"Brian Gaff" <
brian...@gmail.com> wrote in message
news:tajal5$207si$1...@dont-email.me...
> I always thought that semiconductor cameras were merely moving memory
> about in silicon. Every chunk is a frame, and its then handled from the
> place it has gone to depending on the system that is used. One assumes as
> you can now get extremely fast frame rates for video cameras, that this is
> dictated by how fast memory can be copied and processed.
>
>
> The concept of the shutter type is a mechanical construct. I remember many
> years ago watching a picture load into a zx Spectrum. The screen was
> divided into three, and the first row of pixels in the first third filled
> in etc, till that third was occupied, but no colour was there yet. then
> the second, then the third and finally the colour. It appears that this
> was the way the hardware was designed and so you were stuck with it as it
> was, but later on when the machine had a second screen, you could hide the
> screens and sequence the completed ones as an animation, just by switching
> the page of memory you were using mapped to the screen.
It depends how the camera chip converts the analogue brightness of a pixel
to its numerical value. Some sensors "freeze" the analogue value for every
pixel simultaneously, and then sequentially convert the voltages to digital
and read them out. Others convert each value as it is being output on the
data bus.
The former allows a global shutter - like an iris shutter on a film camera.
But it needs sample-and-hold technology for every pixel, or else an
analogue-to-digital converter for every pixel. The latter is a rolling
shutter - like a focal-plane shutter on a film camera. It is simpler because
you only need one ADC, but causes vertical objects to tilt if the camera is
panned horizontally past them.
I presume the effect is more noticeable with a video camera than it was with
a film camera because the "shutter" effect takes a whole frame (1/25 sec) to
pass from the top of the frame to the bottom, whereas focal plane shutters
on film cameras (eg 35 mm still) were usually around 1/125 second (you can
tell what it is because it's the shortest shutter speed at which flash can
be used).
As an aside... Most film cameras made the shutter travel in the vertical
direction because it was the shorter distance and so the shutter took less
time to travel it; some made it travel in the horizontal direction which
takes longer (so slower minimum flash-sync speed) but causes less
parallelogram distortion because you rarely pan (tilt) vertically: a slight
change in the aspect ratio is less noticeable than a tilting of vertical
lines.