I like Live View for critical focusing when using my 1DsMk3 and 5DMk2 on a tripod. I also have used it for framing when I have the tripod elevated for a better camera angle - I am somewhat vertically challenged.
Like others, I have some concern that the energized sensor for Live View will create enough heat to add noise to an image. I decided to try a little test just to see if there was a way to validate this or if indeed Live View doesn't add any significant noise. I think to really determine it full scale would require a lot more tests, varying the ambient temperature, exposure times, ISO etc. But I decided to do a basic test to get an idea. Anyone have any feedback on my method (am I out in the night on this) or what I've found let me know.
Setup was pretty simple. 5DMark2, inside at room temperature. Body cap on camera, viewfinder covered. A series of exposures, each 1 second long at IS0 100. First exposure is before live view is enabled, then with Live View on, 1 exposure each minute for 5 minutes.
The images were brought into LR, all adjusted to 0 black, +4Ev to exaggerate noise. Then opened as one stacked document in Photoshop, where I did a screen grab of the histogram for each layer. I then cropped to a 250x250pixel image from the middle and have show that at 100%.
Here is what I ended up with ... it appears even after 5 minutes there isn't any significant difference.
[attachment=11673:LiveView...cCompare.jpg]