Whether we call them video monitors, displays or screens, they are everywhere. In many cases you choose a technology, size and model, plug it in, and done. But apart from size, weight, mount type, brightness, and cost, there are other subtle factors that might influence the choice of display, whether it works for the application, and how it’s adjusted.
Will This Ever Settle Down?
As I’ve mentioned in other places, the CRT (cathode ray tube) monitor was the only display in normal use from the start of television, let’s say the 1930s, to about 2000. Since then, CRTs are gone and the types of displays keep changing. All non-CRT displays, and projectors, use progressive scanning (as opposed to interlace–see https://www.svconline.com/industry/video-format-basics) but they vary widely in underlying technology and characteristics.
LCD displays use a liquid-crystal panel to selectively pass light from a rear light source—originally fluorescent backlights, now LEDs. Two downsides of this approach are that it’s hard to achieve deep blacks because the backlight is always on (hence the addition of “local dimming” and other enhancements), and they tend to lose brightness and shift color as the viewer moves away from direct center. (Side note: What marketing calls “LED TVs” are LCDs with LED backlights. What will they call real LED TVs?) Unlike LCDs, plasma and OLED displays are emissive devices; the light is actually made by the visible pixels. This was also true of the CRT, and some, including myself, would argue that this produces a richer visual quality. Plasma has come and gone, but OLED continues to produce beautiful images without the off-axis deterioration of LCD. As with CRT, it looks the same from any angle. Two downsides of OLED are higher cost and potential for “burn-in” (ghost images due to excessive still elements on screen). OLEDs usually have algorithms that dim the screen and/ or shift the image slightly to minimize burn-in potential.
Among other display technologies coming and going are Quantum Dots, NanoCells, different refresh rates, etc. Many are attempts to improve existing LCD and OLED technologies and I will defer opinions to colleagues, like Pete Putman, who follow these developments on a regular basis. By we know that the new greatest thing is quickly overtaken.
Then, of course, there’s direct view LED, now shining in videowalls, virtual sets, and signage near you! The most expensive option, but with many qualities that really matter for certain situations: High brightness, pixel density can be chosen for the application, potential for wider color gamut with more LED colors, modular construction. One downside to consider, apart from cost, is potential power usage and heat generated by the huge quantity of individual LEDs.
Meanwhile, real microLED TVs are on the horizon for consumer and AV use. These will likely have similar positives and negatives to dvLED, but the scale and pitch of the LED emitters must be incredibly small since viewing distances are short.
Display Settings and Theory
Some years ago I was helping a client shoot video in a movie theater being used for a live presentation. When the graphics were projected on screen they looked fine to the eye, but through cameras everything looked a bit purple. Here was a case where color temperature—the actual spectral content of white—was causing trouble. The cameras were adjusted (colorbalanced) for the relatively warm (redder) ambient lighting of the theater so that skin tones would look good. But the projection was cooler (more blue), and what came through the cameras was not balanced correctly.
Everyone is familiar with different shades of “white” in paper or paint. The same is true for light, as shown in the CIE chromaticity chart of visible colors (Fig.1). What we perceive as white is in the center, with specific white points defined as Standard Illuminants. Illuminant D65 is 6500K (just Kelvin, no degrees), the color temp of noon-day sunlight, and was adopted as the calibrated white point for television monitors back in the CRT days. Higher Kelvin numbers shift toward blue, lower numbers toward red. This spec is probably familiar from light bulbs.
Even the cheapest monitors usually have a few presets (such as vivid and movie) that change not only black and white levels, but also color temperature, and many have adjustments to match specific color gamut standards that define the range of viewable colors (such as Rec. 709 or sRGB). Using these may require a calibration device. (Captain Disillusion has a quick and clear Youtube video on color: https://www.youtube.com/watch?v=FTKP0Y9MVus)
Color temperature doesn’t matter much for basic KVM use, but displays used in production environments, for content creation or control rooms, should be set to known standards so that colors translate accurately between systems (what happens at the viewing end is another story). And since LCD displays tend toward the cool end, if they are used on-set, or anywhere they’ll be captured with cameras, the lighting and display color temp may need to be coordinated. Often this means making the display as warm as possible, going cooler with studio lights, or actually generating graphics that are pre-compensated for how they’ll look on camera. It may also be possible to insert a video color corrector in the signal line to the display, but these can be more expensive than the displays themselves.
Among the settings on many displays, especially consumer TVs, are functions that attempt to smooth motion in fast-moving images. These go by various names, but their effect can make almost any video look harsh and overly “flat” to some viewers. Some pros pejoratively call this the “soap opera effect,” which derives from a time when most television was actually shot on film and transferred to videotape. Then midday soap operas—daily dramas sponsored by washing product companies—started shooting directly on tape for faster production.
