Vertical Lines in RAW video, ML August (938047c)

Started by mikeinmotion, November 28, 2013, 02:03:25 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.


mikeinmotion

I tested new v2.3.NEXT.2013Nov28.5D3113 and this is a result:

http://www.flickr.com/photos/110050526@N07/11100881773/

Less lines but they are still there!

eyeland

I have the same problem and I have not been able to find much information about it on the forum :(
Daybreak broke me loose and brought me back...

arrinkiiii


This happen also with me, even with low iso but in low light conditions. Noise pattern?

a1ex

Follow this suggestion: http://www.magiclantern.fm/forum/index.php?topic=9454.msg90754#msg90754

and create a thread specific to banding and noise issues, collect samples from all cameras so we can look at them, don't just fill the threads with "me too".

dmilligan

Quote from: arrinkiiii on December 12, 2013, 08:12:33 AM
This happen also with me, even with low iso but in low light conditions. Noise pattern?
http://en.wikipedia.org/wiki/Fixed-pattern_noise

Canon sensors (actually CMOS in general, but Canon more than others) are notorious for having these kinds of issues, but in most normal situations they are corrected for automatically by the camera, and so most 'normal' users never notice them or know anything about them (just as Canon would like). ML RAW, however, captures the RAW data before these corrections are applied, so you're seeing the sensor in it's full, imperfect glory.

Astrophotographers, are very familiar with these issues b/c we see them all the time, due to our exteremly long exposure times and very low SNRs that we have to deal with. In PixInsight (the 'photoshop' of astro image processing software), there's actually a tool called "CanonVerticalBandingRemoval" or something like that.

The simple solution is dark frame subtraction. MLV supports storing a dark frame, but AFAIK there's not actually an automatic method to capture one (yet), and mlv2dng doesn't make use of it (??). One would need to figure out how to make the shutter curtain close while recording.

a1ex, a nice potential feature of raw2dng would be that you could specify a .RAW file for dark frame subtraction. One could shoot a short sequence with the lens cap on, with all the same settings, then specify that .RAW to raw2dng and it would use that sequence (stack all of the frames to create a 'master dark') to correct your other sequences. Temperature and ISO certainly affect these issues, so there might have to be more to it if either of these things are changing. I could look into this myself if you like.

arrinkiiii


Thanks a1ex and dmilligan.

I will open a new thread with some examples later this night. Got in my pc that are render one cinema 4d sequence (more then 12h).




eyeland

Quote from: a1ex on December 12, 2013, 08:16:16 AM
Follow this suggestion: http://www.magiclantern.fm/forum/index.php?topic=9454.msg90754#msg90754
and create a thread specific to banding and noise issues, collect samples from all cameras so we can look at them, don't just fill the threads with "me too".
I started this thread in august: http://www.magiclantern.fm/forum/index.php?topic=7702.msg67836#msg67836 in which I linked to a video I made of the vertical banding problem. I requested that the thread be merged with this thread: http://www.magiclantern.fm/forum/index.php?topic=5614.0 that also seems to deal with the issue. Threads were not merged and the latter is marked as "SOLVED".
I don't want to be pushy so I kind of gave up and the vertical banding issue and the lack of EXIF data made RAW video so cumbersome for me that I went back to shooting h264 once the initial fascination had worn off (fully understanding that one needs time and patience for a non-profit project like ML)
If the time is ripe for dealing with the issue now, just tell me what I can do to help, or where to post what :)
Daybreak broke me loose and brought me back...

a1ex

The nature of banding is different here; the solved issue was for banding in bright areas and for autofocus pixels; yours is underexposed. A dark frame helps here (dmilligan mentioned it somewhere else), but I think it can be solved without dark frame too.

@moderators: can we merge all these threads? (yes, somebody should search for them :D )

I want a single place where I can download sample footage showing all these problems, on different cameras. The solution is a post-processing algorithm in raw2dng / mlv_dump .

RenatoPhoto

http://www.pululahuahostal.com  |  EF 300 f/4, EF 100-400 L, EF 180 L, EF-S 10-22, Samyang 14mm, Sigma 28mm EX DG, Sigma 8mm 1:3.5 EX DG, EF 50mm 1:1.8 II, EF 1.4X II, Kenko C-AF 2X