High dynamic range and wide color gamut combined with 4K resolution and progressive frame rates have catapulted broadcast television to new levels of immersive experience for the viewer. As HDR and WCG are relatively new to television, we need to both understand their application and how we monitor them to enable us to surpass the levels of quality and immersive experience cinematographers’ demand.
Although engineers have worked with creatives and program makers throughout the history of television, HDR and WCG brings a new level of creativity that needs to be understood. The “artistic intent” isn’t just a number we can measure to or calibrate for. Instead, to understand this completely, we must have a much greater awareness of the creative and artistic process.
In television circles, camera log curves seem to be spoken about with almost an air of mystery. They may be used extensively in cinematography but are seemingly new concepts for broadcasters. The irony is that we’ve been using gamma curves for as long as we’ve had television, and these are just a variation on a theme of camera log curves.
PQ and HLG further adds to our interest as they seem to be very similar but are also worlds apart. HLG solves many of the challenges for live television and many argue PQ provides greater creative freedom, and this is even more apparent when we look at future proofing our archives.
HDR has abstracted away the concept of peak brightness in terms of peak white being a particular voltage level. NITs are playing an increasingly important role for measuring brightness and how this compares to voltage levels is often challenging, especially when making programs for PQ and HLG.
As well as providing the new vibrant Rec.2020 color space, WCG has exposed some of the inefficiencies of our traditional YCbCr color difference signal representations. The distortions caused in 4:2:2 and 4:2:0 color subsampling may not have necessarily been apparent in standard dynamic range, but the associated higher quality of Rec.2020 lays these artefacts wide open, becoming even more apparent in post-production.
New color difference systems such as the ICtCp representation provide some solutions to these distortions due to constant illuminance representation. This makes distribution through traditional workflows a challenge and is reason for further consideration as we build cinematography compliant workflows.
HDR is much more than just a marginal increase in picture quality. It opens up a whole new level of creativity that we must work with and embrace.
Television is still constrained by the decisions made in the 1930’s for black and white and 1960’s for color transmission. Frame rates, resolutions and color spaces are all affected. The predominant influencing factor was the cathode ray tube as this was the major source of limitation. The resolution of the CRT was not only limited by the granularity of the phosphors, but also the light spill to neighboring phosphors causing a gaussian response resulting in reduced contrast as well as resolution.
Broadcast television has entered a new and unprecedented chapter due to the need to provide high end cinematography productions. Although high quality productions have been in existence for as long as television has existed, their quality has been constrained by the limits and compromises live television imposes. These articles look at how HDR has influenced our ability to produce cinematography productions.
The advent of CMOS and CCD image gathering devices, along with flat screen television, has greatly improved the viewer experience. A typical dynamic range for a CRT is in the order of 100:1, but modern cameras, with their improved sensors, now have a dynamic ratio of up to 10,000:1. To correctly resolve this range and avoid contouring, 14 bits of resolution is required. These sensor improvements also make wide color gamut available giving a greater and extended color range, especially in the greens.
The human eye’s sensitivity to contrast is non-linear in nature. That is, the greater the brightness, the greater the contrast ratio between objects must be for the eye to detect any differences. This relationship is known as Weber’s law.
We use Weber’s law to help determine the bit depth of the data needed to represent the sample. In digital video, our ability to detect the difference in brightness levels between adjacent quantization levels indicates the effects of contouring. Our ability to detect the difference between brightness levels is proportional to the brightness of the area being considered. This leads to the “just noticeable difference” between light- levels being a fraction of the light level. Also known as the Weber fraction, it determines the response of the cones in the eyes, that is, our color receptors. The eye can detect a change of 2% to 3% in brightness levels.
The 2% Weber fraction leads onto the 8-bit digital coding system used in broadcast television for Rec 709 systems, that is those based on the CRT technology. For cameras with a dynamic range of 10,000:1, then 14 bits are needed giving 16,384 quantizing levels.
Even though broadcast infrastructures tend to use 10-bit SDI, the challenge we have is how do we make a 14-bit camera video feed work in a 10-bit SDI system? Especially when we consider that broadcast delivery systems to the home are 8-bits. If we just truncate the 14-bit data down to 10-bits, then we will have high visibility banding.
In the analog days, broadcast television used the gamma curve to produce a uniform perception of video noise. The noise was generally introduced in the transmission chain which resulted in the CRT display compressing it when the opposite function was applied during display. This inversed gamma function was a natural attribute of the CRT’s current to light response.
One of the properties of the human visual system is that we can perceive noise more in the shadows than we can in the highlights. This also means that we have a greater ability to detect contouring in the shadows. Gamma helps provide greater resolution in the dark areas of the screen to reduce contouring as well as giving better signal to noise.
The primary purpose of S-Log, HLG and PQ is to fit the 14-bit video into a much smaller data space, in the case of SDI this is 10-bits. This diagram shows how S-Log takes a 14-bit video signal and converts it to 10-bits using the logarithmic transform. It has the added benefit of expanding the shadows and sympathetically compressing the highlights so more of the original image can be recovered during grading and editing.
Knee Dynamic Range
Even in SDR, vendors have been able to provide a dynamic range greater than that of the processing and display system. The “knee” is a method vendors use to increase the dynamic range limitations of the 8- or 10-bit broadcast infrastructure and delivery systems. The knee provides a compression function in the highlights to remove the effects of the hard clip and reduce the ballooning of specular highlights.
The primary reason for providing the knee was to extend the effective dynamic range while at the same time maintaining compatibility with existing working formats. The only problem with the knee is that each vendor has their own take on what it should do and how it should work. Consequently, there is no standardization and it’s very difficult to work with in post-production.
Broadcast television fundamentally differs from cinematography as it is live, and the emphasis is on making programs work now, as opposed to recording them and editing in post later on. This allows cinematographers to spend much more time on making programs and editing them later, time live broadcasters cannot afford.
One consequence of this philosophy is that cinematographers record images to a much higher rate than we would generally have in broadcast television. 4KP60 4:4:4 images can be recorded, and the new SMPTE 2082 standards even provide the option for 12-bit distribution. But even at this high resolution, there is still a need to provide a method of compression so that the images can be mapped from 14-bit resolution.
You might also like...
IP is delivering unprecedented flexibility and scalability for broadcasters. But there is a price to pay for these benefits, namely, the complexity of the system increases significantly as we add more video and audio over IP.
Building optimized systems that scale to meet peak demand delivers broadcast facilities that are orders of magnitude more efficient than their static predecessors. In part 2 of this series, we investigate how this can be achieved.
When composing and lighting scenes, DOP’s usually seek to maximize texture and perspective. The rationale is simple: We live in a world that is unmistakably three-dimensional, so DOPs seeking to faithfully represent the natural world exploit a range of w…
Aside from being the first Summer Olympics to be delayed a year due to a pandemic—shifting technical plans and causing strict work-arounds to comply with health restrictions—this year’s live coverage by NBCUniversal (NBCU) is noteworthy for its move …
For many years broadcasters have been working with static systems that are difficult to change and upgrade. This two part series explores the unfolding of a more elastic future based on COTS hardware and flexible licensing.