Well, I'm not sure what the dots are. They could be part of the wall pattern, I suppose. Do you have something more uniform and gray you could try a test DNG with?
In other words, I have a stable script by calling shot_meter_enable(1,1,1,99,99).Of course on the S95 my script runs OK without calling shot_meter_enable(1,1,1,99,99).
FWIW, the colored dots you see in the G11.DNG are definitely bad pixels, known to the canon firmware as bad. Canon sets these to value 0. The reason you see colored spots is that only one R, G or B element is 0, so you get the other two colors when it's debayered.
edit:Also, the black line at the top indicates the active area is not set correctly.
Does the Canon firmware set the known bad pixels to 0 before build_shot_histogram() is called? If I happened to sample one, this means it would show up at position 0 in the histogram, well below the black level, so it's easy to spot.
I 'solved' it by not using aflock and explicitly calling x=get_focus at the beginning of the script and just before I capture an image I call set_focus(x)
if(get_focus_mode()~=1)then press("shoot_half") repeat sleep(10) until get_shooting() set_aflock(1) release("shoot_half") while(get_shooting())do sleep(10) endend
press("shoot_half") -- focuses once and holdsrepeat sleep(10) until get_shooting()-- the above may not be needed.-- Press("shoot_full_only") should focus the first time-- get_shot_ready() only goes true AFTER get_shooting() is true--shoot loop press("shoot_full_only") repeat sleep(10) until get_shot_ready(0) -- waits for histogram to be valid release("shoot_full_only") -- still holds shoot_half down --calculate your new tv96 value from the histogram here set_tv96_direct(tv96) -- works in half shoot immediately now--repeat shoot loop
Started by n6mod
Started by Graystar
General Help and Assistance on using CHDK stable releases
Started by mrblack51
Started by cantain
Started by Bernd R
General Discussion and Assistance