Here were my thoughts on each, as well as some of the blind test results:
The iPhone 4s didn't look as bad as most thought it would. There was a chuckle when it was announced that it was a member of the test. But in the end the iPhone is a $200 device with 1/3" sensor and very small 5 element lens, going up against $80,000 cameras with $50,000 cinema lenses. So given that, it was too bad. In fact I had it as second to worst. The 7d came in last on my list. I thought the iPhone4 was the 7d and the 7d was the iPhone 4s. Go figure.
SONY F65: Looked good, in fact was my number one choice. On further viewings I did see a slight magenta tinge to the post corrected version. This may have been a result of trying to fix the color balance on the skin tones.
RED Epic, Sony F3, Sony FS100, Canon C300 all were in the same neighborhood. Good resolution, decent color. Generally pleasing pictures.
ARRI Alexa: Looked great (My second choice for picture and 1st overall choice, given the fact that it produced the best picture with the least amount of relighting and the least amount of post production tweaking). In the end, the camera that allows you to get a great picture with the least amount of work (and the least amount of expensive color suite rental time!) is the winner for me. Great pictures at the right price point.
GH2 was the real surprise to most everyone there. This $800 camera held it's own with the $80,000 cameras in the area of overall picture quality. Which really opened a few eyes. It was a hacked camera. Didn't catch which hack it was though. That being said, the Raw vs relit and post tweaked images revealed the fact that the GH2 image was really lacking in terms of dynamic range when compared to the other high end cameras. The team that shot the GH2 also bumped the contrast a bit and bumped up the saturation too. Think CSI: Miami. No doubt this was to cover up the stretching of the dynamic range and coverup a bit of noise that was creeping in under the lowered black levels. That being said, it held in there very well. The picture was better than 1/2 as good as the high end at about 1/100th of the price point. $80,000 vs $800. If you controlled the lighting so as not to challenge the dynamic range of the GH2, it would probably be good enough to inter cut with the other cameras. Yes, it was not that far off.
It's not that the C300, F3, Epic, or FS100 looked bad. They didn't. They look very similar to each other. And in fairness the test would have been a different animal if viewed projected at 4k. But then 4k theaters are still a ways away from being standard...