ABSTRArnThis paper describes the Voice Activity Detector standardised by CE€T for use in the Pan-European Digital Cellular Mobile Telephone Service. The difficulties of detecting speech in a noisy environment are discussed and the performance tests carried out to validate the design are described.

No abstract

An important mode of operation for mobile radio systems is known as discontinuous transmission (DTX). In this mode the transmitter is switched on when a user begins to talk and switched off again when he stops.There are two reasons for using DTX: to increase the number of radio channels (by 100% or more) which can be accommodated in a given bandwidth and to reduce the power drain in hand portable equipment.Since system operators rely on the DTX working efficiently (to avoid excessive co-channel interference) it is also necessary to find ways of specifying and measuring the performance.This paper describes the DTX system which has been standardized by the CEPT for the Pan-European digital mobile radio system.

No abstract

Network technology typically contains complex and nonlinear processes, such as low bit rate codecs for data reduction. This nonlinearity means that the perceived performance cannot be predicted with conventional engineering performance metrics. Reliable objective assessment of perceived performance is required for optimal design, commissioning and monitoring of quality. This paper looks at how BT can measure, optimise and control the perceived performance of the service that is provided.The final receiver of a signal is the human user -by modelling the senses of the human it is possible to predict whether an error would be perceptible. An auditory perceptual model has been developed and used to predict the subjective performance of a variety of nonlinear systems. Applications for the auditory model include design optimisation, as a measurement tool and as part of a nonlinear network planning tool.This paper focuses specifically on the auditory system, and how perceptual modelling is providing the basis for improved quality of voice service. However, as the global landscape adapts over the next decade, with multimedia, personal interactivity, and the concept of 'virtuality' then all of the human senses need to be considered together. Future work is planned to include an elementary visual model and the combination of sensory models for the assessment of multimedia products and services.Purely empirical approaches have been described [4][5][6], where a variety of objective measures such as coherence and cepstral distance are used to train a neural network. However, the performance of the analysis with an unknown