next >

Characteristics of Digital Video

Digital video is not like normal analogue video used by everyday televisions.1 To understand how digital video works it is best to think of it as a sequence of non-interlaced images, each of which is a two-dimensional frame of picture elements or pixels. Present day analogue television systems such as the National Television Standards Committee (NTSC), used in North America and Japan and Phase Alternate Line (PAL), used in western Europe, employ line interlacing. Systems that use line interlacing alternately scan odd and even lines of the video, which can produce artefacts when analogue video is digitized. This issue complicates any discussion of digital video and the compression process, and so is best left aside for now. Associated with each pixel are two values, luminance and chrominance. The luminance is a value proportional to the pixel’s intensity. The chrominance is a value that represents the colour of the pixel and there are a number of representations to choose from. Any colour can be synthesised by an appropriate mixture of three properly chosen primary colours. Red, Green and Blue (RGB) are usually chosen for the primary colours. Another system, known as YUV, allows colour to be approximated using only two variables where Y represents the luminance, and U and V the chrominance. When an analogue signal is digitised it is quantized. Quantization is the process by which a continuous range of values from an input signal is divided into non-overlapping discrete ranges and each range assigned a unique symbol. A digitized monochrome photograph might, for example, contain only 256 different kinds of pixel. Such an image would be said to have a pixel depth of 8 bits. A higher quality image might be quantized allowing 24 bits per pixel.

Digital video can be characterised by a few variables:
Frame rate:
The number of frames displayed per second. The illusion of motion can be experienced at frame rates as low as 12 frames per second [Cook90], but modern cinema uses 24 frames per second, and PAL television 25 frames per second.
Frame dimensions:
The width and height of the image expressed in the number of pixels. Digital video comparable to television requires dimensions of around 640 x 480 pixels.
Pixel depth:
The number of bits per pixel. In some cases it might be possible separate the bits dedicated to luminance from those used for chrominance. In others all the bits might be used to reference one of a range of colours from a known palette.

The table below illustrates possible values of these parameters for typical applications of digital video.

Typical parameters for various applications of digital video
ApplicationFrame rateDimensionsPixel Depth
Multimedia15320 x 24016
Entertainment TV25640 x 48016
Surveillance5640 x 48012
Video Telephony10320 x 24012
HDTV251920 x 108024

[Why do we need compression?]

1. This document was first published in 1996.

© Colin E. Manning 1996