Monitors should tell it like it is, warts and all.
Assuring broadcast quality demands high-performance audio and video monitoring. Loudspeakers do not present a problem; specialist monitoring speakers have been around for decades. But video monitoring is a different matter as the broadcast engineering sector represents a very small market for displays. With annual volumes measured in thousands, it is dwarfed by the consumer and office sectors. It is just not profitable for commodity display manufacturers to cater to such a niche sector.
The situation reached a minor crisis with the sudden demise of the CRT. The tube had evolved since the developments of Zworykin and others in the early 20th century. The zenith was reached with the broadcast-grade Sony Trinitron, which became the favored monitor for video engineers. Unfortunately for users, the CRT ceased manufacture before suitable alternatives came on stream.
The broadcast sector is constrained to use popular display technologies, including LCD and plasma. There have been a couple of years of getting by with existing monitors or buying rather unsuitable consumer display panels. The latter have been adapted for professional use with HD-SDI connectors, color calibration systems, and high-quality scalars and deinterlace. Meanwhile, a handful of manufacturers has been developing replacement technologies that meet the needs of video engineers.
For most broadcast applications, the LCD provides a good solution. It is low-cost, has a shallow depth and is lightweight. For monitor stacks, where the LCD is used to check composition and content, it fits the requirement. But for critical monitoring, the commodity broadcast LCD monitor is lacking.
Shading and grading
What do broadcast engineers look for in a monitor? A monitor should tell it like it is, warts and all. Consumer displays are designed with high brightness, for daylight viewing. A consequence of high brightness is that flicker is more visible, so they may use frame doubling. They also use all manner of proprietary picture “enhancement” processing, largely for marketing reasons.
A broadcast-grade monitor should not mask defects with error correction or enhance the picture with any proprietary processing; it is a monitoring instrument, not a viewing screen.
For shading and grading, color accuracy and consistency are key. Video engineers or color graders will have a waveform monitor and vectorscope to monitor the signal, but most importantly, they will have a carefully calibrated picture monitor.
For color graders, the monitor is their primary tool for adjusting the color correction controls. For video engineers, the monitor is essential to match all the cameras in a studio or truck to the same color reproduction.
Calibration usually involves a color measuring probe that can construct a look-up table for each monitor to give a consistent color reproduction. Consistency is vital. Does the picture look the same in the truck as back at the studio? Does the picture look the same in finishing as it did in the grading suite? Sensible quality control needs consistency in the monitoring device to avoid arguments between the director and the post facility.
A big problem with early LCDs was viewing angle. The ideal display would have an isotropic distribution of light. This is achieved by the CRT and PDP, but not so for LCDs. Early LCDs not only had luminance fall off, but the colors change off-axis. The in-plane switching (IPS) technique is one way that panel performance can be improved, although IPS switching speeds are lower than some other LCD designs.
Current broadcast LCDs give a reasonable viewing angle, and by proper room design it can be ensured that critical viewing is not performed well off the display axis.
The television system is based on temporal sampling of the scene. The refresh rates are chosen to give realistic portrayal of motion and to avoid flicker. Each type of display has its own dynamic characteristics that make it difficult to achieve an absolute match for motion portrayal across different technologies.
The LCD motion artifacts, stemming from the sample and hold nature of its operation, are different from the scanning beam of the CRT. In the days of the CRT, achieving consistency in dynamic rendering was easier than today as all points in the chain were using a CRT to view. The different display technologies used today — LCD, plasma (PDP), DLP, SXRD and OLED — each have characteristics, especially dynamic, that make it difficult to absolutely match the way they display video.
The projection devices — SXRD and DLP — find more application in the digital cinema sector, where large displays are the norm. For broadcast, the requirement for critical viewing is for the smaller direct-view displays with typical diagonals of 24in. This rules out the PDP, which has a cell size more suited to displays over 50in diagonal. At the pixel pitch required for a 24in display, the primary choices are LCD and recently the OLED.
Generally, the larger displays found around a TV station do not need the carefully controlled reproduction of a monitor, and either PDP or LCD are appropriate choices.
The LCD displays changing levels at one pixel site as a sample and hold. The OLED can operate in a similar fashion. Contrast this with the short flash of the scanning in a CRT. The PDP cell is either on or off, so different levels of luminance are created by pulse width modulation.
