Pixel Counts & Sensor Sizes

I wrote my first digital camera review when the best cameras were still 640x480 resolution. That was what people called VGA resolution and matched what most monitors used when Windows 3.0 hit the scene in 1991 or so.

Windows 95 had about 90 million users when the first one-mega-pixel camera hit the scene. I already had the Epson Photo-PC and had shot many images for the online magazine, "Connections" where I was editor-in-chief. Those images were small, but, they looked pretty good. You wouldn't want to print them, but, they were posted on the web for readers around the globe.

Since then, I have owned about fifteen digital cameras, but until I moved to digital SLRs, I did not take them seriously, even though I had been using them for years. I had a film scanner and I used it more times than I can count. I posted shots to my online photo gallery and I digitized images for my photography clients, too. I was waiting to see how good digital SLRs would get and held off investment until technology proved that it approached film in quality.

Digital SLRs have advanced to the point that they are now, clearly, competition for high-end film SLRs. There are countless discussions and comparisons between film and digital interchangeable lens cameras online and in print publications.

Canon has 35mm sized "full frame" sensors in their high-end professional cameras. Nikon says it never will, at least at this point. Pentax uses APS-sized sensors manufactured by Sony. Olympus uses 4/3 sensors and Kodak has pledged support for that specification, yet, they produce full-sized sensors.

On an APS-sized sensor, the more pixels (photo-sites) you pack onto it, the more electrical interference you'll generate. This will result in making "noise" that will be hard, if not impossible, to filter. At what point do we say, why bother?

Until some new technology enables us to make more efficient photo-sites that do not create electrical interference, I'd recommend that camera companies stop increasing pixel counts and start working on other refinements. We could benefit from other improvements in image quality, like better color latitude, and better printing technology.

In favor of Nikon's approach to the digital SLR, the APS-sized sensor places the chip in the sweet-spot of a 35mm camera lens. A full-frame sensor would be more prone to light fall-off in the corners and softness. I'll explain... digital sensors are made to absorb light striking them from a 90 degree angle. When you change that angle, color fringing is the result. Your images have a strange color halo effect most visible during night shots around sharp points of light. The most efficient angle, then, is when light strikes the lens at a perfect 90 degree angle.

Film does not work the way digital sensors do. Even at sharp angles, film produces a fairly high quality image. By using a full-frame sensor, the closer to the edge of the sensor you get, the more off 90 degrees the light angle will be. Efficiency drops and the image quality goes down.

Nikon's point is this: Why not keep the sensor in the sweet-spot and let's not get hang-ups about full-frame just because Canon elects to use that size sensor. Nikon image quality is excellent and the D2x is living proof, so to speak.

One last point... When you make enlargements, more pixels store more data, but, if the photo-sites are too small (from having so many in such a small chip surface-area), will the enlargement quality be poor? Think of the sensor like a sheet of graph paper. If you create a picture by filling in the little squares, the picture will be bigger if you fill the same number of squares in one-inch paper as you would with quarter-inch paper. With one-inch paper, each square is much larger than quarter-inch paper (I'm no math wizard, but, it seems like 16 times larger to me) so the enlargement you print would be a lot clearer and better quality.

Think about it.