Toshiba disables DOF-physics?
todays CES announcement by Toshiba about their new smartphone camera module, that is supposed to be able to refocus a image afterwards, is kinda driving me crazy...
The source Newswire doesn't give away that much of the actual technology behind it,
but from a logical standpoint I can't imagine it to be too different from the approach e.g. Nokia uses for their refocus app:
The camera focuses several distances and saves the data for each refocus-area seperately. It does that a bunch of times to have a somewhat reliable depth map and stores it to make it accesible for later uses. Having two lenses can't change the necessity of that, I thought. My guess was that it only uses the two 5MP sensors to process one 13MP image, just to do the same "focus-and-take-multiple-pictures"-process from then on... with the only benefit of Toshiba's module being that the areal view two lenses have going for themselves could suggest the location of things you might want to refocus later, and thus improve the focusing speed.
But then I caught a little detail that bursted my bubble:
the device doesn't require any focus motors
How exactly is this supposed to work?
Sure, when the sensor is small enough (which I'm not certain about, but 1/4" might very well be) the DOF would actually cover all of your sight anyway, but how do they archieve the refocusing, that allows you to blur out parts of the image afterwards?
Is there a way to produce a bokeh-like effect by using two images that were shot slightly apart from each other?... Like, digital depth-of-field?
Is the future for tiny-DOF-fans not a full-frame sensor, but four small sensors+lenses in each corner on a 35mm quadrangle, because the outcome is the same..?
I'm kinda confused about what I thought to know about DOF... What am I missing with this whole thing?