On Tue, Nov 22, 2011 at 5:33 PM, AlunFoto <[email protected]> wrote:
> Revisiting my recently posted Runde GESO, I found that the shots with > longer exposure times showed some very bright pixels in areas where > there shouldn't be any. And they weren't just single pixels either. > Sometimes streaks, sometimes little doodles looking like luminiscent > dust. After two days of research, as in Google (re)search, I think > I've nailed it down to cosmic rays hitting the sensor. I feel > confident enough about this explanation to have blogged about it, but > I am a bit skeptical still. I'd therefore appreciate it very much if > any of you guys have a, literally speaking, down to earth explanation. > > I must refer you to my blog once again, at least for image examples: > http://alunfoto.blogspot.com/2011/11/spurious-noise.html I think that's a very plausible explanation. I used to be an astronomer and have considerable experience observing and processing data from Palomar, Keck, and other major observatories. Cosmic ray hits are routinely seen in the professional CCD instruments--they are likely to occur by the dozens in typical exposures of several minutes, and dealing with them is a routine part of processing. Are these with the 645D? The bigger the sensor, the more cosmic rays will pass through it in a certain period of time, which is why we see so many more on the large observatory sensors than in small-format cameras. The appearance of your streaks is quite reasonably similar to those I'm familiar with, although it was less common (but not unknown) to see them extend over so many pixels in the observatory sensors. I suspect that's because the astronomical CCDs have physically larger pixels, which can also store more charge before spilling over into adjacent pixels. Demosiacing is also likely to affect and possible extend the appearance of the cosmic rays, which was not an issue in the monochromatic observatory sensors. Because comic rays interact with the sensor directly, not the optics, they can be sharper than objects in the scene. Usually in astronomy our sensors were highly oversampled--a point source (like a star) would be blurred over a large number of sensor pixels. Because the CR hits were so much smaller, and rose so much more sharply from the background, they could be identified in a single image, without having to combine multiple images. (We *did* usually combine multiple images, but for other reasons than CR detection.) -- PDML Pentax-Discuss Mail List [email protected] http://pdml.net/mailman/listinfo/pdml_pdml.net to UNSUBSCRIBE from the PDML, please visit the link directly above and follow the directions.
