Pixels and camera sensors do not exist - I compare just the idealized image seen by the camera because there is no reason to include misleading technical details into the picture when talking about motion blur differences between cameras. Obviously this assumes comparing different cameras but equivalent pictures, otherwise the comparison is pointless.

the sensor size is irrelevant because the camera captures angles, not locations. in other words: what matters is angular position in the field of view, not the absolute position. the same shake produces identical blur in a smartphone and in a large format camera (it is obviously easier to shake a smartphone, though)

this has been done already, see makezine "laptopogram": makezine.com/projects/laptopogram

unless, of course, we are taking photos of very hot objects, where thermal radiation enters the visible range ;-) then we could try to infer the temperature from the blackbody radiation colors

top-of-the-line DSLRs from two market leaders (nikon D5 and canon 1D X mark II) are too slow to read the whole sensor at 60fps (using crop 1.5 and 1.4 respectively in 4K mode), so yes, in 2018, the mainstream electronic shutter is still very slow compared to mechanical shutter. of course, smartphone sensors are always a bit ahead (but they are smaller), and there is sony, using multiple channels in a9.

oh, and i mean typical electronic shutters, not the fastest ones. sony A9 seems to be well ahead of the competition here (but still slower than mechanical). and of course it would be nice to see a global shutter consumer camera some day...

they are simply not possible ;-) at full resolution. even some new cameras from top manufacturers don't support 4K video recording! but of course things are getting better each year...

actually it is not a video frame. the linked video shows a smartphone app reviewing a series of photos taken in the burst mode. so even if the picture is extracted from a video, it is still a photo, taken in a photo capture mode. of course, smartphone means electronic shutter, so it's similar to what "serious" cameras do in video mode.

realistic shutter travel time is probably closer to 1/500 sec but yes, i doubt it really makes a difference for long exposures where most of the exposure is captured simultaneously for all pixels.

@MichaelClark "Your camera's light meter can't tell the difference..." - but it can! otherwise it would set the same exposure for both subjects, resulting in completely underexposed or overexposed pictures. therefore i i don't like this particular sentence.

@MichaelClark sure, it's all about the context! but your first paragraph is not true. camera can see, well, we have 0.01 lux (made up number) so it's a black cat in a coal mine, and here, 100000 lux, surely a white cat in a blizzard. yes, your explanation is correct and we agree that knowing the absolute brightness of the subject does not help selecting the "correct" exposure. but it's not because the camera cannot tell the difference between dark and bright subjects. (i can't even say THERE IS a correct exposure. it depends on what effect is desired after all)

the answer seems to contain valuable information on how the automatic exposure works, but the bold sentences in the beginning are very unfortunate and even plain wrong if taken literally. of course the camera can easily tell the difference between the black cat in a coal mine and the white cat in a snowstorm. what the camera sometimes fails to do correctly, is guessing if a given scientifically measured cat brightness should be pictured bright or dark in the photo, and this a different issue.