Newsgroups: sci.image.processing
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!uunet!news.erinet.com!netcom.com!tgl
From: tgl@netcom.com (Tom Lane)
Subject: Re: Lossless JPEG
Message-ID: <TGL.95Apr10221129@netcom8.netcom.com>
Sender: tgl@netcom8.netcom.com
Organization: Netcom Online Communications Services
References: <1995Mar29.171818.12391@lds-az.loral.com> <3mb6pt$kbb@trog.dra.hmg.gb>
	<EDDINS.95Apr10100619@grainger.mathworks.com>
Date: Tue, 11 Apr 1995 05:11:29 GMT
Lines: 22

A clarification and a comment:

1. Yes, there is a lossless JPEG standard.  It has hardly anything
in common with lossy JPEG except that it was developed by the same
committee.  Typical compression is only about 2:1.  I know of two
free implementations, one from Cornell and one from Stanford -- check
the comp.compression FAQ or JPEG FAQ for pointers.

2. The recently proposed PNG format is lossless and gets as good or
better compression as lossless JPEG on most continuous-tone images.
It also works well on colormapped images, whereas lossless JPEG does
not.  Unfortunately, PNG is still vaporware --- free software is under
development but is not yet released.  This will change in a matter of
days or weeks.  Keep an eye on comp.graphics for announcements.

Lossless JPEG has never achieved any great popularity, and I suspect
that the advent of PNG will consign lossless JPEG to the dustbin of
history.  (But lossy JPEG is alive and kicking :-))

			regards, tom lane
			organizer, Independent JPEG Group
			member, PNG development group
