How do they get enough frames for the length of time they need.
Do the little sensors react that fast or do they have to have several sensors panels to spread the exposures around to have that high a frame count?
I have seen digi cameras with multi-burst capabilities and shutter speeds of 2000 of sec. But sustained through several hundred frames or thousands of frames?
If it is just software, why cant little cameras have it, 4 GIG & up memory cards are available at reasonable prices.
Unless I’m missing something obvious about your question, they do it with high-speed video cameras. You don’t use a digital SLR for high-quality video.
They get enough light using the same techniques as film… sensitive CCDs, shooting in bright light(i.e. sunlight), large lenses that gather lots of light. And yes, its just one CCD, which reacts that fast. Expensive though.
They really don’t get enough frames… Watching the making-of about the BBC documentary Planet Earth, they used a high speed digital camera to capture images of sharks jumping, and the camera only had enough memory for 3 seconds, which it would continually overwrite with a HUGE amount of data. When a button was pressed, the current data was saved, and it took them something like 15 minutes to transfer it to a PC for inspection. Whatever was storing the info inside the camera must have had absolutely gigantic bandwidth.
Oh, and film high speed cameras dont have all that much film either… a few seconds worth(depending on the speed).
I’m not sure if I understand the original question. Flashes are not used in in video.
But more to the point, it depends on the video format. For most video cameras, slow motion is done in playback, not in the camera. Take an television coverage of an football game. The video highlights in slow-mo are done by the video decks playing back the video slowly. To keep the image clear, the video cameras will shoot with a high shutter rate, usually 1/1000 or 1/2000.
There’s a large Quicktime video on the page. If you don’t see anything, you either don’t have Quicktime installed or don’t have its plugin installed for whatever browser you’re using.
This is high speed video shot with a digital still camera. The Casio EX-F1 will shoot video at up to 1200 fps and will shoot still shots at up to 60fps. Some DSLRs have video, but right now it’s not really a huge feature.
The instructor for my first film class had been in the Air Force (or something) and told us of a film camera that would shoot one million frames per second. He said this was achieved by having a continuous loop of film run up to speed in front of some sort of special shutter and then opening the lens to record the event.
I don’t know what’s available today, but in the 80’s I worked for Ford doing test crash analysis. They had film cameras that recorded at about 1000 frames per second. They did this with a rotating prism to project the image onto the film, following the film along until the next facet of the prism rotated around to capture the next frame. (Conventional film movie cameras slide the frame into position, expose the frame, then move to the next frame.)
I did not know the CCD’s could react that fast.
So it seems that film is still the fastest in frames / second.
The data transfer seems to be the bottle neck at this point?
How about a cluster of cameras, 5 say, timed at 1/5th of each other all running at 1200 F/s = 6000F/s? A tight cluster would minimize parallax shifting effects and you could then combine them into one detailed strip?
As has been mentioned there are CCD cameras capable of imaging at 1000s of frames per second.
My guess would be that the limiting factor at extremely high frame rates is integration time (shutter speed - sort of) rather than other bandwidth concerns because it is the only step that does not lend itself to parallelization.
If I’m correct that the limiting factor is integration time then the camera cluster idea runs into the problem of shared motion blur. If you understand what you’re filming well enough you might be able to apply some sort of temporal super-resolution black magickery, otherwise staggering camera shutter times will not gain you very much.
Indeed – and one of the things they mention often is that these cameras take an enormous amount of light (which in turn produces an enormous amount of heat) at the high speeds – sometimes limiting the sorts of things they can film.
I believe that in some cases of slow motion footage, digital tweening is used - that is, the footage is shot at something more like an ordinary, sensible frame rate, then a computer (perhaps under human direction/supervision) creates all the missing frames that would have existed, had a higher frame rate actually been used.
Obviously that approach would not work for all possible subjects - particularly in cases where things move in/through/out of frame really quickly.