[OSGeo-Discuss] 'lossless' JPEG2000

[Miles Fidelman]

Michael P. Gerlek wrote:
> Bruce-
>  
> It is not clear to me what sort of "study" you would need to convince 
> you, as the ISO standard for encoding data into the JPEG-2000 file 
> format is by construction mathematically and numerically lossless 
> process.  (Indeed, "compression", i.e. throwing away bits so as to 
> further reduce storage requirements, is actually not defined within 
> the scope of the standard.)
>  
Or to put that just a little more precisely:

- lossless compression involves throwing away redundant bits to reduce 
storage and/or bandwidth requirements

- lossy compression involves throwing away non-redundant bits - 
accepting some level of irreversible loss of quality (e.g., resolution, 
compression artifacts) in order to reduce storage and/or bandwidth 
requirements

As someone else noted, you need to vet the math, and it's implementation 
in hardware/software, to determine whether a process is truly lossless.

Also worth noting: error recovery is a different, but related issue.  A 
lossless compression algorithm may be reversible (uncompression returns 
the original data, exactly) - but it may also be brittle.  Under 
practical conditions, one has to look not just at the level of 
compression achieved, and the costs of that compression (how much 
quality is lost, if any; computational cycles required) - but how 
resistant the coding scheme is to disk and/or transmission errors. 

Which raises the question:  anybody have any idea what the 
characteristics of the JPEG-2000 file format, and particularly the 
wavelet transforms, are vis-a-vis bit- and multi-bit errors?  I.e., how 
much error recovery, if any, is built into the coding scheme?


-- 
Miles R. Fidelman, Director of Government Programs
Traverse Technologies 
145 Tremont Street, 3rd Floor
Boston, MA  02111
mfidelman at traversetechnologies.com
617-395-8254
www.traversetechnologies.com