Within broadcast there has always been a quest for higher and higher resolution with improvements in wider color fidelity. The quest has always been to deliver what we see to the audience, often this is limited by technology or cost of production, but today there is the possibility to increase the resolution to 4K/UHD with High Dynamic Range (HDR) and Wide Color Gamut (WCG) that can convey that window on the world to the consumer.
Ian Valentine, VP of Product Management, Telestream.
This article was first published as part of Essential Guide: Practical High Dynamic Range Broadcast Workflows
However, these advances also pose a problem for content creators as they try to figure out how to produce content that looks great on the variety of displays that exist. It is becoming increasingly difficult to do this without a set of tools that provide some objective measurements rather than subjective observations on how the content will finally appear.
Artistic intent and composition aside, Exposure and Color Management are two key elements that content creators need to master and control to avoid costly mistakes resulting from acquired content that cannot be fixed in Post. How are these new technologies impacting the current workflows?
In its simplest form, exposure management in acquisition is about controlling the amount of light entering the camera and reaching the sensor. It is important to ensure that any captured image is neither overexposed, to avoid picture information being clipped in the highlights, nor underexposed causing the blacks or shadows to be crushed with the subsequent loss of information. Getting this wrong in acquisition will make it very difficult or impossible to fix the image in Post.
To handle exposure correctly Production staff cannot rely on looking at a monitor. A monitor provides no indication of exposure levels. Applying an HDR image to a standard monitor will produce an image that looks “washed out”. Even with a LUT applied to compensate for this, the image may improve but may not accurately reflect the final image desired or the exposure levels because of the response variations of LCD screens and external monitors. Objective measurements or tools based on measurements are required to ensure correct exposure.
There are two common tools used to assist cinematographers or camera operators in exposure management, the Waveform and False Color displays.
To objectively measure exposure (luminance levels) a waveform is used. The horizontal axis represents the frame and the vertical axis represents the brightness level associated with any chosen point in the frame. Traditionally the brightness level is represented by the IRE level that is better represented as a percentage scale where 100% is white and 0% is black. As the exposure level is adjusted the trace or display height will vary with the blacks being (ideally) anchored on the 0% line of the trace. With an SDR (ITU-R BT. Rec 709) gamma applied, as more light is allowed into the camera, the height of the display will increase until the brightest areas of the image hit the 100% point. Clipping will occur at levels above 100% to 109% depending on delivery specifications that define levels for maximum limits.
SDR Displays can be driven to the 100 to 200 Nit range in terms of maximum brightness. Initially 100 Nits (100 cd/m2) was used as the reference white, but this has changed to around 203 Nits in HLG. However, modern displays are capable of handling 1000 Nits and above which allows content creators to take advantage of the greater dynamic range offered by the latest cameras. The cameras utilize Log gamma curves (e.g. S-Log 2, S-Log 3, C-Log, Log C) curves designed to help capture as much data as possible in the luminance spectrum i.e. shadows and highlights. The consequence of this is that when applied to the same scene, the SDR and Camera Log waveforms will look different as the equivalent SDR white point is repositioned at about 60% of the IRE scale on a Camera Log scale, allowing the camera levels to be shown above this (see Figure 1). This makes it difficult to compare the content and to assess if the content being captured is acceptable for both those environments.
1 Figure 1: Shows an SDR (left) and HDR (right) waveform of the same image (a SpyderCube). On the SDR waveform, highlights are clipped and the 90% reflectance white is shown at that level on the percentage IRE scale. On the HDR capture the white levels are adjusted to be at about 60% on the screen. This illustrates the problem of trying to compare SDR and HDR signals on the IRE scales.
