A method for encoding the spectral characteristics of speech, at rates below 180 bit/s, using hierarchical temporal decomposition (HTD) is proposed. A set of the log-area-ratio (LAR) parameters, extracted from a given block of speech, are approximated through Gaussian interpolation between the most-steady frames detected by the HTD. This results in a smaller set of parameters which are encoded using vector quantisation. It is shown that the same spectral distortion is obtained with the new coder at a rate of 180 bit/s as that using a scalar quantisation, TD-based coder, at 600 bit/s.
References
-
-
1)
-
Ghaemmaghami, S., Deriche, M., Sridharan, S.: `Hierarchical temporal decomposition: A novel approach to efficient compressionof spectral characteristics of speech', Int. Conf. Spoken Language Proc. ICSLP '98, 1998, 6, p. 2567–2570.
-
2)
-
Y.M. Cheng ,
D. O'Shaughnessy
.
Short-term temporal decomposition and its properties for speech compression.
IEEE, Trans.
,
6 ,
1282 -
1290
-
3)
-
Atal, B.S.: `Efficient coding of LPC parameters by temporal decomposition', , Proc. ICASSP 83, 1983, p. 81–84.
http://iet.metastore.ingenta.com/content/journals/10.1049/el_19990316
Related content
content/journals/10.1049/el_19990316
pub_keyword,iet_inspecKeyword,pub_concept
6
6