I was able to do one more test last night with the help of my friend Michael from Santiago Filmworks. We tried to keep things a little more consistent. I am still not sure if it is accurate because the final video is on a 29.97 timeline but I also have 30p footage from the 5D mkII mixed with 29.97 7D footage. It seems to play fine but I am sure it would cause some drift issues when syncing footage on a multi-cam shoot. I seem to remember hearing that NeoScene converts the 5D mkII footage to 29.97 when converting to the Cineform codec? I need to look into that.
Anyway, we shot Caltrain in Sunnyvale to try and gage the rolling shutter comparison of the two cameras. Unfortunately I didn't have two tripods, so we set both cameras on top of a box side by side and there is a little shake from the trains. Both cameras have 50mm lenses, the 5D mkII has a 1.8 Nikon and the 7D has a 1.8 Canon. I would agree with some who are saying the 7D is about a stop brighter and on the first shot the f8 label for the 5D mkII isn't accurate, the f-stop matches the 7D at 6.3. The comparison video shows side by side and then each clip individually.
All in all, both cameras have really pleasing images. The 7D isn't as contrasty and doesn't pick up as much detail as the 5D mkII. I do like the filmic quality of the 7D and some of the new features of this camera. If the 5D mkII had 24p, it wins hands down but both are very solid and have there own sets of strengths and weaknesses. I am a little concerned about mixing the footage with the varying frame rates and need to figure out a strategy to keep them in sync in a single timeline.