DocumentCode :
1956418
Title :
Orthogonal analysis of multipulse excitation [of speech]
Author :
Paarmann, Larry D. ; Magner, Christopher A. ; McCarthy, Brian M.
Author_Institution :
Dept. of Electr. & Comput. Eng., Drexel Univ., Philadelphia, PA, USA
fYear :
1989
fDate :
14-16 Aug 1989
Firstpage :
97
Abstract :
A method of multipulse analysis that is sequentially optimal while maintaining high computational efficiency is described. This method exhibits very similar efficiency to standard multipulse analysis with post optimization. This results in higher segmental SNRs for a given bit rate, compared to standard multipulse analysis. Improvement occurs at all pulse densities, but is more significant at higher pulse densities where the pulses become more correlated. This method is based on a nonlinear time-series modeling technique developed by M.J. Korenberg (1985, 1987)
Keywords :
series (mathematics); speech analysis and processing; speech synthesis; bit rate; computational efficiency; multipulse excitation; nonlinear time-series modeling technique; orthogonal analysis; pulse densities; segmental SNRs; sequentially optimal; Algorithm design and analysis; Bit rate; Computational efficiency; Linear predictive coding; Mobile computing; Nonlinear filters; Performance analysis; Signal analysis; Speech analysis; Speech synthesis;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Circuits and Systems, 1989., Proceedings of the 32nd Midwest Symposium on
Conference_Location :
Champaign, IL
Type :
conf
DOI :
10.1109/MWSCAS.1989.101803
Filename :
101803
Link To Document :
بازگشت