Hot Button Discussion The 4K Benchmark By Michael Goldman
The buzzword "4K" was a major theme through the recent NAB 2011 show in Las Vegas, particularly at the DCS 2011. In terms of 4K image capture, the issue was a hot topic partly because of Sony's debut of its new F65 CineAlta Digital Motion Picture Camera. Sony is promoting the camera as a true 4K (4096 x 2160) digital imaging system for high-end movie production built around an 8K 20-megapixel CMOS chip with new architecture that permits almost half of those 20 megapixels to be dedicated to green channel photosites for each pixel, allowing the capture of wider color space, since the green channel is the foundation of the luminance signal on digital sensors. Along with F65 comes a corresponding workflow methodology that outputs 16-bit linear RAW data and is built specifically to accommodate the Academy Image Interchange Framework-Color Encoding Specification (IIF-ACES).
"4K," technically refers to 4,000 pixels of horizontal resolution in a digital image capture format or in scanning 35mm film frames. In practical terms, however, "4K" is a far more complicated question. Until recent years, the notion of image capture at 4K in a digital camera system was more of a comparison concept—a way of evaluating just how close digitally acquired imagery could get to emulating the spatial resolution and certain other properties of a 35mm film image. Then, the industry started exploring ways to build image sensors that could actually record 4K imagery comparable to film imagery. The now-defunct Dalsa Origin technology was the first serious attempt at this, using a 4K (4096 x 2048), 8.3-megapixel CCD Bayer Pattern sensor capable of 16-bit output. The camera system's size and ergonomic limitations made it bulky to use in the field, and then the Dalsa Corporation withdrew entirely from the motion-picture business and the effort was shelved.
Then, of course, the Red Digital Cinema Company significantly impacted the industry with the development of its Mysterium-X sensor, now found inside the compact Epic camera system. Mysterium-X is a 5K (5120 x 2700), 14-megapixel Bayer Pattern CMOS chip capable of capturing a 5K image on the chip before generating a 5K RGB image that can be sub-sampled to any desired output resolution. Red One cameras with Mysterium chips, and now Epics, are in wide use across the industry, having a major impact on the broadcast television industry, and are now also proving quite popular as cameras of choice for 3D stereoscopic rigs due to their relatively compact size and weight, combined with the additional pixels they can capture.
And now, the F65 has entered the fray and is expected to be in the field before the end of the year. The camera represents the most recent development in the arena of 4K acquisition, and cinematographer Curtis Clark, ASC, feels it is a significant development. Clark put the F65 through its paces by recently producing a short film called The Arrival. The film was part of a lengthy testing process for the camera involving Clark on behalf of Sony prior to its debut at NAB. There, the short film premiered at the DCS 2011 (co-produced by SMPTE). Clark was on hand to discuss production of the short, his tests with the camera, and also to moderate a panel examining some of the subjective and artistic impact issues that result from acquiring imagery in 4K.
Clark, of course, is the longtime head of the ASC Technology Committee and, as such, has been working with the cinematography community for years on the issue of how best to develop and evolve higher resolution digital motion picture cameras that can emulate, or someday surpass, the spatial resolution properties of film. He recently spoke with SMPTE Newswatch to discuss the relevance and impact of 4K image capture as a potential standard for the larger industry.
Clark's use of the F65 leads him to personally believe the camera's sensor is "a game-changer" in terms of capturing and outputting true 4K data. But regardless of what approach one prefers, Clark emphasizes that the arrival of 4K-capable cameras generally illustrates the advent of a clear new industry trend on the image capture side. That trend is the fact that digital image capture for high-end motion picture work is clearly moving inexorably beyond what Clark calls "the previous fundamental threshold" of 1920 x 1080, 10-bit, 4:4:4 HD imagery in Rec. 709 color space. In other words, digital cinematography has finally surpassed HD capture, and that reality offers new creative opportunities for moviemakers and new viewing experiences for their audiences, in Clark's view.
"Digital motion picture cameras basically started in the HD format—1920 x 1080—constrained within the parameters of Rec. 709 color gamut with limited dynamic range," Clark says. "Then, some digital cameras evolved capabilities that move beyond the traditional limitations of HD, mainly in terms of dynamic range and the color gamut the camera can capture. But they still had 1920 x 1080 spatial resolution boundaries that were slightly sub 2K. Then, Red and (Arri's) Alexa came along with greater spatial resolution and their own (data formats) like Arri Raw and RedCode. Alexa, though, is still basically a 2K-plus camera, while Red claims a 4K image. And now we have the Sony F65, which I believe unambiguously reproduces true 4K resolution images that incorporate a wider dynamic range of at least 14 stops from highlight to shadow with tremendous sensitivity between 800 EI and 1600 EI. Sony designed an imager that is 20 megapixels with an 8K horizontal pixel count. This exceeds the benchmark references we normally associate with standard 4K in 35mm film. So that means we are now heading toward an imaging benchmark that sets a reference beyond 35mm as the standard. I'm not saying it is necessarily equivalent to 65mm, but it is a significant step in that particular direction."
