When I first got my istD I was very worried about the 3-5 hot pixels, some red, some purple, one blue, that tended to turn up in longish exposures. And the red pixel that showed up in 80% of my exposures at the same position. Well, after about a month, I completely got over the hot-pixel phenomenon. It's nothing to be worried about, and it's nothing that can be done away with. You know, I did the whole 15 second exposure deal with the lens cap on in an unlit room, and yes there are hot pixels that show up in quantities similar to what you posted... Thats also at 15 seconds and with no subject matter, so I can only assume that it doesn't really apply to the real world, unless taking pictures of black emptiness happens to be your game...
Dust causes me more problems than hot pixels do, in fact, hot pixels are the best kind of flaw, they're in the same place every time, so I know where to look by now, and the color signature is usually very easy to pick out. And if it's not very easy to pick out, well, then it must not be very easy to pick out, so why worry?? -el gringo -----Original Message----- From: Brian Dipert [mailto:[EMAIL PROTECTED] Sent: Sunday, July 11, 2004 7:04 PM To: [EMAIL PROTECTED] Subject: More *ist D Hot Pixel Oddness I reran my testing this morning on two different *ist D bodies, with NR both on and off, and across 21 shutter speeds ranging from 1/90th of a second to 10 seconds (yes, 84 TIFFs). My results on the original body were much better this time, which I can't quite explain. The only difference this time around was that I made sure I covered the eyepiece; all other settings were exactly the same as before. Look for example again at the old 1/20th of a second data: [DeadPixelText] Version=1.0 Description= FileType=TIFF NumBadPixels=15 0=Hot,286,340,87 1=Hot,287,340,64 2=Hot,2568,453,65 3=Hot,2567,454,65 4=Hot,2568,454,112 5=Hot,2569,454,65 6=Hot,2568,455,66 7=Hot,2162,1117,66 8=Hot,1666,1213,93 9=Hot,264,1391,91 10=Hot,833,1691,65 11=Hot,832,1692,65 12=Hot,833,1692,132 13=Hot,834,1692,64 14=Hot,833,1693,64 Now here's today's data on the same body, exactly the same settings, but again with the eyepiece covered: [DeadPixelText] Version=1.0 Description=1/20th, old camera, NR on FileType=TIFF NumBadPixels=3 0=Hot,286,340,64 1=Hot,2568,454,67 2=Hot,833,1692,92 Note that the three pixels referenced in the new test are replicated in the old one....the additional pixels in the old test are clustered around them. This mimics what I saw when looking at the old TIFF...almost looked like the pixel matrix 'error smearing' effect of JPEG compression (but I wasn't shooting JPEG). Is light leakage from the viewfinder sufficient to explain this? Other suggestions welcomed! I will toss out that I think there's some randomness in this process; NR doesn't come into play until 1/4th of a second and slower shutter speeds, so above that speed results with NR on and off on a given body with otherwise similar settings should be identical, but they aren't. Good old analog CCD technology...... p.s...the new body, surprisingly, seems to be worse than the old one. Here's its data at 1/20th of a second, again with the eyepiece covered: [DeadPixelText] Version=1.0 Description=1/20th, new camera, NR on FileType=TIFF NumBadPixels=8 0=Hot,1001,796,100 1=Hot,1918,1192,69 2=Hot,289,1360,62 3=Hot,959,1498,100 4=Hot,958,1499,92 5=Hot,959,1499,123 6=Hot,960,1499,92 7=Hot,959,1500,100 Notice the clustering effect again of hot pixels 3-7. How to explain the old vs new body results: more randomness? Or is Pentax loosening its allowable pixel defect rates in order to get costs down? ============================== Brian Dipert Technical Editor: Mass Storage, Memory, Multimedia, PC Core Logic and Peripherals, and Programmable Logic EDN Magazine: http://www.edn.com 5000 V Street Sacramento, CA 95817 (916) 454-5242 (voice), (617) 558-4470 (fax) mailto:[EMAIL PROTECTED] Visit me at http://www.bdipert.com
