I've just finished smoking this thread starting from page 70. First of all I would like to thank Danne, dfort and many others who helped testing releases and iron out the bugs. It is absolutely incredible what you guys have achieved here. I used to be a Sony guy with 3 Sony bodies and only one Canon (6D), but because of ML and mainly the video raw capabilities that have opened up to us recently I now have 5 Canon bodies (withing a month I purchased a 5Dm3 and 3 x EOS M). Raw video is so addictive.
Now, while I've been reading the thread and came across many 'bug' reports, I felt that possibly?? some/many of them could be related to the cards people are using (which is possibly the reason why they couldn't be reproduced by Danne). For raw filming I personally use Sandisk Extreme Pro cards exclusively and have come across very few glitches with ML on my 6D or EOS M and they have been the same - I routinely get one or two corrupted frames at the very beginning of my takes with the rest of the footage being fine. Since we 'overclock' the interface, making it write faster than Canon wanted us to, maybe it's crucial to use the highest quality cards.
I only have limited experience with EOS M - so far I've been using the old Nightly.2018Aug07.EOSM202_mv1080p build because it gave me a quick access to full sensor 1736x976 mode which I used for filming two music clips last weekend and a few test takes in the last two weeks. I wanted to read the thread and get up to speed with all the new development that took place after July last year before installing the latest build, which is what I'm going to do now and dive into all the new stuff.
I hope to be useful to the community that has given me so much and give back something. Would there be any interest if I shared my experience with different raw processing workflows?
baris3 - if all you want is trouble free 4K footage and don't care/don't understand/don't need RAW and just want higher resolution videos - I would suggest you stick with M50 and 4K video it can produce. RAW video, especially using ML is much more involved, especially if you want to push it to extract every bit of detail/color information out of it. Do you shoot raw photos or jpgs? If you shoot raw and understand what advantages it gives then you can understand the advantages of raw video - it's the same. However if you take jpg pictures and are happy with them, raw post processing workflow will be overwhelming for you. I can record only about 27 minutes of 1736x976 14bit raw onto my 64Gb card (are you ready for HUGE files?) If you increase these dimensions by a factor of two (which is close to what you will get when approaching 4K anamorphic), that's 4 times more information, so a 64Gb card would be able to hold only about 7min of recording (a bit more if you record 10bit). Do you want/need that?