clips
from others recorded by cassette.
I would agree with the conclusion stated above, there is no practical
problem with minidisc recordings for nearly all scientific work. And in
the few places where it would be a problem especially calibrated and
designed equipment is used anyway, not general purpose recorders and
mics of any type. (and a lot of that work does not need that kind of
equipment) So a hobbiest recordist will not have the opportunity to
contribute where minidisc (or most all formats) would not be acceptable.
The primary way a general nature recordist can contribute to science is
in documentation of range. For this a clear recording that a expert
listener can identify is what's needed. But, in actual practice it's
often the case we have to tolerate what we can get, like recording a
group of frogs next to a freeway. The real place that making sure your
work is scientifically valuable is actually in the field log. Nowadays
you should document location with a GPS, record notes on abundance,
habitat and so on. Record environmental conditions. The more info the
better. Without those notes your recording is near scientifically
worthless. It's a good idea to develop a standard field card you use so
you won't forget to note things. And then use it.
You might note that the use of call recordings to document frog
populations was a not well accepted idea when I started. Did not matter
what recorder. Bodies in bottles was still the norm. It's much more
accepted now, and I like to think I contributed some to that. It was a
deliberate risk we took when the survey started. The quantity and
quality of the recordings I did made it impossible to talk about frog
distribution in Georgia without referencing some of it. That issue is of
far more import in science than minidisc trivia.
Walt
________________________________________________________________________
________________________________________________________________________
|
