Excitation signal encoding method and device capable of encoding with high quality
First Claim
1. An excitation signal encoding method comprising the steps of:
- dividing a speech signal into a plurality of frames;
carrying out a linear predictive analysis at every one of said plurality of frames to produce spectrum parameters;
dividing each of said plurality of frames into a plurality of subframes each of which has a subframe length;
calculating a weighted speech vector by the use of said spectrum parameters and said plurality of subframes; and
generating a new excitation signal by the use of an adaptive code book comprising a plurality of adaptive code vectors and a sound source code book comprising a plurality of sound source code vectors, said generating step being carried out in a predetermined period,wherein, when said predetermined period is shorter than said subframe length, the new excitation signal is generated by the use of an adaptive code vector that is calculated by using the excitation signal generated in the former period and a sound source code vector of the present period.
1 Assignment
0 Petitions
Accused Products
Abstract
In an excitation signal encoding method comprising the steps of, dividing a speech signal into a plurality of frames, dividing each of the plurality of frames into a plurality of subframes each of which has a subframe length, and generating a new excitation signal by the use of an adaptive code book comprising a plurality of adaptive code vectors and a sound source code book comprising a plurality of sound source code vectors, the generating step is carried out in a predetermined period when the predetermined period is shorter than the subframe length. The generating step is carried out by the use of the adaptive code vector that is calculated using the excitation signal generated in the former period and by the use of the sound source code vector of the present period.
14 Citations
9 Claims
-
1. An excitation signal encoding method comprising the steps of:
-
dividing a speech signal into a plurality of frames; carrying out a linear predictive analysis at every one of said plurality of frames to produce spectrum parameters; dividing each of said plurality of frames into a plurality of subframes each of which has a subframe length; calculating a weighted speech vector by the use of said spectrum parameters and said plurality of subframes; and generating a new excitation signal by the use of an adaptive code book comprising a plurality of adaptive code vectors and a sound source code book comprising a plurality of sound source code vectors, said generating step being carried out in a predetermined period, wherein, when said predetermined period is shorter than said subframe length, the new excitation signal is generated by the use of an adaptive code vector that is calculated by using the excitation signal generated in the former period and a sound source code vector of the present period. - View Dependent Claims (2, 3, 4, 5, 7)
-
-
6. An excitation signal encoding device including a frame division circuit for dividing a speech signal into a plurality of frames, an analyzer for carrying out a linear predictive analysis at every one of said plurality of frames to produce a parameter signal representative of spectrum parameters, a subframe division circuit for dividing each of said plurality of frames into a plurality of subframes, and a weighting circuit for calculating a weighted speech vector by the use of said spectrum parameters and said plurality of subframes, said excitation signal encoding device comprising:
-
an adaptive code book circuit for storing a plurality of adaptive code vectors and for selecting one of said plurality of adaptive code vectors as a selected adaptive code vector in response to an index signal, each of said plurality of adaptive code vectors being calculated by the use of an excitation signal calculated in the past; sound source code book circuit for storing a plurality of sound source code vectors and for selecting one of said plurality of sound source code vectors as a selected sound source code vector in response to said index signal; a calculation circuit for carrying out a predetermined calculation in a predetermined period by the use of a plurality of pitch gains, a plurality of sound source gains, said weighted speech vector, said selected adaptive code vector, and said selected sound source code vector, said calculation circuit producing a calculation result as an excitation vector; a weighting synthetic circuit supplied with said spectrum parameters and said excitation vector for carrying out a calculation on said excitation vector in accordance with said spectrum parameters to produce a weighted synthetic vector; a differential circuit supplied with said weighted speech vector and said weighted synthetic vector for calculating a difference between said weighted speech vector and said weighted synthetic vector to produce a difference signal representative of said difference; and an evaluation circuit supplied with said difference signal for carrying out an evaluation of said difference to supply an evaluation result, as said index signal, to said adaptive code book circuit and said sound source code book circuit, said evaluation circuit repeating said evaluation until it obtains a predetermined evaluation result, said evaluation circuit producing said index signal representative of an index of said sound source code vector and a last evaluation result upon obtaining said predetermined evaluation result.
-
-
8. An excitation signal encoding device including a frame division circuit for dividing a speech signal into a plurality of frames, an analyzer for carrying out a linear predictive analysis at every one of said plurality of frames to produce a parameter signal representative of spectrum parameters, a subframe division circuit for dividing each of said plurality of frames into a plurality of subframes, and a weighting circuit for calculating a weighted speech vector by the use of said spectrum parameters and said plurality of subframes, said excitation signal encoding device comprising:
-
an adaptive code book circuit for storing a plurality of adaptive code vectors and for selecting one of said plurality of adaptive code vectors as a selected adaptive code vector in response to a first index signal, each of said plurality of adaptive code vectors being calculated by the use of an excitation signal calculated in the past; a first calculation circuit supplied with said weighted speech vector and said selected adaptive code vector for carrying out a first predetermined calculation by the use of a plurality of pitch gains, said weighted speech vector, and said selected adaptive code vector, said first calculation circuit producing a first calculation result as a calculated adaptive code vector; a first weighting synthetic circuit supplied with said spectrum parameters and said calculated adaptive code vector for carrying out a calculation for said calculated adaptive code vector in accordance with said spectrum parameters to produce a first weighted synthetic vector; a first differential circuit supplied with said weighted speech vector and said first weighted synthetic vector for calculating a first difference between said weighted speech vector and said first weighted synthetic vector to produce a first difference signal representative of said first difference; a first evaluation circuit supplied with said first difference signal for carrying out an evaluation of said first difference to supply a first evaluation result, as said first index signal, to said adaptive code book circuit, said first evaluation circuit repeating said evaluation until it obtains a first predetermined evaluation result, said first evaluation circuit producing said first index signal for an optimum adaptive code vector and said optimum adaptive code vector upon obtaining said first predetermined evaluation result; a sound source code book circuit storing a plurality of sound source code vectors for selecting one of said plurality of sound source code vector as a selected sound source code vector in accordance with a second index signal; a second calculation circuit for carrying out a second predetermined calculation by the use of a plurality of sound source gains, said weighted speech vector, said selected sound source code vector of the present period, and said optimum adaptive code vector, said second calculation circuit producing a second calculation result as an excitation vector; a second weighting synthetic circuit supplied with said spectrum parameters and said excitation vector for carrying out a calculation for said excitation vector in accordance with said spectrum parameters to produce a second weighted synthetic vector; a second differential circuit supplied with said weighted speech vector and said second weighted synthetic vector for calculating a second difference between said weighted speech vector and said second weighted synthetic vector to produce a second difference signal representative of said second difference; and a second evaluation circuit supplied with said second difference signal for carrying out an evaluation of said second difference to supply a second evaluation result, as said second index signal, to said sound source code book circuit, said second evaluation circuit repeating said evaluation until it obtains a second predetermined evaluation result, said second evaluation circuit producing said second index signal for an optimum sound source code vector and a last evaluation result obtained upon obtaining said second predetermined evaluation result. - View Dependent Claims (9)
-
Specification