We’ve been experimenting with the camera in the iPhone. It’s an amazingly tiny device, but provides rather more useful photos – 2 megapixel, 1600 x 1200 pixels – than any other ‘phone camera’ I’ve owned.

I’m very curious about the details of the camera. Judging by depth of field and some other clues, I’m guessing the lens aperture is around f 2.8 (in fact, I just did a ‘Get Info’ on an iPhone pic and it reports f2.8, but, curiously no shutter speed or other data).

Which got me to thinking about how this camera must work. I’ve noticed that the low light performance is good – very little shake is visible, even in low light situations (unusal for me because I usually can only manage to hold it with one hand). My guess is that this is really a video camera (it does have a video mode): there’s no mechanical shutter – the pictures are really framegrabs from what may be an analog video stream.

Since framegrabs are essentially instantaneous regardless of light level, low light ’shake’ performance should be good (we’ll be experimenting to see if our hypothesis is correct). Exposure, then, would be handled entirely in software, amping up or down depending on scene brightness.

This in turn means the lens needs to be relatively high quality – with focus fixed just this side of hyperfocus – the closest point to the lens where objects at infinity are also essentially sharp. The lens is noticeably sharper in the center two-thirds of the image (typical for relatively wide apertures like 2.8), which gives many pictures the appearance of being shot with a focusing lens, since the subject of most photos is near the center of the frame. Neat trick, and good use of the physics of optics, if I’m right. Voila 2 pix from iPhone: Peet’s Coffee in Menlo Park and tulips at Stanford Shopping Center…

{ 0 comments }