EP3764355B1

Encoding method, decoding method, encoding apparatus, and decoding apparatus

Abstract

This record has no abstract on file.

EP3764355B1, drawing sheet 1
Sheet 1 of 20

Term

6.8 yearsleft in the term

Expires 25 July 2033.

  1. Priority and filed
  2. Granted
  3. Today
  4. Expires

18 claims: 4 independent, 14 dependent

  1. 1
    An encoding method (100) for encoding a speech signal, comprising:dividing (110) the to-be-encoded time-domain speech signal into a low band signal and a high band signal, according to a frequency threshold;performing encoding (120) on the low band signal to obtain a low frequency encoding parameter, wherein the low frequency encoding parameter includes an algebraic codebook gain;performing linear predictive coding, LPC, analysis on the high band signal to obtain a LPC coefficient, predicting a high frequency excitation signal using the low frequency encoding parameter, obtaining a synthesized high band signal, using the high frequency excitation signal and a synthesis filter that is determined according to the LPC coefficient (130);the method being characterised by further comprising: performing filtering processing on the synthesized high band signal by using a pole-zero post-filter to obtain a short-time filtered signal, wherein a coefficient of the pole-zero post-filter is set based on the LPC coefficient;and calculating (150) a high frequency gain based on the high band signal and the short-time filtered signal.
  2. 4
    The encoding method according to any one of claims 1 to 3, wherein a z-domain transfer function of the pole-zero post-filter is calculated by using the following formula:H s z = 1 − a 1 βz − 1 − a 2 β 2 z − 2 − … − a M β M z − M 1 − a 1 γz − 1 − a 2 γ 2 z − 2 − … − a M γ M z − M wherein a 1 , a 2 ,...... a M is the LPC coefficient, M is an order of the LPC coefficient, β is 0.5 and γ is 0.8.
  3. 5
    The encoding method according to any one of claims 1 to 4, wherein the encoding method further comprises:generating an encoding bitstream according to the low frequency encoding parameter, the LPC coefficient, and the high frequency gain.
  4. 6
    A decoding method (200) for decoding a speech signal, comprising:differentiating (210) a low frequency encoding parameter, a LPC coefficient, and a high frequency gain from encoded information, wherein the low frequency encoding parameter includes an algebraic codebook gain;performing decoding (220) on the low frequency encoding parameter to obtain a low band signal, the frequencies the low band signal are lower than a frequency threshold;restoring a high frequency excitation signal according to the low frequency encoding parameter, generating a synthesis filter according to the LPC coefficient, performing filtering, by using the synthesis filter, on the high frequency excitation signal to obtain a synthesized high band signal (230);the method being characterised by further comprising: performing filtering processing on the synthesized high band signal by using a pole-zero post-filter to obtain a short-time filtered signal, wherein a coefficient of the pole-zero post-filter is set based on the LPC coefficient;adjusting (250) the short-time filtered signal by using the high frequency gain to obtain a high band signal;and combining (260) the low band signal and the high band signal to obtain a final decoding signal.
  5. 9
    The decoding method according to any one of claims 6 to 8, and a z-domain transfer function of the pole-zero post-filter is calculated by using the following formula:H s z = 1 − a 1 βz − 1 − a 2 β 2 z − 2 − … − a M β M z − M 1 − a 1 γz − 1 − a 2 γ 2 z − 2 − … − a M γ M z − M wherein a 1 , a 2 ,...... a M is the LPC coefficient, M is an order of the LPC coefficient, β is 0.5 and γ is 0.8.
  6. 10
    An encoding apparatus (300) for encoding a speech signal, comprising:a division unit (310), configured to divide the to-be-encoded time-domain speech signal into a low band signal and a high band signal, according to a frequency threshold;a low frequency encoding unit (320), configured to perform encoding on the low band signal to obtain a low frequency encoding parameter, wherein the low frequency encoding parameter includes an algebraic codebook gain;a high frequency encoding unit (330), configured to perform linear predictive coding ,LPC, analysis on the high band signal to obtain a LPC coefficient;a synthesizing unit (340), configured to predict a high frequency excitation signal according to the low frequency encoding parameter, and enables the high frequency excitation signal to pass to a synthesis filter that is determined according to the LPC coefficient, to obtain a synthesized high band signal;the apparatus being characterised by further comprising: a pole-zero post-filter (410), configured to perform filtering processing on the synthesized high band signal to obtain a short-time filtered signal, wherein a coefficient of the pole-zero post-filter is set based on the LPC coefficient;and a calculation unit (360), configured to calculate a high frequency gain based on the high band signal and the short-time filtered signal.
  7. 13
    The encoding apparatus (300) according to any one of claims 10 to 12, wherein a z-domain transfer function of the pole-zero post-filter (410) is calculated by using the following formula:H s z = 1 − a 1 βz − 1 − a 2 β 2 z − 2 − … − a M β M z − M 1 − a 1 γz − 1 − a 2 γ 2 z − 2 − … − a M γ M z − M wherein a 1 , a 2 ,...... a M is the LPC coefficient, M is an order of the LPC coefficient, β is 0.5 and γ is 0.8 .
  8. 14
    The encoding apparatus (300) according to any one of claims 10 to 13, wherein the encoding apparatus further comprises:a bitstream generating unit, configured to generate an encoding bitstream according to the low frequency encoding parameter, the LPC coefficient, and the high frequency gain.
  9. 15
    A decoding apparatus for decoding a speech signal (500), comprising:a differentiating unit (510), configured to differentiate a low frequency encoding parameter, a LPC coefficient, and a high frequency gain from encoded information, wherein the low frequency encoding parameter includes an algebraic codebook gain;a low frequency decoding unit (520), configured to perform decoding on the low frequency encoding parameter to obtain a low band signal, the frequencies of the low band signal are lower than a frequency threshold;a synthesizing unit (530), configured to use the low frequency encoding parameter to restore a high frequency excitation signal, use the LPC coefficient to generate a synthesis filter, and use the synthesis filter to perform filtering on the high frequency excitation signal to obtain a synthesized high band signal;the apparatus being characterised by further comprising: a pole-zero post-filter, configured to perform filtering processing on the synthesized high band signal to obtain a short-time filtered signal, wherein a coefficient of the pole-zero post-filter is set based on the LPC coefficient;a high frequency decoding unit (550), configured to adjust the short-time filtered signal by using the high frequency gain to obtain a high band signal;and a combining unit (560), configured to combine the low band signal and the high band signal to obtain a final decoding signal.
  10. 18
    The decoding apparatus (500) according to any one of claims 15 to 17, wherein a z-domain transfer function of the pole-zero post-filter is calculated by using the following formula:H s z = 1 − a 1 βz − 1 − a 2 β 2 z − 2 − … − a M β M z − M 1 − a 1 γz − 1 − a 2 γ 2 z − 2 − … − a M γ M z − M wherein a 1 , a 2 ,...... a M is the LPC coefficient, M is an order of the LPC coefficient, β is 0.5 and γ is 0.8.