Read this off another forum,
Comments and feedback was all by Bjørn Rørslett.
Bjørn Rørslett said:In my test shots I can get signs of "banding" if the image is grossly underexposed (more than -3 stops), then has the shadows lifted over 5 stops. When an image is mistreated like this some banding is not your main problem, since the image as a whole is in a pretty badly shape anyway.
Bjørn Rørslett said:Yes, tried to shoot directly into the sun to provoke "banding", and still had to abuse the image grossly to get the weakest signs of anything. So I'm not convinced this is a real issue. Blooming of the sensor was a real issue for D70 (CCD) on those subjects but not on D200(CCD) or D2X (CMOS).
I've seen those allegedly supportive images on the web and am puzzled as to how the comb-like pattern arises. From most DSLRs you can get "banding" which really is a precise recording of the phases of alternating current (AC), 50 Hz here in Europe. You can easily see this if you shoot rain on a dark night in the vicinity of a street light, the raindrop trails get "zebra-striped" crosswise due to the AC. I observed this first time with D1 and have recorded similar will all later Nikon models.
The comblike weave on D200 might be a "beat" or interference pattern between CCD read-out and AC current of the light source. Really hard to tell for the time being and not an easy matter to isolate for testing.
Bjørn Rørslett said:This is an example of the degree of abuse I need to employ before any sign of the tiniest banding occurs. See legend below for details.
Image illustration
Legend as follows,
A. Original. Exposure -3 stops below meter reading to show sun rays.
B. Added +2 EV to basic exposure in Bibble Pro 4.5, then increased "exposure" further by curve tool so as to bring RGB-levels above 12 into clipping.
C. 200% crop of original. Note no signs of CCD blooming, no signs of shadow noise.
D. 200% crop of abused image (B). Barely visible signs of banding noise.
Taken with 28 mm f/2 Nikkor AI at f/22, 1/160 sec, ISO-equivalent 100.
Does this test show that there is an issue with D200? Absolutely not.
Comments and feedback was all by Bjørn Rørslett.