When I compare the two videos in my mind the ML video much look better? In fact I see very little difference.
AFAIK H.264 encoding is done by hardware.
ML has no influence on noise reduction, chroma subsampling, sharpening done there.
You *may* see differences in motion blur/artifacts if ML is able to manipulate I-/B-frame ratio.
The video on the ML camera has a *.MOV file much larger then the standard EOS (which is understandable).
Well, I don't understand that one. Maybe dev team is able to shed some light on this.
Ciao
Walter (not using H.264 that much)