March 2011
Digital video on the desktop – part 2
In part 1 of this series we looked at the early days of digital video. In this second instalment, we’ll discuss the nature of video and some important characteristics.
An eye on the world
Video is Latin for “I see”. Every video camera sees the world through a lens and captures the image on an electronic sensor. A modern sensor is an array of millions of picture elements (pixels) and is commonly made in one of two chip design technologies ? CCD or CMOS –each with its advantages and drawbacks.
Most high quality (CCD) cameras have three sensors: the incoming image first passes through a prism and is split into a red, green and blue (RGB) image ? with one sensor per primary colour. This results in more accurate colour processing than in a single-sensor camera design.
The image, captured by the sensor(s) is electronically optimized and converted into a standardised video signal. For many decades, this signal was analogue and very sensitive to degradation, interference, tape imperfections, etc. Today, video is mostly distributed in a digital format, which is much more robust.
This video signal can be transmitted through the atmosphere or via satellites, sent to the homes via cable distribution, streamed over the internet, displayed on CRT, LCD, OLED or plasma screens, or stored on tape, disc or Flash memory.
Standardised… or is it?
A video signal has many characteristics, and although it is relatively well standardised, there are variations that can lead to incompatibility between equipment and digital files. Especially when editing video footage from different sources, it’s important to know the characteristics, because not everything mixes and matches easily.
-
TV system & frame rate
In early TV technology, many of the frequencies inside the system were based on the mains frequency ? 50 Hz in Europe, 60 Hz in the USA and Japan. This resulted in a frame rate (number of full images shown each second) of 25 f/s in Europe, 30 f/s in the USA. When colour TV was invented, this regional gap was widened by introducing incompatible colour systems as well: PAL in Europe, NTSC in the USA, and SECAM in France and Russia.
The advent of digital media like DVD and digital displays like LCD and plasma could have been the opportunity to introduce one global standard, but political and commercial interests stood in the way. So even today, TV standards still obstruct an easy exchange of video material. When standards do not match, a hardware device (standards converter) or a software conversion is needed to adjust the frame rate and colour system.
-
Interlaced/progressive scan
Before flat screens where introduced, there was a CRT (cathode ray tube) in every TV and monitor. The principle of this device is to shoot a beam of electrons towards a fluorescent phosphor layer at the front of the screen. By controlling this beam appropriately, a moving dot “writes” the TV picture from left to right, line by line. Because the phosphor has some persistence and the beam travels quickly, our eyes see a complete picture instead of a moving dot.
To conserve bandwidth and reduce flicker, it was decided to write the picture frame in two halves (fields). First, all the odd lines are written, then the dot returns to the top left and writes all the even lines. This process is called interlacing. It made a lot of sense in the analogue world, but creates many problems in the modern digital world.
LCD and OLED screens are progressive by nature: the frame is written line by line, and not as two consecutive fields. Interlaced video therefore has to be “de-interlaced” before it can be displayed. If this is not done well, visible artefacts are created. In an interlaced frame, a fast moving object will be in a different position in the odd and even field, and poor de-interlacing will result in a “combing” effect of two superimposed objects.
Many digital cameras can be switched between interlaced or progressive recording. It is always best to keep the whole production chain in the same format: recording interlaced video when it will eventually be broadcast, or recording progressive when it will end up on the web.
-
SD/HD resolution
Generally, standard definition (SD) refers to a picture with a resolution of 720x480 in the NTSC world, and 720x576 in the PAL world.
High definition (HD) is standardized as having a resolution of 1280x720 pixels (720p) or 1920x1080 pixels (1080i/1080p). The latter is sometimes called TrueHD. The “p” or “i” refers to progressive or interlaced, and the frame or field rate is often added. So a 1080i50 video recording has a resolution of 1920x1080 and consists of 50 interlaced fields per second. A 720p25 recording has a resolution of 1280x720 pixels and shows 25 progressive frames per second.
-
Aspect ratio
Most older TV sets and many current computer monitors have an aspect ratio of 4:3, meaning that the width of the screen is approximately 1.33 times the height.
When DVD-video came onto the market, it was a boost to the adoption of “widescreen” TV’s, with an aspect ratio of 16:9, or roughly 1.78, closer to the average ratio used in feature film production.
When HDTV was introduced, the 16:9 aspect ratio became an integral part of the standard, so there is no 4:3 (standard) HD format.
When showing a 4:3 video on a 16:9 screen, the image is either stretched to fill the entire area (leading to distortion) or “pillar boxed” (a vertical black bar at each side).
When showing a widescreen image on a 4:3 screen, there are three options: cropping (cutting away the excess part of the image at the sides), pan-and-scan (cutting away parts, while dynamically shifting the image to lose as little as possible of the information) or letterboxing (a black bar at the top and bottom of the screen).
It’s important to remember that every video recording is defined by a combination of these parameters, and that a mismatch between different footage will require some form of adjustment that can take a toll on the quality of the end result. In an ideal world, all the material that needs to be edited together to create the final video is identical in terms of these parameters.
Also, when displaying digital files on a screen, it’s important to verify ? and correct if necessary ? that the source and display are in the same mode. Anyone visiting a trade fair will have seen stretched videos, black bars where they shouldn’t be, or clear mismatches in resolution or aspect ratio… Mistakes that degrade the company image instead of boosting it!
In part 3, we will look at methods of compressing the video information. Without compression, even today most computers, storage media and distribution channels would not be able to efficiently cope with the massive file size of uncompressed video.