Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!apple!fernwood!lia!jgro
From: jgro@lia (Jeremy Grodberg)
Newsgroups: comp.compression
Subject: Re: Decompression algorithm for CCITT Group 4
Message-ID: <1991Mar26.234038.3855@lia>
Date: 26 Mar 91 23:40:38 GMT
References: <1991Mar25.210709.21316@cuff.uucp>
Reply-To: jgro@lia.com (Jeremy Grodberg)
Distribution: na
Lines: 28

In article <1991Mar25.210709.21316@cuff.uucp> tom@cuff.uucp (Tom Flaherty) writes:
>
>	Does anyone have or can anyone point me to a fast
>	algorithm for decompressing CCITT Group 4 data?
>	Does anyone know of any references for compression
>	/decompression techniques?

The Group 4 spec can be found in CCITT Fascicle VII.3  - Rec. T.6.
Unfortunately, I don't know where to get it.  Maybe Global Engineering
Documents in Southern California (Irvine?).

Sun sells some fair code for decompressing Group 4 compressed Data.  
I've seen a number of proprietary implementations, and they all hover 
around 1.25 Million pixels per second output from the decompressor on
a Sun IPC.  Sun's code is considerably slower, but I don't have any good
figures on exactly how much.   

The speed of the decompression is largely based on how effectively it
utilizes the computer's native bit instructions, memory I/O, and caching 
architecture, so you may well find that a "wicked fast" piece of code on
one machine is a real dog on another.  For example, Sun's code is slow
on a SPARC machine, but may very well be fast on a Sun-3 (Motorola 680X0)
machine (I haven't tried).


-- 
Jeremy Grodberg      "Show me a new widget that's bug-free, and I'll show
jgro@lia.com         you something that's been through several releases."