|
|||||||||
|
Thread Tools | Search this Thread |
December 17th, 2008, 12:04 AM | #1 |
Major Player
Join Date: Mar 2006
Location: Chicago
Posts: 706
|
What does "Shutter" mean?
What does an electronic shutter really mean? How can a camera that apparently has a 1/2000th electronic shutter have a slow read/reset? How can there be "jello" with a 1/2000th shutter? Is it technically surprising that the Canon seems to have such a fast electronic shutter?
|
December 18th, 2008, 10:05 PM | #2 |
Regular Crew
Join Date: May 2006
Location: Daegu, Korea
Posts: 180
|
An electronic shutter is essentially the effect of turning the sensor on and off again (I'm not an electrical engineer, so hopefully someone with more knowledge than I will come along and post up a more detailed description).
The Canon 1D has a maximum electronic shutter speed of 1/16000, and the 1D Mark III's fastest speed is 1/8000. One-two-thousandth of a second does not impress me at all. |
December 18th, 2008, 11:54 PM | #3 |
Major Player
Join Date: Feb 2006
Location: Philadelphia
Posts: 795
|
My understanding is that the electronic shutter refers to the amount of time each photosite is allowed to gather light.
On a CCD all sites start gathering light and then are read out at once. On a CMOS each line is read sequentially - and the light gathering period and read period are tied together. So each line may be gathering light for 1/60, 1/125, 1/2000, 1/n of a second, but the time it takes to read all of the lines from top to bottom will always be the same and is determined by the speed at which it can process the data. The amount of skew is determined by how long it takes to read all the lines, not by how long each line gathers light.
__________________
My latest short documentary: "Four Pauls: Bring the Hat Back!" |
December 19th, 2008, 07:40 AM | #4 |
Major Player
Join Date: Mar 2006
Location: Chicago
Posts: 706
|
So doubling the ability to read the sensor may cut the skew in half.
The term read/reset refers to the whole sensor? It seems to me that the number of channels and processor is a major factor in reducing skew. That's good because that's controllable. I'm not bothered by the 5D skew as much as some people here. |
December 19th, 2008, 11:46 AM | #5 |
Inner Circle
Join Date: May 2006
Location: Camas, WA, USA
Posts: 5,513
|
Me either.
But, if I were going to film a ping-pong match, I'd grab a CCD camera.
__________________
Jon Fairhurst |
December 19th, 2008, 01:31 PM | #6 | |
Major Player
Join Date: Feb 2006
Location: Philadelphia
Posts: 795
|
Quote:
You probably can build a CMOS with no skew, but it might only be suitable as a studio camera because the processor would draw too much power to be battery powered, and would then generate too much heat to be passively cooled - especially in a compact camera body. The EX1 has a slot through the body with a fan to keep it cool as it is and it still isn't able to eliminate skew. I would imagine it takes a lot more processing power than the 5DmkII since it has 3 chips, which is probably why the 5D can do what it does in such a small body with no active cooling and smaller batteries. However they've clearly had to push it close to some heat limit and thus had to add the emergency shutdown capability to prevent people from frying the sensor.
__________________
My latest short documentary: "Four Pauls: Bring the Hat Back!" |
|
December 19th, 2008, 01:33 PM | #7 |
Major Player
Join Date: Feb 2006
Location: Philadelphia
Posts: 795
|
Me either, but now I have this sudden desire to go out and shoot a ping pong match with my 5D just to see the skew...
__________________
My latest short documentary: "Four Pauls: Bring the Hat Back!" |
December 19th, 2008, 02:36 PM | #8 |
Inner Circle
Join Date: May 2006
Location: Camas, WA, USA
Posts: 5,513
|
Check out Beijing Hoop Dreams by Dan Chung from about 1:40 for an idea of what ping pong might look like: 5DmkII - Beijing Hoop Dreams on Vimeo
I find that tracking motion in one direction isn't much of a problem. The skew can feel a bit organic. But once the camera reverses direction, it's jello-land. I think it's also worse when telephoto. A short lens ads its own distortions, so it masks skew, but a long lens can really make problems apparent. Just last night, we were trying to get two photos to fit a given frame size. We cropped as much as we could artistically, but neither quite fit. We found that we could simply squeeze our angled, wide-lens building shot, and it looked completely natural. We tried that on the straight telephoto shot, and it didn't work at all. Maybe if we stretched the edges, but left the middle clean, it would work...
__________________
Jon Fairhurst |
December 20th, 2008, 05:10 AM | #9 |
Regular Crew
Join Date: Mar 2008
Location: Paris, France
Posts: 80
|
I guess a slower shutter speed decrease the perception of this effect, it's probably hardly noticeable with 1/60 shutter speed (the same skew happen, but as motion is more blurred with longuer exposure, human eye will hardly see the difference on low frequency informations). So this effect should not be much of a problem for film-like (slow shutter) applications.
|
December 21st, 2008, 10:52 PM | #10 |
Major Player
Join Date: Jan 2007
Location: Portland, OR
Posts: 949
|
It has many meanings.
In the context of the 5D2, I believe the correct meaning is "movable louvers on a pipe organ". The slowness of the pipe organ explains why the camera has skew/wobble, and also the AGC and noisy pre-amp in the onboard sound. |
| ||||||
|
|