A study of enhanced PAL television, including a new coding technique, and new methods for analysing and appraising coding techniques

Abstract

The introduction of colour television in the 1950's was an impressive achievement producing a full colour picture on new receivers whilst maintaining substantially complete compatibility with monochrome receivers. This was achieved by placing two colour difference signals on top of the existing monochrome (luminance) signal. The disadvantage of the technique was the lack of full separability of the signals in the receiver, resulting in signal cross-effects and some loss of picture quality. Although acceptable at the time of introduction improving source quality (due to improved cameras, digital sources, and improved recording) coupled with increasing quality expectations has caused these artefacts to become more noticeable and less acceptable. Consequently considerable work in the last twenty five years has concentrated on finding improved encoding and decoding techniques for creating compatible PAL signals. Unfortunately until very recently with the introduction of PAL plus no satisfactory solution had been agreed upon. Work in the field has been made difficult through the complexities of the multi-dimensional PAL signal, the lack of common framework, and the difficulty of assessment and comparison. This thesis provides new work in several areas in the field of enhanced-PAL coding including: i) the development of a complete multi-dimensional representation for a television coding system, ii) the description of a new technique for determining the spatio-temporal characteristics of coding systems, iii) the description and application of a new approach to objective quality assessment, and iv) the detailed study of a new improved PAL coding technique. This thesis also i) provides comprehensive background on the multi-dimensional interpretation of the PAL signal and television coding techniques, ii) develops a model for the human visual system chrominance threshold response, iii) places all current enhanced PAL coding techniques into a common theoretical framework, and iv) provides a detailed mathematical analysis of WC-PAL. Work on all these areas was greatly assisted through the development by the author of a multi-dimensional graphical interface (PHIGSdraw) and image processing system (TVPROC).