Lourens Veen wrote:
On Wednesday 05 March 2008 10:57:28 James Richard Tyrer wrote:
Just a nit pick, the q factor in JPEG compression is not a
percentage, it's a number between 0 and 1.
So, if the range 0 to 100 is used (as most applications use) it is percent.
http://www.oreillynet.com/pub/a/javascript/2003/11/14/digphoto_ckbk.h
tml
IIUC, JP2 gives better quality for lossy compression at a given
compression ratio. A major advantage is that it doesn't compress
in blocks so doesn't have the blocks like JPG has. And, lossless
JPEG 2000, the JPF file format, can be used to store the image in a
RAW file.
The only problem is that it is slower than regular JPEG. The
computations are faster but there are more of them since it does
the image pixel by pixel rather than in blocks.
Does it do more than 8 bits per channel?
Yes
What about floating point for HDR imaging?
AFAIK, ADCs don't output floating point (i.e. only <= 24 bit integers)
so this isn't a concern for a camera.
However, part 10 will implement direct (i.e. numerical) compression of
IEEE float data. Other compression methods compress IEEE float as
though it were an integer which leads to issues if there is a wide range
of exponents.
http://portal.acm.org/ft_gateway.cfm?id=1284114&type=pdf
{Membership Required} but there is fair use. Contact me by private
email if you want a copy and aren't a member.
And wasn't there a patent issue with wavelets in image coding? Does
JPEG2000 work around that somehow?
Yes, there are patent issues. However, this (actually part 1) has been
adopted as an ISO/IEC standard making it royalty and license-fee free.
--
JRT
_______________________________________________
Open-hardware mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-hardware
