Wednesday, August 20, 2008

Photography - being precise about the unprecise

I feel I have been somewhat limited in my understanding of the how the sensor size and pixel count in a digital camera is related to the properties of the lens. As an example, the statement "6-8 megapixels is optimal for a compact digital camera" is not self-evident, perhaps it is not true at all. Perhaps 10-15 megapixel sensors are the way to go.

First of all, there is an excellent article about sensors vs. lenses at Luminous Landscape. The article is quite technical but rather rewarding, if you are interested in the physics/optics side of photography.

Secondly, at Dpreview there is a long-winded, fascinating and really, really technical discussion on this resolution topic related to the micro 4/3 format of Olympus/Panasonic.

I learned a lot from the discussion, and then noticed that I understood it all wrong - several times.

Especially interesting were ideas about re-invention of the digital camera. Do we really need a mechanical shutter? Why not use an electronic one? Then you could provide a speed of 30 FPS at full resolution. A complete rethinking of a (serious) digital camera is possible, and probably unavaidable, once the new technology gathers speed.

But let us return to the resolution question. What I learned is that the interaction between an imaging element consisting of discrete elements (binning the light into pixels) and the resolving capacity of a lens is a far from self-evident question. There is room for creative/artistic decisions.

An good example of this is the "bokeh" of a lens, that is the patterns of light produced by a lens from subject elements which are not within the focus region. Here you could say that you don't need a big resolution to solve the details, because they are blurry in any case. But the point of a photograph may well be to capture precisely the non-sharp part of the picture.

Thus, photography is not only about capturing a precise image, it is about the relation between the precise and the unprecise. This topic merits some further thinking, I feel.

No comments: