American Cinematographer - May 2011 - 64

◗

Decoding Digital Imagers: Part 1
standard as: 720 (H) x 480 (V). The CCD cameras invented at that time were all horizontally sub-sampled [with imager photosites numbering less than the horizontal output standard]. That was all they could build! But as the ’80s progressed, manufacturers learned how to build more sophisticated CCDs, and by the late ’80s, they started to oversample, or super-sample, in the horizontal direction. To my knowledge, however, even today there is no professional standard-def camera that has [a 1:1 ratio of] optical photosites and [output] digital pixels. “Super-sampling facilitates subsequent digital filtering and helps produce a very nice ‘aperture,’ as we call it,” Thorpe continues. “In HD, however, in both the single-chip and tri-imager systems, you will find a lot of linkage between the photosites and the digital pixels. One reason for that is that when the first 3-chip CCD HD cameras came out, in the early ’90s, the creators decided to go as high as they possibly could [in terms of definition]. Even today we still see lots of HD cameras whose optical sites match their digital pixels, but, surprisingly, we also see lots of sub-sampling in HD camera systems. Additionally, there are several Ultra High Definition camera systems emerging on the market that are now super-sampling.” Resolution Clearly, there is some room for interpretation in talking about photosites, pixels and resolution. “The resolution of a digital camera is often defined as the number of pixels in the signal that’s being delivered, but that is inaccurate,” states Thorpe. “However, it is true that the topic of resolution and the specifications for it are bound up in discussions of pixels.” Looking at the specs of various camera systems on the market today, you see a range of different photosite dimensions utilized in various HD, 2K and even 4K camera systems. “There are a lot of differences to be seen in each of these specs,” notes

says Thorpe. “The resolution, the dynamic range and the tonal reproduction, all of the really important imaging attributes, have to faithfully be represented digitally. With some sensors, such as CMOS imagers, that analogto-digital conversion is done inside the sensor, but it’s still a two-step process: analog transformation followed by digitization. From that digital sampling structure, you then do a lot of processing and formulate interface signals that can go to the outside world in the form of a video-output signal to be viewed, or to be sent to a recorder or other various systems. “Now,” he notes, “is there a link

between the total number of imaging photosites and their resultant digital pixels? There can be, but there might not be!” In fact, it is not mandatory to have a direct 1:1 relationship between the number of photosites on a sensor and the digital-pixel dimensions derived from that photosite array. Indeed, the concept of an imager’s photosites directly corresponding to its digitalpixel output specs is fraught with confusion. Thorpe recalls, “In 1981, when the world created the first digital standards for standard-definition video, SMPTE defined the North American

May 2011

American Cinematographer

American Cinematographer - May 2011

Table of Contents for the Digital Edition of American Cinematographer - May 2011