Looks like the 5D mk II isn’t immune to the skew produced by a rolling shutter, although it performs better than the D90 for sure — and I’m sure a careful shooter can produce some very impressive footage with either one. Although you mainly notice it in the video above when abnormal shaking movements are made, that kind of issue can really mess with the natural look of a medium-speed pan — say, tracking a car passing the camera.
Another problem occurs in flickery light, for instance some fluorescents, which may be lighter or darker in certain parts of the frame due to interference between the light’s “frequency” and the camera’s electronic shutter speed. Oh well, the camera’s still sweet as hell. [Thanks for the tip, Mike, and tell me if I'm getting it wrong]
I might be missing something, but isn’t this just a case of the autofocus system’s inability to keep up with video (probably because with video, a different and less effective AF method must be used, compared to still shots).
Would switching to manual focus resolve the issue?
It actually was on manual focus (says the guy) but that’s not what I wanted to point out. when the camera shakes or moves at any amount of speed, the image “leans” away from where it’s moving due to the way the camera pulls information off the sensor. I’ll add a link to the d90 video, it shows it way more.
id be interested to see what an XL1 or XL1H looks like with similar shake
jas
As a long time Canon shooter I can tell you that this mysterious “Jello” effect can be simply and effectively eliminated. What you are seeing is the Image Stabilizer in the lens attempting to stabilize the shot. Simply turn OFF the stabilizer and bingo no more “Jello” effect.
Regarding Grant Brummett’s assertion that the “jello” effect is caused by the Image Stabilization function.
While the IS function can be problematic on quick pans or rises it is not the cause of jello distortion. IS can cause a pause at the start of a pan followed by a jump in the fottage to catch up with the pan (or any other camera motion). Some of this effect may be present in the sample clip, but the shaking of the camera is so violent that I can not be sure. The jello like twisting or bending of the image is the result of the CMOS sensor which is sampled sequentially from one edge to the other (top to bottom I think but it could just as well be bottom to top or side to side) This results in one edge of the picture being exposed at a different time than the far edge. They call it a rolling shutter and it functions much the same as a focal plane shutter on a still camera. If you have seen early 20th century photo’s of cars racing, you may have seen the apparent leaning of the car in the direction of travel. The top of the radiator leading the bottom of the radiator. These photo’s were shot on large format film with fairly slow focal plane shutters and this distortion is analagous to the jello effect seen here.
Hey Fred. The exact thing is happening to me. When I hold the camera straight, the footage comes out great, but when I do a large tilt or pan, the footage pauses and jumps. It’s really bad, and unusable when it jumps. I tried turning off the IS on my f4 24 to 105 lens, but that doesn’t fix the problem. Any ideas? Fred, I was looking for a way to contact you to ask you personally, but couldn’t find your contact info. Please email me with your answer to djp76 (at) me.com. I would really appreciate it.
Pauses and jumps? That does not sound like the jello effect. That sounds like the playback device not being able to keep up with the video (much more info must be processed during rapid whole-frame picture changes, like when a camera pans).
If the “jump” is in the pan itself, then it sounds like lens IS, but if you had that swicthed off, I suspect your playback processing is falling behind during decompression/presentation.
The 5D Jello is much less of an issue than the D90 Jello — with the D90, I kept having it ruin clips of medium speed pans (in fact, every shot in this mode had VERY exagerated Jello, but in the 5D using the same rig-motion, I could only detect it if I was trying to see it. I had to pan much faster to make it really noticeable on the 5D, to the point where the shot was borderline unusable because of the pan speed anyway.
Fred,
Thanks for the informative input. Perhaps the camera was shaken pretty hard from side to side. Maybe harder side to side then anyone trying to actually record a watch-able video?
I know I have seen this effect through my 40D but it wasn’t much of a factor because I was trying to track a flying bird and each shot is a still.
Grant:
I couldn’t agree more. The camera was shaken more violently than any reasonable person would do trying to make a watchable video.
I’ve seen the jello distortion most clearly on video taken from a helicopter, with the camera mounted firmly to the air frame. The vibration caused the whole picture to be full of ripples, still nice and sharp, but rippled like waves on a lake. It also tends to show up if ridgidly mounted to a moving automobile, especially a racing car. Canon’s HV20 and HV30 video cams also use CMOS sensors with rolling shutters.
I’m sorry guys, this conversation has gotten a little too informative and civil, I’m going to have to ask you to leave.
Hmm, wtf do u mean by “too informative and civil”?
Hello! generic cymbalta
I am interested in the new 5D Mark II and these posts have shed some light on a subject I’ve only just heard of.
I was thinking of using this camera from a helicopter but curious about the mount that was used via Fred’s post? The Tyler Nose mount (No Gyro’s) might have caused this to shake… but most other mounts are aptly gyro’ed or stabilized to limit those guaranteed helicopter vibrations.
Any info on this particular subject would be great.
Cheers
Yeah that “Jello” effect ruined 2 important shots.
Make shure youre dolly is absolutely flat.
horrible jello vision. Not as bad as the d90? Jello is jello and it is annoying to see. You can still notice it pretty badly even when camera just shakes a little bit. Even my $250 cam corder performs better and shakes less. Sure it doesn’t have the lens advantage but I won’t have to haul a rig around just to take some video. I guess neither canon nor nikon was able to bring practical DSLR video. Great chance for sony IMO to step up to the plate and produce jello-free video.
I’ll pass thanks.