SYNTHESISING A 12bit IMAGE FROM 8bit EXPOSURES
AN IMPROVED AVERAGING TECHNIQUE
The averaging method used in the previous page can be improved to give greater accuracy at low levels and bias the final image towards the frames with the longest unsaturated exposures at each brightness level. The formula used at each pixel is
16* (sum of values at 1,2,4,8,16 unit exposures /sum of exposure times)
(NB individual exposures where value exceeds 255 for that pixel, are ignored)
The revised graphs are shown below
