During the recent NAB Show in Las Vegas, I attended a panel discussion and screening of a series of tests called the Single Chip Camera Evaluation. The SCCE shootout, produced by an independent, ad hoc group named Image Quality Geeks, compared 11 single-chip digital cinema cameras, along with two 35mm film emulsions.
These extensive tests, designed for “apples-to-apples” comparisons, provided a comprehensive look at the following cameras:
- 12-bit Sony F35
- ArriRaw Alexa
- Kodak 35mm Vision3 film stocks: 5213-200T and 5219-500T
- RED One M-X
- Weisscam HS-2 high-speed camera
- Phantom Flex high-speed camera
- Panasonic AF-100
- Sony F3
- Canon 5D Mk II DSLR
- Canon 1D Mk IV DSLR
- Canon 7D DSLR
- Nikon D7000 DSLR
Cinematographer Bob Primes, ASC, directed the tests and chaired the panel, which included Stephen Lighthill, ASC, of the American Film Institute, and a half dozen other postproduction and lab honchos. Over a hundred technicians were involved in shooting and posting the tests over a period of several months before NAB. The screening featured a 25-minute film, which described the tests and showed direct comparisons among cameras, projected brilliantly on a large screen.
Each camera was shot at its highest quality level, rated at an ISO between 800-1250, using the least compression possible and RAW files when available. Cameras were tested for sharpness, sensitivity, exposure latitude, color space, compression, skin tone, and shutter artifacts. In some cases, images were recorded through the cameras’ HD-SDI outputs, for less compressed files than onboard recording.
The IQG group shot high production value scenes with sets, actors, and extensive careful lighting, as well as specially devised technical tests. Some results were shown graphically on screen, some (like skin tone) required more subjective judgment in a split-screen comparison. Offered as a work in the public domain, the presentation provided filmmakers the opportunity to evaluate cameras with meticulous standards. Full results will be posted soon at thescce.org, but these trends were clear from the screening:
- The Arri Alexa and the Sony F35 both looked great in most of the tests, as befitting these cameras which sell in the $80,000-100,000 range. The Alexa had great low light sensitivity and film-quality latitude.
- The F35’s color space was similarly gorgeous, though I thought it looked a bit noisy, especially compared to the Alexa.
- The two film emulsions showed the best resolution and great latitude, especially in the highlights, not surprisingly. Some of the digital cameras actually exhibited better latitude in the shadow areas. And the film shots looked grainy, especially side-by-side with pristine digital images. Film grain has never bothered me before; I’ve always accepted it as a form of analog integrity, but here, for the first time, I found it bothersome.
- The Red One M-X did well in many of the tests, especially resolution and latitude, but looked average in others.
- The Sony PMW-F3, which sells for about $13,000, looked great in nearly all the tests, clearly the best bang for the buck among the less-expensive cameras. The F3 scored especially well in sharpness and latitude.
- The Panasonic AF-100 (which sells for under $5000, the least expensive actual video camera in the tests), looked over-compressed and over-enhanced in this high-end environment, on a large, theatre-sized screen.
- The DSLRs generally scored poorly, especially in resolution, sharpness and shutter artifacts, though they tested well in latitude.
Primes and Lighthill stressed that the aim was not to establish which was the best camera. Any camera could be the right camera for a particular shooting situation, depending on budget, logistics, size, mobility, postproduction workflow, and a host of other factors.
Other panelists included Michael Bravin, Mike Curtis of Pro Video Coalition, William Feightner of EFILM, Bill Hogan of Clarity Image, Jack Holm, Matthew Siegel, and Josh Siegel of IQG. DP Nancy Schreiber, ASC, also participated in the tests but was unable to attend the event.