It does not make much sense to build a destructive overlay on a 16 bit image. The overlay hides parts of the original image. Hence it is only of limited use for further processing. I think the main reason for the overlay are documentary purposes. RGB is completely sufficient if you print or view such a documentary image. There is perhaps an exception for medical images which will be displayed on 12 bit displaying monitors.
I tried Adam_H's solution, but I am dealing with U16 images. If I typecast from RGB to U16, I get a black image with the text. What steps do I need to add?
Have you found a valid solution? I'm also troubled by the U16-problem. I use 2 cameras with 10 and 12 bit. The cast from U16 to RGB32 seems to be broken (and does not work as single function either).
