4K/HDR: The state of the art - Part 1
Chris Bone
Issue: August 1, 2015

4K/HDR: The state of the art - Part 1

As the motion picture and television industries begin to work with higher resolutions, such as 4K (4,096x2,160 pixel image raster), and UHD (Ultra High Definition 3,840x2,160 pixel image raster), and examine delivery methods for increased dynamic range, some questions arise as to what the additional resolution and the higher dynamic range can mean to the production processes.

For television, such as sporting events, higher dynamic range can mean increased freedom of camera placement and framing, but the compositional display choices are not necessarily convertible to standard range viewing. As well, for the television and movie industries, it can be argued that there is an impact as to some creative choice points of potentially accommodating, or not, greater color representation, since the perceptions of colors can change when the colors are represented in a lower light state, including states of desaturation, or when extended in high saturation. Further, due considerations should be given for adequate bit depth to preserve image fidelity and prevent adverse display issues. Considerations can also be given for the added choice points in the adaptation of — or conversion of — high dynamic range compositions, similar to aspect ratio adaptations based on areas marked on camera viewfinders for “safe viewing,” or similar to the use of viewfinder distinctions for exposure or contrast. Arguably, high dynamic range is perceived as having a greater impact on viewing than additional resolution, although it could be pointed out that the depth of modulation could in theory be needed to be twice as high for 4K/UHD viewing of normal contrast dynamic range in order to properly see that resolution even in square wave representation.

When observing adjacent complementary hues, for example, it has been said that the viewer can perceive a strobe effect where they meet, but thin black lines (can) keep them under control (1). The additional resolution of the 4K image presentation allows for more subtle transitions in line, form, color, and tone. In visual effects layering, the treatment of the edges of forms, when they are set over complementary hues, may necessitate subtle blending to separate the forms from each other, and time variable manipulation, and relative transparencies of edge color correction. Various production techniques to design or photographically-treat subjects and objects to be “read” against backgrounds, and the compositional aspects of framing and deep focus are potentially more important as to maintaining continuity and sensibility when viewing the combined images in higher resolution presentations. In theory, with more resolution in front of the camera, the cinematographer’s work becomes all the more challenging to maintain enough obvious and subtle control and directed consistency to influence the image presentation in the various forms and derivatives it might eventually take.

As mentioned, the additional resolution can bring more transitional areas into presentation form, and when photographing natural subjects, control of color harmony, or the notion of maintaining or manipulating “optical color balance” can be all the more challenging. There is no reason colors can’t be assigned and used more or less at will, but historically, certain types of color techniques, such as in classic painting, have been appreciated all the more over vast periods of time for their perceived color harmony, or balance. Predictability of color, when mapped with tone and coming through the image capture process, has been observed to be improved through the use of accurate color temperature control in lighting, including the spectral continuity of the light sources themselves. As objects seen in the natural world can emit or reflect colors outside of a conditional color space boundary, selective manipulation in both cinematography and presentation is sensible to try to control or encourage an end result that approximates the desired colors in both the bright and the dark parts of the image, and all the way through to presentation. No doubt through history, many a famous painter complained how his or her work was being seen, when the lighting conditions of the viewing were not the lighting conditions in which they worked. If the color and viewing lighting were all in harmony, when the painting (or artist’s palette) were spun fast enough in a circle, it would appear to be the same grey (2) in both lighting conditions.

When images are changed through digital filters, image processing and color management manipulation, some considerations for presenting higher resolution content to the viewer, such as 4K, should probably include being able to observe the changes while viewing the native resolution of the image itself, and then viewing the effects of the changes within various presentation forms and image scales. This necessitates proper and appropriate image conversions for image and color space scale, and making or approving choices that are satisfactory in both critical and practical monitoring. Differing contrast capabilities of image display technologies for 4K/UHD have to be taken into due consideration. With the minimum standard for on/off film print contrast at 4,000:1, and 2,000:1 for digital projection, monitoring on/off contrast at these levels with monitors set for matching brightness and the same viewing conditions is necessary to be on par with a projection experience. Moreover, the net result representation of average reflectance may change based on the display gamma or the projection or viewing conditions.

Certain monitor technologies employ local dimming, and the emergence of new technologies and presentation manipulation can exceed the contrast in theatrical settings. In any case, the shadows and the highlights, the midranges and the colors, can all differ between technologies, environments, and viewing trends.

Starting in the digital camera, the digital image files themselves are often coded differently to address various mappings of color and tone coming off the image sensor, or in conjunction with the manufacturer’s varied choices in color management. Observing the coding issue, as well as observing spectral differences in sensor response and from new forms of lighting, the Academy of Motion Picture Arts and Sciences (AMPAS) generated a great deal of research for which, through the AMPAS Science and Technology Council, subsequently presented to the industry in the form of recommendations for best practices of image exchange, color management and long-term archiving, the Academy Color Encoding System (ACES). Among other things, ACES outlines the normalization of these effects such that camera image data can be brought into digital manipulation more consistently, given the interest of the camera manufacturer to provide the appropriate input device transforms for their camera sensors.

Along these considerations, however, are the related concepts and complexities of moving color and tone through appropriate conversions when necessary, given the permissible color palate of either the original color space, or the new color space, and maintaining desirable results or avoiding abrupt transitional artifacts such as clipping. This, as of now, is left to the manufacturer to shape their sensor data into a “raw” form, and then to provide the appropriate Input Device Transform, which takes into account the manufacturer’s sensor coding efficiency, debayering, etc., and when used for ACES workflows, such that the image can be convincingly adapted into the ACES workspace as a linearly-encoded starting point.

For spatial conversion, both predictive methods and recursive methods exist outside of ACES, as well as various adaptive methods, and due to many factors, different methods do not necessarily produce the substantially same observable results as each other, especially given limitations in display technology and differences in viewing conditions.

It is said to be possible to create practical expanded gamut displays, such as those emulating ITU-R BT.2020 (3) color gamut, that are nearly or at least appropriately approximate enough to those that use more precise or narrow wavelength Rec.2020 primary colors (4). For example, if you can reasonably represent DCI P3 (movie) color space primaries on a given flat panel display, that may translate into being able to represent up to 98 percent of DCI P3 color space, for instance, although such an approximation might aggregate into some visible errors as the limits of the spectral locus are approached with high saturation. For the cinematographer, graphic artist, or anyone vested in the outcome of their imaging through the various changes that may occur to it, close attention to all of these factors and to viewing conditions themselves, is likely desirable.

Post’s Special Report on 4K/HDR continues in next month’s issue.  

Credits:
1 p. 134, Color Harmony, A Guide To Creative Color Combinations, Hideaki Chijiiwa, Rockport Publishers, 1987
2 Ibid. p. 137
3 International Telecommunications Union Radiocommunication Sector Recommendation ITU-R BT.2020
4 Quantum Dots and Rec.2020: Bringing the Color of Tomorrow Closer to Reality Today, SMPTE Motion Imaging Journal, May/June 2015, p. 24 by James Thielsen, James Hillis, John Van Derlofske, Dave Lamb, and Art Lathrop

Chris Bone is the CTO of VTP (www.myvtp.com) in Burbank, CA.