Incompatible image geometry - Another issue

[MikeOates]

I foolishly made a set of flats at too high an exposure and this means i can't use them to calibrate a batch of images.

So I thought I would adjust the DSLR CR2 images in Lightroom (reduced exposure by 2 stops) and export them as TIFFs.

But when I try and use them in calibration I get the "Incompatible image geometry" error. I even converted the TIFFs to FITS and tried those, but that did not work.

Is there a workround for this, I can't understand why I get that error when the sizes have not changed.

Update: I just checked the image sizes and they have changed from 4770 x 3178 to 4752 x 3168. That is looking from within PI using the File Explorer, but when in Lightroom the size for both CR2 & TIFF files are 4752 x 3168. Does anyone know why this is happening? The camera is a Canon 500D

Thanks,

Mike

[pfile]

well... the images are being debayered by Adobe Camera Raw when you do this in lightroom. DCRAW is probably interpreting the data slightly differently. that would account for the different pixel dimensions.

but you can't properly calibrate debayered DSLR images so i'm not sure what you are trying to do will work. the closest you can come would be to debayer using superpixel mode in PI, save the files as tiff, then put those images into lightroom and fix them and write them back out.

but i still question the need for lightroom; you can do pixelmath on the flats to try to reduce their intensity, even in bayer mode. regardless i'm not convinced it would work. if the flats are truly overexposed then you can't recover data that's not there.

what happens if you load these flat subs into PI and look at the histogram?

finally if you are trying to use the debayered flats coming out of LR against your raw CR2s, that will never work...

[MikeOates]

Thank you very much pfile, You have given me a few things to think about and try. This is what I needed some good advice, I have no idea why I did not think of using pixel math, that should work.

The flats are not overexposed in that all the data is there, but just too bright so it would not work correctly.

I am pretty new to PI and learning all the time.

Many thanks,

Mike

[pfile]

are you sure that they will not work? i think that the first step that PI does when preparing to divide the flat is to normalize it, so as long as a flat is well exposed it should not matter what the absolute pixel values are.

rob