Originally posted by acoufap DNG is only a standardized container format where the camera puts in the sensor data and it‘s icc camera profile and some additional data.
Not sure what any of that has to do with what I wrote.
Originally posted by acoufap So the Adobe DNG converter simply won’t cut out a square or any other rectangle from the sensor data.
Well, it does omit the masked pixel data from original PEF files, for instance.
(Most people don't need that masked pixel data anyhow, but it part of the original data that is filtered out by the Adobe DNG converter.)
If that kind of data reduction occurs, I think it is plausible that only visible pixels (those inside the 1:1 mask) could be actually stored in the DNG file (still not demosaiced, of course).
Originally posted by acoufap But you can do your own experiments and prove me to be wrong.
I won't be doing any experiment myself because I shot about three square images with my K-1, mainly for testing purposes.
I just thought I mention the potential for someone who shoots a lot of 1:1 images and is keen on saving data to check whether the Adobe DNG converter may do the trick for them. TBH, I doubt the converter would go through the trouble of filtering the data but it is not entirely inconceivable that it does.