Originally posted by jeffkrol Unfortunately, the same pattern can be seen at iso 400. Download Herb's test DNG. That's at 400 -2EV, of a grey card w/ even illumination. I pushed it 3 stops in Silkypics.
See this post and screen capture. Honestly, if I was paranoid I'd think people were ignoring me...
It is reasonable to see the same pattern also at lower ISO values, it is the same sensor and technology used regardless of ISO value. I'm not that surprised. Then it will always be something to discuss: is it reasonable to exposure a grey card with -2EV and later push it 3 steps? Some will say yes (with your motivation), some will say no. I know that when I have pushed shadows in DS pictures I have run into noise pretty quick. So, I am one of those wanting to be able to push the dark areas to make details visible. Getting pattern noise like that is not what I want. Once again I have to nag about Photo_mom's salt package picture: There we had a picture not pushed at all, and with no attempts to lighten up the white wall. Still, that picture was ruined by the pattern noise.
The color shifting... If this is a real problem, shouldn't wee see it in nearly all pictures then? For the moment I guess that the shutter runs horizontelly in the K10D and that the shutter time and the fluoroscent light didn't came out well together. Can the same be observed using a longer shutter time? If so it is real bad and I would guess the camera is faulty. If all K10D cameras behave slike that, then it's a question about bad engineering. I would be surprised to find a modern dSLR camera to behave like that. No, I would like see this with continuous light, or with a longer shutter time, before drawing any conclusion at all.
Of course you are ignored: I quote this from a email conversation I had with a member at DPR (who will remain anonymous here): "30% have their monitor calibrated too contrasty, so the noise is buried. Another 20% hear "banding" and think of the pixel-banding the D200 had, not pattern noise. Another 15% cannot tell wall texture from noise patterns, 15% upgrade from a P&S digicam and are happy with that "little bit" of banding, 10% are blind brand loyalists..."
My reply to that was: "...brand loyalists are more than 10% and max 50% can tell wall texture from pattern noise. 75% doesn't care about what the words they type really means and 90% don't read the threads before posting"
regards,
/Jonas