Hey, DavidP, was absent from computer for a short while. Excellent film you made, more than excellent use of existing light, your cam position is not always optimal, but nevertheless you suck an enormous amount of action out of it. Wow, one of the better films to be brought to this forum. You certainly set the bar high for what I hope to achieve. So I understand, that pinks are totally annoying.
And... DeafEyeJedi may be on to something.
Card speed is best determined by benchmark in-cam. Go to debug menu, remember to benchmark in 'play' mode, test your card. The "official" markings are read speeds, not write speeds - and by the way I don't know if pinks have anything to do with card speed, but nobody seem to count it out. Pinks happens so often, that there are several threads in this universe dealing with them.
But then again... Cam makes pinks, recovers and continues as if nothing happened. There are a lot of variables; which one sets off the pinks? or which combination? A1ex and many other of the most insightful guys have tried to recreate the precise conditions for pinks, not only with 5d3 but with several cams, and to the best of my knowledge haven't found a common denominator yet.
I'm rather green in this environment and I'm still experimenting with my (new) 5d, and the settings I mentioned above have worked so far. but I also shut off all irrelevant modules and features (and probably more than necessary), so I don't know if there is a culprit hidden somewhere.
Base line:
Go ahead. Use ML for pro jobs - you're doing a helluva good job with it already, color and effects are yummy. If there is a pink spark somewhere, let it be. Check footage every once in a while if at all possible - and have a good buddy with a cam2 running for really important work. The chance that cam2 has a pink at the exact same spot is nil. So splicing cam2 in is just like having... er? a cam2!!!!!
