Hey everyone
I just came across the FFMPEG software (https://www.ffmpeg.org/); which I think some of you may already know about and I began wondering if there might be a way to make Magic Lantern take full advantage of its incredible features, especially the compression in one way or another. Besides it is even a great software for multi-platform streaming (talk about having your own broadcasting).
What do you think?
There's a thread in the raw recording section about ffmpeg decoding of MLV files.
As for encoding, no. Plenty of threads on this subject.
There are already several ML projects (post processing utilities) that utilize ffmpeg:
MLRawViewer (http://www.magiclantern.fm/forum/index.php?topic=9560.0)
cr2hdr-r (http://www.magiclantern.fm/forum/index.php?topic=13512.0)
(there may be others)
As Audionut says, using it in camera is more or less impossible.
I like working with ffmpeg and ProRes encoding. However, the 10-bit conversion loss even as high quality as Prores444 is rather significant compared with staying with a dng workflow (DaVinci Resolve, Premiere Pro CC). Exposing ettr style will keep good clean shadows but dynamic range is never gonna be as good as with native dng 14/16 bit. There are some really nice features like 3d lut implementation and a rather good sharpening effect(for H.264 mainly) so if not working in extreme conditions ffmpeg encodings could do just fine.
The best Prores exports I get is when going dng to Prores through AE but that workflow has a lot more to wish for if working larger amounts of files.