NikonWeb.com

As you know, Kodak does not provide software support for the DCS200 and the twain drivers are a bit old. You can transfer files to disk, but then must "change file" one at a time without a mass preview function.

So I decided to write a ".kc2" to ".bmp" convertor TODAY. I will probably move it to .tif's later, but .bmp is easier and works. I spent a few hours staring at Hex Dumps of the .kc2 files. It finally occurres to me that the raw images were bigger, and had calibration rows in them. I also found out the raw images themselves have a more active image area than what the Twain driver imports into Photoshop AND the original files look better! The twain driver applies some grey-scale correction that gives up contrast! I like my original better!

Raw file made into .bmp:
http://mywebpages.comcast.net/brianvswe ... ds1BMP.JPG

Raw file imported into Photoshop using the twain driver: no image corrections used.
http://mywebpages.comcast.net/brianvswe ... DS1TWN.JPG

What do you all think? Anything to the "we do not want to be locked into one Raw file processor" debate? At least to me, The .kc2 images are bigger (1536 columns vs 1524) and better (grey scale) than "twained" image.

I am going to write a "batch" .kc2 to .bmp or .tif convertor. It will probably be a DOS compatible .exe, written in FORTRAN and MACRO. Nothing fancy, but at least you can use a modern version of Photoshop with the raw data.

Well done Brian!

I agree that your version is better. Will it work with DCS200 color images as well?

(When ready, I'd be happy to host your batch converter on nikonweb.com if you need a home for it.)

Thanks,
Jarle

I believe that the raw format for the color images are the same, so an interpolation routine would be required to build the 24-bit color "pixels". The .bmp format handles RGB easily. I'll read into what algorithms are used, I have cubic spline routines available. Tiff would allow storage of meta-data. The data for date, SN, firmware rev dates, camera manufacture date, are in readable ascii in the file header of the DCS200 and DCS420c. The 420c uses a "compression" of sorts. I am going to Email Kodak for possible format information. Otherwise I will be playing with it as well. Just curiousity, Kodak provides support for it. I would like to squeeze out those extra pixels from the CCD though (if they are there)!

You know, looking at the top of the raw->bmp image, maybe that is not calibration data but are really "Virtual Sprockets". This camera was made in 8/1992 according to the stored date in the file. Just made the Kodak guys feel better.

TIFF with metadata would be my preferred output format. Now I just need to get my DCS 200

Jarle