Abstract: Input speech is coded in an encoder (11), the coded speech is decoded in a decoder (12), compensatory speech which compensates the speech of the current frame is generated in a compensatory speech generating part (20) by using past decoded speech, the quality of the compensatory speech is evaluated by using the input speech and the compensatory speech and a duplication level is generated the value of which increases incrementally with decreasing speech quality evaluation value in a speech quality evaluating part (40), and as many identical packets as the number specified by the duplication level is generated for the coded speech in a packet generating part (15), and the packets are transmitted, thereby reducing the possibility that packet loss will occur at the receiving end.

Abstract: In coding and decoding an acoustic parameter, a weighted vector is generated by multiplying a code vector output in a past frame and a code vector selected in a present frame by weighting factors respectively selected from a factor code book and adding the products to each other.

Abstract: At the speech encoding end, upon generation of an fixed excitation vector, the shape of an excitation vector output from pulse excitation codebook 301 is identified in pulse excitation vector shape identifier 302, a dispersion vector used for excitation vectors of the shape is output from dispersion vector storage 304, and, in dispersion vector convolution processor 303, dispersion vector convolution processing of the excitation vector is performed. In particular, when a pulse excitation vector having a specific shape of high frequency of use is output from pulse excitation codebook 301, pulse excitation vector shape identifier 302 controls dispersion vector storage 304 in such a way that an additional dispersion vector prepared dedicated to the pulse excitation vector is output. By this means, it is possible to provide a technology that improves the quality of decoded speech and that decodes speech more natural and audible to the user.

Abstract: First determiner 121 provisionally determines whether a current processing unit is a stationary noise region based on a determination result on stationary characteristics of a decoded signal. Based on the provisional determination result and a determination result on periodicity of the decoded signal, second determiner 124 further determines whether the current processing unit is a stationary noise region, whereby a decoded signal including a stationary speech signal such as a stationary vowel is distinguished from a stationary noise, and thus the stationary noise region is detected accurately.

Abstract: In coding and decoding an acoustic parameter, a weighted vector is generated by multiplying a code vector output in a past frame and a code vector selected in a present frame by weighting factors respectively selected from a factor code book and adding the products to each other.

Abstract: The present invention carries out pre-selection on many LPC codevectors stored in an LSF codebook 101 using a weighted Euclidean distortion as a measure and carries out a full-code selection on the LPC codevectors left after the pre-selection using an amount of distortion in a spectral space as a measure. This makes it possible to improve the quantization performance of the LPC parameter vector quantizer and improve the quality of synthesized speech of the speech coder/decoder.