Nova Patents
EP1760695B1

Orthogonalization search for the CELP based speech coding

Abstract

This record has no abstract on file.

EP1760695B1, drawing sheet 1
Sheet 1 of 57

Term

Term ended

Expired 22 October 2018, 7.9 years ago.

  1. Priority
  2. Filed
  3. Granted
  4. Expired
  5. Today

2 claims: 2 independent, 0 dependent

  1. 1
    A CELP speech coder using an orthogonal search, for determining a search reference value comprising a numerator and a denominator term, the speech coder comprising:an adaptive codebook that generates an adaptive vector representing a pitch component;an ACELP type random codebook that generates a random vector representing a random component;a synthesis filter that uses filter coefficients obtained by analyzing and quantizing an input speech signal;and a distortion calculator that calculates a distortion between the input speech signal and a synthetic speech signal obtained by the synthesis filter excited by the adaptive vector and the random vector, and selects one adaptive vector and random vector that minimize the distortion by maximizing the search reference value, the numerator term being: Nort = a 0 ⁢ φ l 0 + a 1 ⁢ φ l 1 + … + a n - 1 ⁢ φ l n 2 wherein a i is the sign of the i - th pulse;l i is the position of the i - th pulse;n is the number of pulses;and φ is p t ⁢ H t ⁢ Hp ⁢ x - x t ⁢ Hp ⁢ Hp ⁢ H ;where x is a target signal obtained by differentiating a zero input response of the synthesis filter from the input speech signal, characterized in that the distortion calculator calculating the denominator comprises: a system that computes power p t H t H of a signal, Hp, obtained by synthesis by synthesis filter using the adaptive codevector, computes an auto-correlation matrix, H t H, of the filter coefficients of the synthesis filter, and calculates a first matrix, N=(p t H t Hp) H t H, by multiplying each element of the auto-correlation matrix by the power;a system that calculates a second matrix, M, by providing a time reverse synthesis, r t =p t H t H, to the signal, Hp, obtained by synthesis in the synthesis filter using the adaptive codevector and taking an outer product, m=rr t , of the resultant signal by the time reverse synthesis;a system that calculates a third matrix, L=N-M, by using the first matrix and the second matrix;and a system that calculates the denominator term of the search reference value by a double summation calculation over the elements of the third matrix using a value obtained by multiplying an element, in the third matrix, associated with two pulse positions of a random codevector, by a value in accordance with two polarity values for each pulse at said two pulse position of the random codevector;wherein p is the adaptive codevector, H is the synthesis filter coefficient matrix, t denotes transpose.
  2. 2
    A CELP speech coding method comprising:selecting by an orthogonal search an adaptive vector of an adaptive codebook and a random vector of a random codebook having an algebraic structure by minimizing a distortion between an input speech signal and a synthetic speech signal, the synthetic speech signal being obtained by exciting a synthesis filter with the adaptive vector and random vector, the orthogonal search minimizing the distortion by maximizing a search reference value, the search reference value comprising a numerator term and a denominator term, the numerator term being: Nort = a 0 ⁢ φ l 0 + a 1 ⁢ φ l 1 + … + a n - 1 ⁢ φ ⁢ l n - 1 2 wherein a i is the sign of the i-th pulse;l i is the position of the i-th pulse;n is the number of pulses;and φ is p t ⁢ H t ⁢ Hp ⁢ x - x t ⁢ Hp ⁢ Hp where x is a target signal obtained by differentiating a zero input response of the synthesis filter from the input speech signal, the orthogonal search method characterized by the following steps to calculate the denominator: computing power p t H t Hp of a signal, Hp, obtained by synthesis in a synthesis filter using an adaptive codevector, calculating a first matrix N = ( p t H t Hp ) H t H, by multiplying each element of the auto-correlation matrix by the power;calculating a second matrix, M, by providing a time reverse synthesis r t = p t H t H, to the signal, Hp, obtained by synthesis in a synthesis filter using the adaptive codevector and taking an outer product, M=rr t , of the resultant signal by the time reverse synthesis;calculating a third matrix, L=N-M, by using the first matrix and the second matrix;and the denominator term of the search reference value by a double summation calculation over the elements of the third matrix using a value obtained by multiplying an element, in the third matrix, associated with two pulse positions of a random codevector, by a value in accordance with two polarity values for each pulse at said two pulse positions of the random codevector;wherein p is the adaptive codevector, H is the synthesis filter coefficient matrix, t denotes transpose.
Independent claims2