Darktable - creating HDR merge

[kmilos]

16-bit half float DNG output of HDRMerge is more than enough to preserve all the data.

16 bit float export looks weird in other applications, but 32 bit float is fine

Yes, it’s less precise than 32-bit float (duh). However, half float is more than adequate for photography - you get approx. 30 stops of dynamic range (good luck representing anything more than 16 stops with uint16) and constant quantization per stop (agreed limited). The quantization error is still smaller than the photon shot noise. As @agriggio points out, there is no need to represent every single integer in highlights because those measurements are noisy anyway (see e.g. why NEF lossy compr…