The issue is one of perception, not absolute quality. Over 100 years the “look” of film, due to frame rate and other characteristics, has come to be associated with higher quality, with storytelling (in the case of fiction), and general pleasantness. In fact, many would say that the motion blur inherent in the film look is a desirable effect, and motion compensation in displays attempts to “fix” this blur. Since I’ve watched the film look my whole life I generally agree with this position; for me, the motion compensation in displays looks bad. Even 60fps video looks odd to me (not “more real” as some would say) because I’m used to 30fps video.
But subjective preference isn’t the only reason to turn off all the motion comp, and anything else that messes with the frame rate or cadence of video. We really don’t want the display to change anything about the image that was not in the source content, particularly in ways that we cannot predict or control. For that reason, I also start by turning off black level compensation, automatic brightness adjustment, detail and noise processing, etc.
Artifacts
I used to worry a lot about scaling artifacts when the video resolution did not match the native resolution of the display. For example, watching 1920×1080 on a 1280×720 display. The monitor had to down-convert that video with processing that could result in visual artifacts. At this point the processing in displays is so good that it’s mostly a non-issue. And it’s nearly unavoidable because native 4K panels are replacing HD panels in most products.
However, it’s still quite common to see another artifact: contouring (Fig. 2). This can occur when the display (or the production/transmission path) does not have enough digital bits to represent subtle changes in level. The classic effect is seeing “steps” in what should be a smooth gradation of brightness or color. The difference between 8-bit and 10- bit display panels may seem inconsequential, but in binary math those last two bits take the number of possible levels from around 16 million to over a billion. Choosing a 10-bit (or even 12-bit) display depends on the application and budget, and whether contouring in the display might be mistaken for a problem with the content.
Something we often forget about video is that nothing is continuous. Our eye-brain system is fooled into making motion out of individual still pictures, and solid images from millions of individual pixels lit up at very high speed. The same is true for cameras, which capture millions of pixels every frame, at 24, 30 or 60 frames each second (or more if slo-mo is being recorded).
Where these characteristics become problematic is in situations that create conflict between what the display is actually doing, and what we want it to do. One example is the appearance of moire patterns; basically an interference artifact between two conflicting areas of detail. Camera sensors have a fixed pattern of pixels. If the camera is pointed at a subject that also has a very fine pattern of detail, such as a piece of clothing, an overlay of dark, often wavy, lines may appear at certain distances and zooms (Fig 3).
Shooting displays with video or still cameras can produce moire due to interference between the display pixels and camera pixels. This is of particular concern in virtual production applications with large LED videowalls. The camera and display can interact in ways that our eyes don’t perceive standing in the room.
Color interaction can also be an issue. Cameras don’t “see” a continuous spectrum of light because they only have sensors for red, green, and blue. When displays, or LED lights, also rely on red, green and blue (which is typically the case) the image coming through the camera may be different than what we see looking directly at the display (or the lit scene).
One Last Thing
Everyone is familiar with the assortment of input types on displays—everything from VGA to DVI to Displayport and HDMI. SDI is still a necessity on broadcast and production monitors, and a few still have analog composite. New connectivity options are still coming and going, so that needs another article! But if you work with “cable” (CATV) distribution or off-air antenna systems, or need to specify monitors that can handle those inputs, be aware of a few potential hazards.
First is the difference between old analog CATV and digital CATV. Both are modulated RF signals, with “channel” numbers assigned to various frequencies, but the content of the modulated signals is not compatible. If the source system is analog, does the monitor support it? For digital CATV (and satellite services) there are several modulation schemes, of which QAM (quadrature amplitude modulation) is commonly used in the US. Clear-QAM means there is no encryption in use, as was sometimes the case with cable or satellite set-top boxes.
The same goes for traditional broadcast (over-the-air) TV. Digital broadcast, DTV, is alive and well, but uses different channel assignments than QAM. Some televisions may still have channel assignments for analog broadcast, but this was discontinued in the US in 2009.
On top of this, broadcast television is in the midst of another change, moving from the current DTV scheme to a better one that is, naturally, incompatible. The new system, known as ATSC 3.0, has huge improvements and is starting to appear in consumer TVs, but it’s a slow process that is not being pushed by the FCC, as it was in 2009. Nor is ATSC3.0 on the air yet across the whole country. But if a TV is intended to last quite a few years, finding one that supports ATSC3.0 might be worthwhile (and make you look like Nostradamus to your client).
And don’t forget to scan for channels! This is standard practice when setting up a new TV, so that the tuner can find all the available channels it is capable of receiving. When changes are made to the system, rescanning channels can save a lot of frustration.
