To: moore(_at_)cs(_dot_)utk(_dot_)edu
Subject: Re: Massive Content-Type definition ideas & Gopher
Date: Mon, 7 Jun 93 11:39:02 PDT
Keith Moore wrote:
It's not anywhere nearly that bad. First of all, we don't need multiple
compression algorithms -- we need one that everyone can use freely.
I slightly disagree with this. There are 2 types of compression: 1)
content-type specific compression algorithm, 2) general purpose
compression.
The content-type specific compression is tuned specifically for a certain
type of data, e.g. audio uses adpcm audio compression. Unless we create
a new subtype for the compressed audio data [...]
I propose that we do exactly that. audio/basic is already compressed in the
sense that 14-bit samples are mapped onto an 8-bit mu-law encoding. Most
audio compression algorithms are lossy, and stacking one compression
algorithm on top of another really trashes the signal. I conclude that we
don't want to add further compression to audio/basic (the only audio type
defined so far), and that other audio compression schemes should go in a
different audio subtype.
Keith