-
Øyvind "pippin" Kolås authored
All other precisions are still stored as 32bit float - perhaps we should store 8bit integer files as half as well? The only case where this can lead to slight fidelty loss is 16bit integer. Fixing gimp#10778.
All other precisions are still stored as 32bit float - perhaps we should store 8bit integer files as half as well? The only case where this can lead to slight fidelty loss is 16bit integer. Fixing gimp#10778.