Use this URL to cite or link to this record in EThOS:
Title: An intelligent multiplexer architecture for thin route communications
Author: Shah, Asadullah
ISNI:       0000 0001 3395 4060
Awarding Body: University of Surrey
Current Institution: University of Surrey
Date of Award: 1997
Availability of Full Text:
Access from EThOS:
Access from Institution:
The main aim of the work reported in this thesis was to investigate and develop ways of increasing the throughput of a multiplexer for voice calls by introducing some intelligence into it so that the voice frames with least subjective significance are discarded. The discarded frames of the corresponding users are faithfully reconstructed at the receiver. The proposed intelligent multiplexer simulations are based on total five frame discarding techniques; three criterion and two non criterion based techniques. It is concluded that frame discarding techniques and Lost Frame Reconstruction (LFR) have a great potential of bandwidth savings. Generally criterion based frame discarding techniques perform better than none criterion based schemes. Digital Speech Interpolation (DSI) advantage) is achieved for monologue speech at 3% frame discarding rate, i.e., 1.33. For dialogue speech a DSI advantage of 2.0, can be achieved at 3% frame loss rate as indicated by analysis. The bandwidth efficiency can be increased by designing LFR schemes that can cater for changing characteristics of the speech signal, such as transitional regions of speech. The additional increase can also be achieved by investigating frame discarding techniques based on the perceptual importance of the speech signal. This thesis consists of 7 chapters. In the first 2 chapters the introduction of thesis and the background of Speech Interpolation (SI), systems employing Time Assignment Speech Interpolation (TASI), Digital Speech Interpolation (DSI) and LFR techniques are described. In the third chapter, the low bit rate speech coding is covered. In chapter 4 statistical measures of speech temporal parameters, such as talkspurts and silences are presented. These measures are then used for the simulation of talkspurts and silences by 2-state Markov model to represent speech sources. Also the tools and techniques developed for simulation are presented in chapter 4. The chapters 5-7 of the thesis consists of simulation results and suggestions for further work.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: Voice calls; Intelligence; LFR