have sent out the following email to a dozen registered testers
hi 50D testers,
per a1ex request, we need to confirm the black borders on DNGs generated in Silent Pics mode.
http://www.magiclantern.fm/forum/index.php?topic=5586.msg46436#msg46436i have made a special build to do this - it's not intended for general shooting and most .raw files will have the black borders and crazy colors.
it is to put the camera in Silent Picture mode, take single silent pictures, and get the exact pixel count of any black borders.
this build has set the borders to 0 - hence the name NULL-05June13
if you haven't loaded new builds onto an existing card before, be careful, format it first and delete the ML config file.
not all of us need to do this, so whoever has time tonight should give it a go and report results on the forum.
the values that look correct to me are:
default image size -
top 26
left 74
right 0
bottom 0
in 5x zoom mode
top 0
left 74
right 0
bottom 0
please shoot some Silent DNGs and report back any discrepancies.
it may be helpful to confirm that nothing changes in the other cases:
- try 10x zoom
- move the zoom box to various positions top,left, center, bottom right (didn't JulianH report that thee borders changed when he did a raw video pan?).
- turn on raw_rec and try at various aspect ratios - i'm not even sure this will give images of different resolutions, if it does check for black borders.
the code assumes that these values are always the same, so if they are not there will be black borders.
this build may be slower that 28May13
file spanning will work.
.raw files will all have the black border issue.
weird things may happen with HDMI.
please refer back to forum post and provide credit in any public video.
regards,
mentioned on another thread - have Silent Pics and took 2 DNGs after setting skips to 0 - one default, one zoom mode.
inserted measured skip values into src/raw.c now in unified branch.
tested zoom and regular video no longer have wild black point (fixable in gimp or RPP).
zoom mode test video -
( ufraw-batch -a ) color slightly magenta - roses were a little less pink to my eye.
i see some cameras have different skip values for mv1080 and mv720 - which i don't think apply to 50D.
is there value in testing this further -
zoom at different aspect ratios?
zoom with box top left vs. center (wonder about this because of some comments on color changes when using raw panning)?
please let us know how much additional testing should be done and i will send a zero skip build to some of the experienced testers.