CMOS Darks and Flat Darks: Part 2 --Pixel Math Trouble

Hi Adam,

I was viewing the above and experimented on my own flats and this is the problem I ran into:

My MasterFlat using the Statistics Process showed a Mean= 19371.  After applying this Pixel Math expression "$T - .15" the Mean now=34329. Why does not subtraction take place?

If I turn off "rescale result" which produces the Mean value I was looking (14335) when used as the replacement Master Flat it fails to calibrate in BPP by generating this error:  *** Error: Zero or insignificant scaling factor for flat frame CFA component 0. (empty or marginal master flat frame data?)

If I enable "rescale result" the output is far higher than expected. My goal was to obtain a Mean value in the 14000 range.

Thanks,
Bill

Comments

  • Perhaps a little thinking about concerning what subtraction is doing...
    Let me give you an example...(I cannot see your image... so this is just a thought experiment).

    Lets say your flat has some strong vignetting towards the corners. Those values where it is vignetted might be around 0.15. So if you subtract 0.15 and make them zero...does this make the average rise or fall. Remember... with these images you never get a negative value. In addition, (literally) I think the average is of non-zero values. 

    I think the issue are the zeros you are creating when you subtract a large value like 0.15 from your data.

    Does my thought experiment gives you pause?

    -the Blockead
  • Indeed, I decided to go back to Fundamentals --Pixel Math videos as I have not given them the time they deserve.
  • I found your troubleshooting the "moire" pattern in the problem image a useful tool from pixel math.

Sign In or Register to comment.