Tweet
Originally posted by bstolle
View Post
-
Not correct. The name "rolling shutter" comes from the blinder-type mechanical shutter of the reflex cameras (film or digital sensor doesn't matter).
--- Judge what is said by the merits of what is said, not by the credentials of who said it. ---
--- Defend what you say with arguments, not by imposing your credentials --- -
CorrectOriginally posted by Evan View Post
Correctyou can do this with film.
(It's a joke, I hope you don't take offense. We may disagree in 1000 things but "not thinking" is not something I associate with you).
--- Judge what is said by the merits of what is said, not by the credentials of who said it. ---
--- Defend what you say with arguments, not by imposing your credentials ---Comment
-
-
More that you ever wanted to know about shutters (rolling shutter comparison between mechanical shutters and electronic shutters included).
But it is a very well made and entertaining video.
--- Judge what is said by the merits of what is said, not by the credentials of who said it. ---
--- Defend what you say with arguments, not by imposing your credentials ---Comment
-
Quite the production for nerd-adjacent topic that a tiny sliver of us will find interesting.Originally posted by Gabriel View Post
But it pretty much confirms that a CMOS rolling shutter sensor is needed to create my profile photo (with the camera-phone held sideways and the shutter travelling left-to-right). I think even an electronic film shutter would be too slow. My Contax has the fastest shutter of any of my film gear at 1/6000, but I don't think even it can create that photo. Or am I wrong?Comment
-
First of all, I am not an expert in photography (especially not as in being a photographer, although I suspect that I understand how cameras do their magic better than some very good photographers), but I am an an engineer so I feel entitled to talk about anything technicalOriginally posted by Evan View Post
Pay attention to the video in the 7:43 / 8:33 segment.
This is what I understand:
1/6000, 1/6 or 1/600000000000 doesn't matter. That only tell you how long each pixel is collecting data (electronic shutter) or exposed (mechanical shutter) or how long each grain of film is exposed. I will use "pixel exposed" going forward for simplicity.
That will make a difference on the "motion blurriness" of the image if the camera or what is being photographed is moving. For example, during its exposure time, a pixel may receive prop coming from the propeller and then light coming from the sky once the propeller moved out of the way.
But it tells you nothing about the rolling shutter effect. Rolling shutter is related to different pixels (or specifically different rows of pixels) being exposed at, and hence capturing, different moments of time. In principle, you can have 1/6000, 1/6 or 1/600000000000 exposure with all the pixels exposed at the same time (what is called a global shutter in contrast to a rolling shutter, and there arre digital cameras with global shutter and hence zero rolling shutter effect), or with any exposure you can have the first row exposed a few hours before the last row (that will never happen in reality, unless it's intentional and even then I don't know if any camera would have this flexibility, but you can always capture high-speed video and then replicate this effect).
Captain D says that a mechanical blinder-type shutter typically swipes the whole frame (both the bottom blade moving down hence opening the slit as the top blade moving down hence closing the slit) in 3 milliseconds, regardless of the exposure (the exposure would be given by the width of the slit which in turns depends on how much later the after the bottom blade starts to move down the top blade starts to move down). That means that the bottom row of pixels will start to be exposed, 3 milliseconds after the top row of pixels starts being exposed (and the same goes for "ends being exposed). In other words, the bottom row of pixels will capture a range of time (how big that range will be determined by the exposure setting) that is 30 milliseconds offset (later) than the range of time that the top row of pixels is exposed, hence the bottom row will capture the propeller in a position 3 milliseconds later than the top row, creating the rolling shutter effect.
Now, 3 milliseconds is fast. A prop at 2400 RPM will turn 40 degrees per second or 0.12 degrees in 3 milliseconds. If the prop diameter was 1.5m, that would mean that the tip would have moved 3 mm (nerd fact, actually pi millimeters) between the time when the top row of pixels was capturing and the bottom row was capturing. That's almost nothing.
But he also claims that a typical CMOS will do the same (digitally) in 10 milliseconds, which would be 0.4 degrees or 10mm at the tip. Also not much.
So either Captain D is mistaken with these typical times, or you your profile photo was taken with a cheap digital camera
--- Judge what is said by the merits of what is said, not by the credentials of who said it. ---
--- Defend what you say with arguments, not by imposing your credentials ---Comment
-
To simplify that, the lower blade moves down when the shutter is fired and the shutter speed (exposure time) is the delay of the action of the upper blade to follow it. The travel time for each blade is typically 3ms regardless of shutter speed. At some point, you reach a limit whereby the speed of the rolling shutter makes faster exposures impossible. So, what is that fastest possible exposure with that 3ms shutter rolling over a 24mm frame?
Comment
-
That matches my understanding.Originally posted by Evan View Post
That doesn't. The exposure itself can be much less or much more than 3ms (1/300 sec approx). Because the closing blade can START moving much sooner than 3 ms after the opening blade (so you have a rolling slit) or much later that 3ms after (opening balde goes down, you have the full frame exposed for some time, closing blade goes down).At some point, you reach a limit whereby the speed of the rolling shutter makes faster exposures impossible. So, what is that fastest possible exposure with that 3ms shutter rolling over a 24mm frame?
This old film reflex camera has exposures ranging from 1 second to 1/2000 seconds (0.5 ms). Yet the opening and closing shutter blades swipe the frame always IN the same time (say 3ms) -but not AT the same time-.
--- Judge what is said by the merits of what is said, not by the credentials of who said it. ---
--- Defend what you say with arguments, not by imposing your credentials ---Comment
Comment