I'm not expert of imaging tech, but I can observe the video
preview is wrong and useless at night, when my eyes are
able to see targets well. (E.g. a fox chasing a cat which
I missed because black preview is useless for fast targeting.
Canon failed badly!)
In computer graphics they have Cornell's box: side-by-side
a real box with objects inside and a computer display.
They make the display and the CG rendering to match
the real box. That kind of match I would call wysiwyg
view finder when done in camera.
The essential question is: can I do it with CHDK or not?
At CHDK development level, I mean.
Can I add pixel operations, such as brightness, prior
view finder or encoder for video?