"So all of these developments mean the industry is reaching a new plateau, where the benchmark is raised and it is possible to digitally capture at true 4K, which has long been associated with the highest film imaging quality generally available in the digital intermediate environment. We can now get there—we can make 4K the primary target going forward for digital image capture. What changed was the realization that the technology behind these cameras could move beyond the restraints of the (original) HD broadcast parameters (for which they were originally developed)."
Clark emphasizes that Sony's and Red's ongoing initiatives, and other industry developments include many important, related strides forward in areas ranging from workflow (Sony's endorsement of the IIF-ACES spec is a major development, according to Clark, as the industry moves toward IIF-ACES as the basis of a standard data workflow encoding scheme), to size and weight/ergonomics, where Red is currently offering all sorts of innovations, from recording media (Sony is offering new SRMemory Card technology for its workflow), to easier/better future proofing of content, and more. Some industry types think these breakthroughs are even the start of sorting through another industry Holy Grail—how to develop a stable and long-lived digital archiving medium.
But at the end of the day, Clark emphasizes what should be obvious—the impact on the creative process and the audience's reaction to that process is all that really matters, no matter how impressive technical paradigm shifts may be. In that area, Clark sees a ream of creative possibilities resulting from the 4K acquisition shift in general.
"For the cinematographer, it can be more liberating because these cameras can work at low light levels without the consequences that sometimes happen with film if you get too brazen with under exposures," Clark says. "This is especially true of night scenes. Wider contrast, huge dynamic range—these developments will enable us to pursue certain aesthetic paths that filmmakers will want to explore without having to resort to remedial procedures in post to reduce grain, noise, and other 'fixes' of that nature. With 4K cameras like F65, you can see shadow detail more clearly in low light, but with far better spatial resolution. This addresses the major categories we've been striving to deal with in terms of matching the benchmark qualities of film—dynamic range, color bit depth, and spatial resolution."
As a result, Clark opines, cinematographers will be able to "resolve finer details and produce nuanced textures and tonalities better than in the past, which will facilitate a more seductive viewing experience with the prospect of creating more compelling, i.e., cinematic, visual narratives. These cameras are giving us a more expansive canvas and color palette than we've had with previous digital systems, or even with 35mm film.
"So that's what we mean when we say 'true 4K,' " he adds. "Expanded dynamic range from highlight to shadow—a minimum of 14 stops with the F65 and they are still testing that—in conjunction with tremendous low light level sensitivity without noise, wider gamut (than film print reproduction), along with true 4K spatial resolution. These exceed the benchmark references we normally associate with 35mm film. This really sets the stage for greatly enriching the expressive capabilities of filmmakers. We can enrich our ability to utilize these attributes for the purpose of expanding our vision with fewer obstacles."
Another interesting technology development at NAB—this one in the 3D realm—was presented by a startup company called Meduza Systems (owned by 3D Visual Enterprises of the U.K.). Meduza introduced what it is calling a digital imaging system for stereoscopic 3D capture—a single camera, in other words, with a single set of electronics, rather than a rig designed to unite two cameras for stereo capture. The camera, called the Meduza, received an award from TV Technology Magazine at NAB. The concept of a single, high-resolution camera with a single set of controls, built-in remote-controlled inter-axial convergence capabilities and interchangeable lenses at a relatively light weight (15 pounds) is one industry experts have long believed would be a major step forward in overcoming some of the complications of 3D image capture in the field. Of course, the camera still needs to be put to use on major real-world productions before it can be determined if Meduza has provided the right solution.
If current 3D rigs aren't your preference, and you aren't ready to try a brand new system like Meduza's, another option has percolated into the marketplace—the Multiple InterOcular 3D approach. A company called MIO 3D™ is offering it as their solution to the 3D acquisition challenge. MIO 3D™ is a company founded by cinematographer Sean Fairburn, which takes the approach that three or more different cameras can be used to capture stereoscopic imagery, increasing options for image convergence in post. The solution is achieved by offering solid fixed mounting points for three or more cameras in succession on mounting plates called MultiPlates, along with new slating technology (called MultiSlate) to allow creation of a 3D alignment chart to provide data for synching and converging image pairs later, resulting in multiple options for which pair of images to converge, depending on a project's needs.
Meanwhile, cinematographer and longtime industry journalist David Leitner has been comprehensively documenting his NAB impressions in categories related to moving image capture and processing for various publications and websites for several years. This year, his extended impressions can be found in a three-part blog for the Filmmaker Magazine site. Part I covers NAB's press conferences and business news; Part II analyzes Apple's ballyhooed release of Final Cut Pro X editing software; and the fascinating Part III takes a look at numerous developments and trends with recording media and storage technologies.