Very interesting. I think i'm gonna keep my camcorder as long as this rolling shutter won't be solved on DSLR. Thanks for this test.

typo alert "please check my block for how this was shot "

I guess that should be 'please check my blog for how this was shot'. Posting a link to your blog would be a good idea right now ;-)

More people should do these comparisons, maybe then we can see the true range of these camera's!!! Worth a 'like' in my book ;-)

D90 = blueh, 5DM2 = nice sharp detail, GH1 and 7D = absolute identical picture, bad details.

Good job, Mike. Thanks for sharing.

Thanks Mike!

How would rate the 5D versus 7D in terms of sharpness?
Is the preset sharpness lower by default on the 7D?
Can both camera deliver equal sharpness?
Got a 7D on preorder but that seems to be by biggest concern about the camera.

Thanks for the test.

Was the 7d shooting 30p?

I wonder if the rolling shutter performance is any different between 24p and 30p modes on the 7d.

I don't think anything useful can be easily gleaned from this test. It wasn't controlled well enough. Further, I don't believe all the cameras were shot at 1/80 second.

I was going to gather stats of lean vs. motion blur, which would have given a useful rolling shutter comparison. However, there is virtually no motion blur in the Nikon D90 clip. It wasn't shot at 1/80 second but a much shorter shutter speed.

An alternative would be to compare lean vs. motion between frames. The problem with that is that the video is 24 fps, and not all the cameras shoot at 24 fps (5D, for example), so consecutive frames in this assembly of clips does not represent consecutive frames in the original.

I appreciate you taking the time to do these tests.

I've seen quite a few "rolling shutter" tests and they all involve wagging the camera back and forth on a tripod at arbitrary speeds/rates. IMHO the most conclusive test that can be done is to get in a car (have someone else do the driving) drive the same stretch of highway at the same speed for each camera. the sign posts would easily exhibit the varying degrees of rolling shutter and could easily be measurable.

unfortunately, this really isn't a very good test, and only proves that all vdslr's suffer the same terrible rolling shutter problems when you shake them vigorously. None looked perceptively better than the other.

I think I'll be making my next short film, titled "Mad Wagging on a Tripod," with a ccd video camera.

Wow i think I just threw up

hi mike i asked u a question which do u like better the d90 or lumix gh1 and u said lumix, ( i do love ur video times square a nyc tourist moment), now i ask u your prefrence between the lumix gh1 and the canon eos 7d, by the way thanks for your comment on my music video dreaming

I like your test. But you're next test should consist of walking around while recording, and a few quick pan here and there to mimic daily use. Because that's what most people do, So it would be the most natural way to test rolling shutter because most people try to use it like a camcorder

Digging out my old ASC manual, it would be interesting to actually time the pan motion of a static scene--say a 90 degree sweep at various frame rates with various lens focal lengths. Say shooting at 24 fps with a 50mm lens the camera would take 23 seconds to pan 90 degrees and 80 seconds for a 150mm lens. (35mm film). I also noticed a stuttering of the image during slow er pans. I wasn't sure if that was because of the camera or the delivery system. Thanks for conducting the tests. They were insightful. Leaning buildings in N.Y. equal leaning palm trees in LV.

how come the colors of the 5d are so warm?

thanks for the test. was a lot utility