I'm not certain I'm asking this correctly as I don't entirely understand how such a camera sensor works, so help me out with a couple of areas please!
I want to program something that makes use of the effects caused by the line-by-line scanning of an electronic rolling shutter camera sensor, as is used in modern smartphones. What I don't understand is:
Does each line of pixels only "accumulate exposure" at the moment it's being scanned, thus capturing only what visually occurs at the very microsecond (or so...) of the single line's scanning?
Can the scanning (and exposure) time for each line of pixels be varied to reach a different scanning time from one end of the image to the other?
Are long-exposure photos (or any-exposure in general) made of many quicker whole-sensor scans superimposed on one another? This I'm asking because you don't end up with sectionally-distorted images when long-exposing a moving scene but rather end up with smooth trails of movement, and that means the exposure can't involve only a single slow scan through the array of pixels.
And my last question is, what is a usual scanning or "readout" time for a line of pixels? One article I read claimed it's up to 8 microseconds, thus saying an average (according to article) 2,500 pixels wide smartphone camera takes 1/50 of a second to read out from first line to last (and then saying you'll thus have a 1/50 sec rolling distortion). But if I take a photo outside with my Xiaomi Mi 6 while dimming the exposure down, the phone chooses a declared exposure time of 1/90,000 sec. So... what does that mean? Are the explanations I've been reading wrong or just badly explained?
Thanks in advance.