[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Information, Compression, and storage.



Some random thoughts...


After having looked at imagery produced with lossy compression it is hard to
argue the fact that, for human-viewable consumption, the images produced can
be superb.  The only problems with lossy compression should be in a very
narrow field of application where compositing/effects/coloring need as much
info as possible.  Archival is another area of concern.

On the question of lossless compression:
True lossless compression does not exist (refer to the USENET Compression
FAQ below, among other sources).  That is, given a perfectly
random data set, no algorithm can be had that will represent that unique
data set in a more compact form without incurring some loss.  The smallest
representation of the data set is the data set itself.  The fire test of
lossless compression algorithms is recursion.

All lossless algorithms take advantage of repeating values, value patterns
or bit patterns in one way or another.  All compression algorithms also take
advantage of having lots and lots of data to look for these patterns.  Video
is great on all of the above.

I don't know of any lossless compression algorithm that can guarantee a
fixed compression ratio and fixed processing time per unit of data.  That's
why it would be difficult to use in a real-time imaging type application
with current technology.

As a valuable experiment, try to think of a way to reliably compress five or
ten bytes of random data by just 20% (in other words, turn five bytes into
four bytes, etc.).



Some neat compression sites I've learned from:

Lossless compression
http://www.cs.sfu.ca/cs/CC/365/li/squeeze/

Wheeler's block algorithm
http://www.research.digital.com/SRC/JCAT/jdk11/wheeler/

A nice list of compression links
http://www.internz.com/compression-pointers.html

Signal processing at Stanford University
http://www-isl.stanford.edu/~gray/iii.html

USENET Compression FAQ
http://www.cis.ohio-state.edu/hypertext/faq/usenet/compression-faq/top.html



-Martin





---
Thanks to Rich Torpey for support in 1999
No advertising/marketing allowed on the main TIG.  Contact rob at alegria.com
anonymous messaging now at http://www.alegria.com/HyperNews/get/ubique.html
1046 subscribers in 41 countries on Thu Jul  8 01:50:18 CDT 1999 
subscribe/unsubscribe with that Subject: to telecine-request at alegria.com
complete information on the TIG website http://www.alegria.com/tig3/