Inquirer reports from the last week's Future World Symposium in London: Bob Gove, president and CTO of Aptina, said pixel sizes on both CCD and CMOS camera sensors are shrinking roughly in line with Moore's Law, pushing up resolutions.
Changes in pixel microstructure will make them more efficient, enabling them to capture scenes in low light. And, by processing multiple images captured in rapid succession, cameras can capture details of both shadows and highlights in a scene. Speaking after his address, Gove reckoned that by 2014 cameras will be able to match the human eye in its ability to easily capture a scene in virtually any lighting condition.
BSI?
ReplyDeleteIs the wavelength of light also shrinking in accordance with Moore's law?
ReplyDeleteYes, wavelength of light is also shrinking by a factor 10power0 every year. It started about 15 years ago.
ReplyDeletedoes anyone know where the slides can be downloaded?
ReplyDeletewhat do you folks consider to be the ultimate limit to a practical sized pixel for 2D imaging and why?
I'd think it is in the range of the wavelength of light...
Half wave dipoles are a wonderfully resonant structure. At 560nm (green light) that's 280nm, or about 7.3 gigapixels on a 24x36mm sensor. You can even use resonant pixels to do your color separation, without all this screwing around with filters...
ReplyDeleteOn the other hand, that's a bit past the diffraction limit of an f1.0 lens. The resonant pixels might still make that pixel count worth the bother.