I was doubtful whether to post this article in Digital Signal Processing forum, or here in the Shooting. Actually what Fuji's Super CCD does is rather a shooting technique than a processing.
For the first time I have been able to fiddle in a HDR shot provided by the Fuji S3 camera where a high contrast scene was shot, and I am amazed at the real captured dynamic range; definitively a Canon or Nikon is far from being able to do this in just one shot.
Fuji's sensor generates two independent images in just one shot. In DCRAW's manpage is said that they are 4 f-stop apart (does this figure mean anything to you?). I have blended both with my bending routine and the program calculated that the two images were 3.6EV apart, and this is the exposure correction that was used. In fact if we load both linear images in PS and correct the most exposed by -4EV, it becomes a bit too dark so I trust more the calculus yielded by my program.
The image is acceptably free of noise in all its pixels, so looking at the
logarithmic histogram:
These are the partial images:
And the blending result (ACR and other developers perform this blending automatically, in DCRAW I had to extract them one by one with the -s option):
I think what the Fuji sensor does is fantastic. Exactly the same concept as shown here:
ZERO NOISE PHOTOGRAPHY.
NOTE: I know this is nothing new, but this is the first time I put my fingers on a Fuji RAW and wanted to share because I was really surprised at the results.