Additionally, for most people behind the camera the %IRE scale is meaningless as it essentially references mV level electrical signals from standards defined in the days of NTSC. Light levels and Stops are the common language of camera personnel. Converting the waveform to display in light levels now means that the reference levels, whether working in SDR, Camera Log or HDR, are consistent in vertical position and the waveforms are the same shape for easy comparison. PRISM can display these waveforms as a STOP Display where luminance levels are shown in NITs or STOPs. Unlike standard luminance waveforms, when using the STOP Display in acquisition, changes to the exposure settings will move the whole waveform trace up or down on the vertical (light level) scale and the operators can easily avoid highlight or shadow information being unexpectedly lost through clipping or crushing when trying to measure exposure levels. It also means the STOP display allows direct comparison between different cameras on different inputs of the instrument if the light levels remain unchanged. There are some key values to note when using this system; in acquisition 90% reflectance whites are normally set to be between 100 and 203 Nits, and 18% greys at around 26 to 32 Nits. These values will become important when using False Color to manage exposure.
Figure 2: Shows the same SDR and HDR signals as in figure 1, but in this case, they are displayed in a STOP display with levels being set by light or luminance levels. Comparison is easier because both images look very similar, the light levels can be compared. The key difference is that the specular highlights are shown on the HDR waveform, but are cropped (as expected) on the SDR waveform.
False Color images provide an easy to interpret and powerful interface for looking at exposure for both SDR and HDR content. These displays work in conjunction with the luminance levels shown in a waveform but the False colors show regions of similar luminance ranges. This allows the camera operator to easily identify regions of the images with similar luminance levels and to easily measure exposure levels at different points in the image.
In acquisition False Color makes it easier for operators to quickly and accurately set the exposure of selected image elements (e.g. skin tones, white lines or green grass on a football field) to the correct level by simply adjusting the exposure until the correct color appears on the image element. This approach removes the need to subjectively set the exposure by eye when looking at the image.
False Color screens are often shown as a rainbow of colors from one extreme to the other (representing different luma levels from darks to whites in a wide array of colors) which can be distracting. PRISM allows users to select their colors in user defined bands so that they can limit the number of colors on the image to only those that are most important to them – for example 100 to 203 Nits for whites and 26 to 32 Nits for 18% Grey (see Figure 3).
The use of HDR is not simply about producing brighter images. To create realistic images HDR should be used to emphasize pinpoints or small bright areas (specular highlights) in the image. PQ and HLG with maximum luminance level and screen gamma defined are the two most common screen referenced HDR standards. While a modern display may be rated at 1000 Nits, only a small percentage of the display can be driven at this level at any one time. This means that understanding not only the exposure levels of these highlights, but what area of the screen is being driven in the HDR region is very important to those responsible for acquisition as well as those working in Post Production. PRISM provides a luminance False Color display that will highlight which areas of the screen are being driven in the HDR region (normally above 203 Nits) and will also provide measurements of what the minimum level of the brightest 1% and 10% of the screen and the maximum luminance level of the darkest 1% of the screen (see Figure 4). This allows cinematographers and video engineers to acquire correctly exposed HDR images and Post Production staff to objectively adjust levels to give a comfortable and realistic viewing experience.
Figure 4: Left display shows lightest and darkest areas in the HDR zone. Right display shows selected key luminance levels in the image.
Traditionally the tool used for color management is the Vectorscope. This is an X-Y plot of the R-Y and B-Y color difference signals. A specific color is represented as a vector plot of hue by the angle and saturation by distance from the center. There are normally target markers for each of the colors and these (for broadcast) mark the safe saturation level of that particular hue.
The Vectorscope, an effective tool for color measurement and management in SDR, becomes harder to use in an HDR environment. As different color spaces are applied it has the effect of shifting the colors which manifests itself as a rotation in the position of the color markers. Then as the gamma curves (PQ/HLG) are applied for the signal the vector trace appears as a smaller trace roughly 50% of a trace for a Rec. 709 SDR signal. Although a LUT can be applied to the HDR signals allowing users to “zoom” into the HDR signal to make it easier to read on the screen, these color shifts and the trace compression increase complexity when trying to master content for SDR and HDR in different color spaces.
