On 10 Sep 2004 at 21:15, Caveman wrote: > It is squeezing, at least those extra bits as from 8 (jpeg) to 12 bits > (sensor) and you also get rid of some compression artifacts. However, if > the camera did a good mapping of those 12 bits in 8 in the first place, > and used a good sharpening method without ugly artifacts, and would > allow you to set the desired jpg compression level in fine steps, etc > etc you wouldn't need to do that yourself on the PC.
You've answered the question, there is no good way to map 12 bits to 8 bits particularly if you wish to tweak the image (any kind of gamma correction for instance) during post processing. Reducing bit depth compresses the colours (depending upon the colour space reference) and the brightness range which translates to larger steps between any colour and brightness level. > So yes, there is a place for better firmware, and if this is "Canon > thinking" then IMHO Canon is the way to go. There will still be a place for RAW. The current implementation of in camera RAW or JPEG/TIFF (with their inherent compromises) works quite well. I think that the Canon concept of writing out a RAW and JPG of the same file is a waste of time and resources too, why not also produce a TIFF and PNG as well? It's so easy to batch these functions after the fact, if you have a RAW file you aren't limited in how you can post process the image (which in camera processing is too it just happens sooner and leaves you without access to the original data). Rob Studdert HURSTVILLE AUSTRALIA Tel +61-2-9554-4110 UTC(GMT) +10 Hours [EMAIL PROTECTED] http://members.ozemail.com.au/~distudio/publications/ Pentax user since 1986, PDMLer since 1998
