Rob Studdert wrote:
On 2 Mar 2004 at 14:42, Robert Gonzalez wrote:
Does a CCD suffer from reciprocity failure also? If so, are the curves published anywhere?
I believe that they are pretty linear from their minimum exposure time (determined by the matrix read speed) out to lengthy exposures. The main problem is self noise which will eventually saturate (fully expose) all pixels regardless of light hitting them or not.
I agree that self noise is probably the limiting factor on the long exposures. I'm guessing that it makes sense that they are linear. But, the matrix read basically only limits how fast you can continually take pictures (added to other parts of the capture pipeline), I don't believe that it limits the minimum exposure time, if you can get the shutter speed to 1/20000, I believe you can capture photons (in bright enough light!) that will represent the image. I think the matrix reading happens post-capture, and it has to undergo A-to-D transformation (after scaling for ISO), followed by memory write, and finally conversion if necessary to JPG or other formats and final write to card.
rg