Figure 5: Vectorscope displays – top is for a Rec. 709 and bottom shows the impact of Rec. 2020 (for color bar input).
The Vectorscope is typically used for color matching between scenes, brand and logo matching (e.g. ensure labels and graphic colors for a brand are correct and consistent) and setting skin tones. However, gamut errors (image colors outside the declared or specified color space) are a key reason for having content rejected after Post Production and so it is essential that colorists and QC personnel know when this is a problem.
When working across multiple color spaces and using different gamma curves (e.g. PQ/HLG) an alternative or complementary tool for color management is the use of a CIE chart display based on the CIE 1931 color space chromaticity diagram. The CIE system characterizes colors using a luminance parameter (Y) and two color coordinates (x and y) which define a point on the chromaticity diagram. The CIE “horseshoe” shape represents all the colors the human eye can see. In most diagrams and displays triangles are added to the chart to represent the colors that can be matched by combining a set of three primary colors (Red, Green and Blue for televisions and displays). The PRISM CIE chart shows triangles for Rec. 709, DCI-P3 and Rec. 2020 and each triangle contains the gamut for those standards.
Acquisition normally uses the widest gamut possible (based on the camera gamut) since limiting the gamut at this stage will prevent expansion of the color space in Post. The captured content is then converted to the appropriate display gamut (Rec. 709, DCI-P3, Rec. 2020) in Post Production and Mastering. At this point it is important to know that the conversion has not placed colors outside the chosen display gamut (avoiding clipping and unpleasing images). Full 2020 is not practical to display on monitors today so limiting to P3 has become a practical choice with Post houses being asked to master in 2020 but limit to P3 colors.
For this application, PRISM provides a special false color display that highlights any parts of the image that are outside the P3 gamut for a Rec. 2020 encoded signal. Information on the area of the screen impacted by these issues is also provided (see Figure 7).
The Changing Workflow
In the current, and predominately HD, world there are a set of established tools for handling both exposure management and color management. These tools include a waveform for checking luminance levels, a Vectorscope for managing color and a picture display to check the appearance of the content. As demand for 4K/HDR/WCG content grows the traditional tools are beginning to hit limitations. Traditional tools need to be extended to address the new technologies and processes. It is possible to address some of the issues that arise by applying LUT’s that perform the appropriate conversion to allow the continued use of these tools. However, as Production and Post Production personnel are asked to work with multiple camera log gamma curves, multiple HDR display gamma curves, and at least 3 color spaces the existing workflow will need to evolve to use the next generation of production tools.
The new workflow will be built around a STOP waveform that works using light levels and will provide consistent measurements regardless of the camera gamma, a CIE chart that allows Post staff to easily work across multiple color spaces / gammas and check for conversion and encoding gamut errors and a false color display for different applications. False color can be used to show luminance levels, which parts of the image are in the HDR region, what area of the screen is in the HDR region, and where color gamut errors have occurred.
The Telestream PRISM provides all the traditional and future tools through a range of options that will allow customers to transition their workflow at their own pace (see Figure 8).
1 Image resolutions in this article are governed by the PRISM image capture function. These are real images from a real instrument making real measurements.
You might also like...
Program delivery to mobile devices and smart televisions has fueled the growth for internet delivery. But one of the challenges broadcasters and media content providers face is that the internet was never originally designed to stream large amounts of video…
Monitoring has always been the engineers’ best friend as it turns apparent chaos into order and helps us understand what is going on deep inside a system to deliver high-quality pictures and sound. As OTT continues to play a more p…
The current social and medical situation with lockdowns and distancing is unleashing new ideas at local TV stations. Some will become the new normal.
For many years broadcasters have been working with static systems that are difficult to change and upgrade. Although we have video and audio routing, the often-tangled mess of jackfield patch-cords is testament to how flexible broadcast systems really need to…
SDI is one of those technologies that is so well established and ubiquitous it can almost be taken for granted.