Dieter wrote:
How do you "fix" a defective pixel? Set it to the average of the surrounding pixels? The extreme perfectionist would prefer that the data include a list of defective pixels, so that they can be dealt with as the photographer chooses in post-processing.
IIUC, it is done by interpolation. I think that this is 3rd degree rather than 1st degree (linear). Perhaps it even uses fuzzy logic to avoid the problems with higher order interpolation. Yes it would be good if the camera would let you download a list of "stuck" pixels. Can this be included in DNG metadata? I don't know if there are methods that could do better than interpolation, but the nice thing about this (having the camera fix pixels)is that it doesn't destroy any data, you can always fix them again if you have the necessary list of bad pixels for the camera. The camera does need to fix the pixels since it is very necessary for lossy compression of the image.
For some professional "still" applications, photographers like the ability to shoot a rapid burst in hopes of getting one shot that is just right.
I think that the camera stores the True RAW data in RAM and then processes it after the burst.
I still worry about condensation. :-(
Yes, it would be a problem. I wonder what professional (6x6) backs do. Since the digital back has to replace a standard back (e.g. Hasselblad, Bronica) I don't think that there is any room for an Argon filled cell in front of the sensor. I don't own a Bronica but I used to sell them and although the pressure plate doesn't retract like it does in 35mm Contaflex or Contarex IMBs (and I think the Rolleiflex SL2000), there is very little distance from it to the front of the back. Although a digital back doesn't need a dark slide. This lack of a dark slide might make it possible. I would think that they have glass to protect the sensor (would have to be very thin) in any case. Obviously, space isn't an issue with the CCD for a large telescope. :-)
I don't like the idea of the camera mangling the data before I can get my hands on it, even if is an improvement 99.9% of the time. Apparently accuracy is not valued.
Yes, the only other exception to this that I see is having the camera do dark noise compensation. Dark noise is not random but rather the tendency of different pixels to have a different value for black in a long exposure. But, you should be able to turn this off if you don't want it. -- JRT _______________________________________________ Open-hardware mailing list Open-hardware@duskglow.com http://lists.duskglow.com/mailman/listinfo/open-hardware
