This is a fascinating paper. It takes a look at the fundamental limits of digital camera sensor performance from a theoretical physics point of view. The quick tl;dr summary: for 35mm full-frame cameras, low-light performance isn't going to get much better than 3200 ISO; if use an aperture smaller than f/8, you will start losing resolution due to diffraction effects; and it doesn't really make sense to go bigger than 21-24 megapixels. (Doubling the number of pixels just halves the number of photons that can fall on a particular pixel site on the sensor, thus increasing noise for a given ISO level.)

The paper is very thorough. It covers sensor sizes from full-frame 35mm to APS-C to small digital point-and-shoots. And life gets even worse for APS-C and small point-and-shoot cameras, in terms of physical limits.

The bottom line is that it appears that digital SLR's may have reached maturity in terms of sensor performance. Not only do most modern sensors have a performance where the optical performance of most lenses (even professional grade, uber-expensive lenses) are the limiting factor, modern sensors are pretty close to optimal from a physics perspective. After all, as every single Star Trek enthusiast knows, "you canna change the laws of physics!"

Addendum: the reasoning of this paper does make me question Nikon's decision to go to 36 megapixels on the D800. And it also made me pretty happy about my choice of purchasing the 5D Mark III; I can be pretty confident that at least for a 35mm full-frame sensor, it's not likely the camera will be obsoleted any time soon (especially since I'm not all that interested in trying to use a DSLR for video).
Shared publiclyView activity