Yeah, picture style does not get applied to the DNGs when shooting ML RAW, but I do believe such meta-data is recorded and can be applied to the footage in post (so, maybe they did apply picture style in post without realizing it. This could easily occur if not proficient in ML RAW post processing workflow).
But either way, what was with all those purposely over and under exposed shots? Why bother comparing anything but correctly exposed footage???
Also, why include the GH4 and the 1DC? He states it's all ungraded footage, OK, but they don't shoot RAW and apply picture styles in-camera, so in a sense these cameras' footage have been color graded.
Seems to be quite an unscientific excercise in camera comparison in which the authors themselves admit to many mistakes. So nice idea, but poor execution
