Abstract
In recent years the rapid growth in the number of telephone users world-wide, particularly in the area of mobile communications, has highlighted the Quality of Service offered by the service providers as an increasingly important issue. The perceived level of service and the accompanying cost offered to the users will often be a major issue in deciding on which operator to choose for a particular application. Owing to the advantages that digital transmission offers over the original analogue systems, it has become the dominant current/future technology in both the PSTN for trunk telephony, and the emerging PCN and DMR standards, owing to its robustness to channel degradations, signal regeneration, cost, flexibility, increased capacity, switching etc. However, the adoption of digital technology for applications such as satellite and land-mobile systems will require the original speech information to be compressed in order to allow for more available channels within the limited bandwidth allocation. The compressed data will suffer from various levels of distortion caused by the deep fades and multipath signals that are experienced during mobile connections. The effect of these distortions can be reduced by the application of a suitably optimised channel coding and frame substitution strategy. This thesis will present a detailed description of the development and real-time implementation of an 11.4 kbps combined speech and channel coding scheme, that was specifically designed to meet the stringent system constraints of a DMR application. The trade-offs required to allow the entire system to be realised in real-time are discussed, which demonstrated how important it was to include complexity issues in the initial design of any future standards. It also highlights one of the major degradations that may occur when the current and future digital standards are operated in tandem. The problem of tandeming digital voice codecs can cause increased distortions that reduce the overall link quality, owing to limitations in the design of the adaptive predictors which are essential to enable the reduction in coded bit rate. This will have a direct bearing on the quality of service offered by a network operator, as transparent interoperability of the different systems will be expected by the users. The development of a unique system designed to allow the network gain to be measured for a digital mobile to PSTN connection is also discussed. This type of test equipment will be required for future systems as the traditional techniques involving pure tones are no longer applicable. This is because of the assumptions required in the models used by the voice codecs to enable the high levels of compression, whilst still providing the necessary perceptual quality.