Title :
A CELP-based speech-model process controlled by trained Markov chains
Author_Institution :
Inst. for Network & Syst. Theory, Kiel Univ., Germany
fDate :
29 Jun-4 Jul 1997
Abstract :
A novel speech-model process (SMP), which is based on a modified CELP-structure, is proposed as a test signal for speech processing applications. The applied CELP-structure and the generation procedure for this SMP are presented. The controlling part of this SMP involves several kinds of Markov chains to adapt the time-varying properties of the process to those of natural speech. The measurement results have shown that the characteristics of “Markov-CELP”-SMP agree well with those of natural speech. This concept renders it possible to produce a test signal with characteristics of several languages and/or to generate a test signal which includes required special features if needed, like male/female-type speech or weak/strong variation of pitch frequency
Keywords :
Markov processes; adaptive systems; decoding; linear predictive coding; speech coding; speech processing; CELP coder/decoder; CELP-based speech-model process; Markov-CELP; adaptive transmission systems; generation procedure; languages; male/female-type speech; measurement results; modified CELP-structure; natural speech; speech coding; speech processing applications; test signal; time-varying properties; trained Markov chains; weak/strong pitch frequency variation; Character generation; Decoding; Electronic mail; Natural languages; Process control; Signal generators; Signal processing; Speech processing; Stochastic processes; System testing;
Conference_Titel :
Information Theory. 1997. Proceedings., 1997 IEEE International Symposium on
Conference_Location :
Ulm
Print_ISBN :
0-7803-3956-8
DOI :
10.1109/ISIT.1997.613383
