On the heels of yet another 4k camera announced by Canon this week, I think it valid to spend a bit of time on the subject of resolution. 4K and 3D are both popular buzz words for camera manufactures these days. As with everything else in the world of retail, being as good as or better than your competition, at least in print adds, is going to sell cameras. So why 4K, considering most cinematic digital projection is still 2K.
Red, Sony, and now Canon are all in the 4K+ camera game, so why not Arri? Technology is always demanding bigger, better, and at least in cameras, more sensitive to light. That is the big drawback of high resolution. The physical size of the light collecting sensors directly effects sensitivity.
Taking a step back, Arri produced a research paper on the topic in 2008 called 4K+ Systems (sorry I don’t hava a link). They looked at two important factors, what is the resolution of 35mm film, and what is perceivable by an audience. An important point made was that the perception of sharpness puts more weight in contrast than resolution.
A Simple Resolution Chart can tell us a lot about the resolution of a film negative and further more about the resolution of the human eye. The idea is to find the finest detail before the lines blend together, thus determining the ability to resolve fine detail. With a chart like this, Arri was able to conclude that a s35mm film negative holds an equivalent resolution of 4153×3112 pixels. The resolving power of a human eye can also be determined in this manner. Based on their tests, Arri produced the following chart showing what projected resolution an audience could perceive. Their conclusion was that over half of the audience would in fact benefit from 4K resolution.Over Sampling an image, has a major benefit when talking about a digital sensor. Pixels create a hard grid pattern, they have a hard time dealing with fine detail that does not mesh with that pattern. My favorite example is watching Robin Williams walk out onto the David Letterman show wearing a pinstripe suit designed specifically to make the cameras go wild. When fine detail doesn’t quite mesh with that hard grid you get parts of the detail blending together and other parts holding up, an effect known as Moire or Aliasing. Capturing at a high resolution and downsampling can eliminate this effect as detail is lost. Over sampling on a Bayer Pattern sensor has benefits as well. Half of a Bayer Pattern sensor is dedicated to green (luminance) pixels. Red and blue (colour) pixels only have a quarter each. Sony’s new F-65 camera claims a new buzz word, True 4k. Sony over samples up to 8k (20mega pixel) so that every pixel within 4K resolution has red green and blue information.
As sensor technology grows, and hard specs need to be improved to sell cameras, keep in mind that the end result is achieving a quality image. I think that far too often quality gets confused with quantity, but a good argument can be made to find the best of both worlds.