Exposure Tools for Digital Video, Part 1
On the formal level, exposure is the amount of light that reaches the image-capturing medium. It is determined by the sensitivity of the medium, the illuminance at the image plane and the exposure time. Setting exposure is one of the most important artistic decisions and probably the most important technical decision for a shot because it governs the distribution of tones in the image. With film, exposure choice requires good knowledge of the available exposure latitude and the dynamic range distribution of the emulsion. Light levels are measured with a light meter and the exposure can be further adjusted for optimal tonality based on properties like scene mood, vision or simply personal preferences.
The instant feedback of digital video is its greatest advantage over film. On a shoot, this feedback relates to exposure more than anything else. Exposure errors can be fixed immediately (no need for rushes or dailies) and exposure choice can be feedback informed and guided. A calibrated field monitor can be of great help, but it needs both good viewing conditions and excellent rendering capabilities in order to offer a decent representation of the recorded video. Exposure is often hard to judge based on monitor feedback alone. But a number of tools can process the captured signal and display it in various ways to aid the exposure decision. This article is an overview of the most popular of these exposure tools, and their strengths and weaknesses.
Most of these exposure tools take as their input the image data after the transfer curve is applied, so they directly represent the recorded video. In some cases – mostly with higher-end equipment – the display data may differ from the actual recorded data. This is usually done when log space video is recorded. The flat image is not very useful for monitoring, so it may have some LUT applied for preview and exposure purposes. This LUT will generally simulate to some degree the final look of the video.
These exposure tools are available through the camera firmware, on field or studio monitors, or through specialized software fed with the camera signal. In any case, they need a screen to display info on, be it a camera LCD, an external display, or just a computer screen. They are at their most useful when combined with good knowledge of the camera transfer curves (gammas).
The histogram is familiar to most photographers and Photoshop users. The histogram is a simple chart diagram. It displays the distribution of image pixels over the possible range of the video signal domain. The more pixels falling into a particular value bin, the higher its column in the histogram. The histogram doesn’t really contain any information about what part of the image falls where. But it is useful for quickly identifying the presence of clipped highlights or crushed blacks, as well as determining the overall key of the acquired image. A histogram skewed to the right suggests a high key image, a skew to the left suggests low key. Histograms usually come in two modes – luminance (brightness) and RGB – which are often switchable according to the preferences of the camera op. RGB histograms are actually compound histograms featuring a histogram for each of the three color channels, either overlaid or in parade. They help identify color channel over-saturation or color balance issues.
The main strength of the histogram is that its graphic representation can be scaled down without significant information losses because it is the overall shape that delivers meaningful exposure information. This allows to get it overlaid on the actual image, tucked somewhere to the side.
Zebras are a video camera invention and one of the first exposure tools to appear in video cameras. The concept is simple. Colored stripes (hence the name) are overlaid on the actual image, marking the position in the image of any clipped whites or crushed blacks, or both. Unlike the histogram, zebras localize and visualize the problematic areas so it is easier to judge if the out-of-range pixels are important or can be painlessly sacrificed. Depending on the implementation, zebras may have their warning triggering clip/crush points customizable; they may blink or feature other improvements.
The waveform monitor (WFM) is one of the most useful tools in video production. Both during shooting and in post. Traditionally, the WFM is a monitoring device that displays the level of a video signal over time. In analog video this means voltage. In the digital world the WFM concept has been expanded to handle discrete values. In a sense, the digital waveform monitor is an extension of the histogram. But instead of binning pixels from the entire image, the waveform monitor bins pixels from each image column separately. These bins are then represented by pixels in a column of the output of the WFM. On the vertical axis of the waveform monitor are the values defining the bins. These can be the actual video space values, values scaled from 0 to 100%, or IRE units. The horizontal axis runs over the columns of the input camera image. The intensity of the output pixel corresponds to the saturation of its respective bin. The more pixels from the corresponding input column falling into a value bin, the brighter the bin’s pixel on the WFM.
The waveform monitor can do pretty much any job the histogram can handle, plus more. Because each image column is processed separately and there is one-to-one correspondence between image columns and WFM columns (assuming a non-scaled WFM image), any offending areas are easily horizontally localized on the waveform monitor, and then in the actual camera image. This is especially useful for faces, grey cards or just large suspicious areas. Moreover, important values on the vertical axis can be signified with lines parallel to the horizontal axis for quick judgment of the image tonal distribution. This is a powerful tool when coupled with good knowledge of the transfer curve used.
Waveform monitors can be used with a variety of input color spaces. Luminance is generally enough for exposure and is one of the channels of the typical YCbCr digital video stream. RGB, YRGB and others can be used for variety of tasks, most notably for color correction. These are either displayed in parade, or overlaid in different colors.
The exposure tools overview continues with false color, spot displays and light meters.