naturerecordists
[Top] [All Lists]

RE: MD v. DAT

Subject: RE: MD v. DAT
From: Gianni Pavan <>
Date: Mon, 18 Aug 2003 15:30:16 +0200
Hi all,
        I have a lot of respect for both Raimund and Walt. I learned a lot from 
both and I'll like to read again their postings.

On the technical side, I should agree more with Raimund than with Walt. 
ATRAC artifacts "may" be negligible to our ears, though they "are" there in 
unpredictable ways as unpredictable are the infinite combination 
possibilities of incoming sounds. Though I've no direct experience with MD 
as I never found an ATRAC software to make my own tests without all the 
possible artifacts induced by the other components of a recording chain. 
Probably next winter I'll buy a MD machine to play with as I played (...for 
testing, of course) with a lot of DAT decks and sound boards (I discovered 
a lot of interesting things but I never had the time to publish my results).

I made some tests with MP3 and I found that it behaves (apparently) 
unpredictably when different and non musical sounds combine together. I got 
artifacts very similar to those showed by Raimund. MP3 and ATRAC are both 
based on psychoacoustic models to cut details supposed negligible to our 
ears. By definition, they are based on algorithms that must cut something.

And now just an example of one of the artifacts in MP3 I found intolerable 
for my needs.
I compressed sperm whale clicks in recordings with >90dB dynamic! Clicks 
are very loud, very short (and thus broadband) impulsive sounds. Just after 
the main click there are low level components generated inside the head of 
the whale, including reflections that carry an information about the size 
of the spermaceti organ. The reflection to be used for assessing the 
spermaceti size is after 4 to 7 ms the main click with a level that is 
often 60 dB less than the main click!! The MP3 encoder "thinks" we can't 
hear anything after a loud broadband click and thus cuts everything for 
more than 10 ms! Thus after each click I have a short silence instead of 
the subtle components I'm looking for.  Unfortunately, I never made the 
same test with ATRAC.

This example is just to say that absolute perfection is not of this world 
and that each real technical solution 1) may have different negative and 
positive aspects and 2) may fit well certain tasks and not others.

Walt and Raimund, I hope to learn more from both you,
Gianni

--------------------------------------------------------------
Gianni Pavan
Email 
Centro Interdisciplinare di Bioacustica e Ricerche Ambientali
Universita' degli Studi di Pavia
Via Taramelli 24, 27100 PAVIA, ITALIA
Tel/Fax   +39-0382-525234
Web       http://www.unipv.it/cibra




________________________________________________________________________
________________________________________________________________________

<Prev in Thread] Current Thread [Next in Thread>
Admin

The University of NSW School of Computer and Engineering takes no responsibility for the contents of this archive. It is purely a compilation of material sent by many people to the naturerecordists mailing list. It has not been checked for accuracy nor its content verified in any way. If you wish to get material removed from the archive or have other queries about the archive e-mail Andrew Taylor at this address: andrewt@cse.unsw.EDU.AU