Each of these methods gives a different portrayal of motion. The LCD “look” can be modified by flashing the backlight, and a similar result can be achieved with the OLED by switching a cell on and off during the frame. Such techniques are a compromise. The short flash of the CRT appears to the eye as flicker, more so for 50Hz scanning than 60Hz. This frequency is around the point at which the brain fuses flicker to a continuous image. The sample and hold effect adds a stepping effect to motion, which appears as judder.
Ultimately, the brain adapts to any of these display mechanisms, but it has been shown in tests that having a mix of displays in a control room causes viewing fatigue as the eye changes from one display to another during the working day.
The optimum is to use one technology for all the displays. So if there is a large monitor for clients to view, or a monitor stack has varying sizes, they should all be of one type. This rules out PDP monitors, as they are produced only in large sizes. Currently, OLEDs can only be economically manufactured in sizes up to 25in. So, LCD is the current choice for a room with a wide mix of sizes and where the avoidance of visual fatigue is a primary factor.
Monitors have to cope with a range of inputs:
- SD, PAL and NTSC;
- HD, 720 and 1080 lines; and
- Progressive and interlace.
The SD standards have non-square pixels. This is not a problem with the analog scanning of the CRT, but they need scaling to match the square pixels of an LCD or OLED. The SD standards also use different chromaticity coordinates specified in 1974 CCIR report 624, commonly referred to as EBU phosphors (for PAL) and SMPTE C phosphors (RP145). HD television is now standardized in the ITU-R BT.709 recommendation. The other standard in use today is SMPTE RP-231 for digital cinema, which has a slightly larger gamut than 709.
All these various video input formats must be scaled and the colorimetry corrected before presenting to the panel drivers.
LCD panels are inherently progressive, so interlaced video must be converted. This can be through a deinterlacing filter, the most common way, or through the artifice of emulating an interlace by displaying fields as frames, with black lines alternating with the picture lines. The latter will give a good impression of interlace complete with artifacts like interline twitter. However, there are drawbacks. There is a long processing delay, which means the monitor picture is lagging behind conventional monitors and the video signal itself. This will give the TD and sound mixer sync problems.
The CRT was an emissive display; the phosphors emitted light when excited by the scanning electron beam. In contrast, the LCD is a light valve. Each picture element controls how much light passes from the rear illumination. This difference accounts for some of the characteristics of each, notably the rendition of black. With the light valve, there can be slight leakage, preventing a proper black, or absence of illumination, being achieved. The OLED is also an emissive display, so it can reproduce black in the same way as the old CRT.
There is a choice of backlight for the LCD, with the cold cathode being the original. LEDs are now an alternative and have proved popular in the consumer market, where a display should be bright and highly colored; consistency of color is not a requirement. White LEDs can be used for backlights, but separate RGB LEDS allow better control over colorimetry.
Conventional LCD panels have a resolution of 8 bits per color, less in laptops at 6 bits. This can be masked by technologies like dither to improve apparent color resolution, but a professional broadcast monitor needs a minimum panel resolution of 10 bits per color. This will avoid any possibility of quantizing effects on smooth gradients, and matches the bit-depth of the SDI signal.
Ideally, the monitor's native gamut should match the application — wide for film and 709 for television. Constraining a film display panel to 709 space using a look-up table in the driver will waste potential bit-depth in the panel.
All displays age. It can cause a simple loss of brightness, or more unwanted effects like color shifts or burn-in. The LED backlights used with LCD panels exhibit color shifts as they age, and they require feedback systems to maintain constant color. CCFL back lights also age, but just lose brightness, without any significant color shift.
The different display technologies have a usable life between 30,000 and 60,000 hours. For the broadcast professional, it is essential that during the life of the display, it can achieve the necessary peak white luminance of 100cd/m2, that there is no color shift, and that there is no burn-in of static images like logos or VITC.
THE PRAGMATIC APPROACH
No display technology is perfect. The CRT had drawbacks. It was more suited to interlace display and was not available in larger screen diagonals. It could not achieve the scanning speeds necessary for 1080p/50 operation. The LCD has its drawbacks, notably reproduction of black and viewing angle, and OLEDs suffer from screen burn and aging.
In 2012, the LCD has the widest application for broadcast operations, and most of the artifacts have been controlled. In five years' time, there may well be another technology that has better performance. But the performance must also be balanced against cost.
Is what you see on the monitor a true representation of the video signal, very close or just an approximation? Since the LCD has replaced the CRT, answering that question requires careful viewing tests of a well-aligned display.
For the portrayal of motion, ultimately the display is limited by the fact that television uses temporal sampling. So, at best, motion is an illusion. Video monitors are an engineering compromise between what is needed and what can be manufactured at an affordable price.