Thank you reyalp for your prompt reply.
I added both version of DNG files in case it is needed there
Thanks. Looking at the 1.3 DNG, there are in fact no 0 valued pixels, so from a CHDK POV, everything appears to be working correctly. This doesn't mean your sensor has no bad pixels (unlikely, as you say) but the kind of pixel badpixel.bin would deal with is not present in this image, so the 0 bad pixels is correct.
There are ~350 pixels with values below the nominal black level of 127, but it doesn't look to me like the camera is using a different value to mark "bad". There are none with a value less than 81, and the number increases with higher values, which is a pattern seen on other cameras. I wouldn't expect these to affect image quality much, but you could patch them with a
manual badpixel list if you wanted.
I notice the exif says this was shot at ISO 400. It's possible that this camera will exhibit different behavior at other ISO values, for example at high ISO the cameras own processing might be doing something with the bad pixels. I would try at the lowest ISO (50 according to the specs, which is a bit unusual)
Examining the image in raw therapee, I only see a relatively small number of hot pixels, which CHDK badpixel.bin would not address anyway. Again, these could be addressed with manual badpixel removal.
Converting with Adobe's DNG validate gives extremely ugly results (speckles of saturated green and purple) but I suspect this is due to the IR filter removal. You can find some previous discussion of DNG with IR modded cameras in this threads
http://chdk.setepontos.com/index.php?topic=11475.0http://chdk.setepontos.com/index.php?topic=10520.0I seem to remember there were some more specifically discussing color but I didn't find them.
