Magenta edge tinging at high ISO and a possible solution...

Started Aug 17, 2009 | Discussions thread
GordonBGood Veteran Member • Posts: 6,308
Re: What about K10d raw files?

KL Matt wrote:

I still have my good old k10d and it exhibits very similar behavior. Two differences in what I've seen so far are that it is the left and right sides that are more affected than the top and bottom of the frame in landscape mode and that the magenta cast transitions into a green cast on the opposite side. Perhaps that is simply the result of some WB setting I have made, I'll have to do more tests to see. How feasible would it be to adapt your code to work on k10d files as well, and how much more work would that cost you?

Well, the problem may be related. With the K10D, it is due to the red and blue photosites being while the green photosites are simultaneously scanned by the other readout channel from the other, with the offset levels for all increasing with dark current over the time of the scan. These makes the red + blue = magenta slightly stronger on one side and green slightly stronger on the other, all in landscape orientation. The K20D scans differently and I don't know if I am correct that it scans from the top and bottom to the center or not, but the problem may be more related to the heat of the various areas of the sensor over the scan.

I recall the issue being quite pronounced with my camera while i was living in Asia. I wonder if the heat had anything to do with it. Let me know if you would like me to send you any files.

Yes, because the rate at which the offset drifts is (exponentially) proportional to sensor temperature, the problem will be worse for the K10D for warmer environments. I'm not sure this applies to the K20D or not, as it this effect may just be caused by differential temperatures across the sensor, which may be more or less constant no matter what the absolute sensor temperature is.

Great work BTW for the k20d files, this and pattern noise are my biggest technical complaints with Pentax image quality. A highly accurate and automated solution to one of them is a real upgrade for me. I may switch to a k20d sometime in the future.

I would be able to fix the left/right colour "shading" for the K10D even easier than I fix the border tingeing of the K20D but with a separate program as the sensor layout and scan patterns are quite a bit different. I don't know how much interest there is in this though.

I'm sorry that I am unlikely to be able to do anything about pattern noise in either the K10D or very little with the K20D. I've tried and failed to "back out" the vertical pattern noise of the K10D, due to there not being enough information left in the noise patterns of the masked black border after being clipped to zero black. The K200D and K20D have much less of this pattern noise but it is still there. Again, I am unlikely to be able to do anything about this pattern noise for the K20D other than I may be able to correct for a particular artifact I call "Wide Pattern Vertical Banding" in the case of the K20D as in godfrog's case presented here.

Regards, GordonBGood

Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow