Speech decoding method and apparatus which generates an excitation signal and a synthesis filter

• US 8,249,866 B2
• Filed: 03/31/2010
• Issued: 08/21/2012
• Est. Priority Date: 04/04/2003
• Status: Active Grant

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Miseki, Kimio
• Primary Examiner(s): Smits, Talivaldis Ivars
• Application Number:
  US12/751,191
• Publication Number:
  US 20100250262A1
• Time in Patent Office: 874 Days
• Field of Search: None
• US Class Current: 704/223
• CPC Class Codes: G10L 19/18 Vocoders using multiple modes

Abstract

A speech decoding method which generates an excitation signal and a synthesis filter from coded data and which obtains a speech signal based on the excitation signal and the synthesis filter. The method includes acquiring identification information used for determining whether the speech signal to be decoded is a narrowband signal or a wideband signal; and modifying the excitation signal based on the identification information by controlling strength or presence of emphasis of pitch periodicity with respect to the excitation signal generated from the coded data, so as to generate the speech signal by use of the modified excitation signal and the synthesis filter.

Claims

1. A speech decoding method which generates an excitation signal and a synthesis filter from coded data and which obtains a speech signal based on the excitation signal and the synthesis filter, said method comprising:

acquiring identification information used for determining whether the speech signal to be decoded is a narrowband signal or a wideband signal; and

modifying the excitation signal based on the identification information by controlling strength or presence of emphasis of pitch periodicity with respect to the excitation signal generated from the coded data, so as to generate the speech signal by use of the modified excitation signal and the synthesis filter.

2. The speech decoding method according to claim 1, wherein:

the excitation signal includes an adaptive code vector and a noise code vector, and

the excitation signal is modified by controlling strength or presence of emphasis of pitch periodicity with respect to the adaptive code vector or the noise code vector.

3. The speech decoding method according to claim 1, wherein the identification information is acquired from the coded data or data attached to the coded data.

4. The speech decoding method according to claim 1, wherein the identification information is acquired by analyzing a signal reproduced in the decoding process or a spectrum parameter representing the outline of the speech signal.

5. The speech decoding method according to claim 1, wherein the identification information is acquired by a predetermined input unit of a decoding side.

6. The speech decoding method according to claim 1, wherein when the identification information represents a narrowband signal, sampling rate conversion is executed by maintaining a speech signal band assumed after the decoding processing and by using a fewer number of signals by down sampling.

7. A speech decoding method which generates an excitation signal and a synthesis filter from coded data and which uses a decoding process in which a speech signal is generated from the excitation signal and the synthesis filter, said method comprising:

acquiring identification information used for determining whether the speech signal to be decoded is a narrowband signal or a wideband signal; and

modifying the excitation signal based on the identification information by controlling strength or presence of emphasis of formant with respect to the excitation signal generated from the coded data, so as to generate the speech signal by use of the modified excitation signal and the synthesis filter.

8. The speech decoding method according to claim 7, wherein:

the excitation signal includes an adaptive code vector and a noise code vector, and

the excitation signal is modified by controlling strength or presence of emphasis of formant with respect to the adaptive code vector or the noise code vector.

9. A speech decoding method which generates an excitation signal and a synthesis filter from coded data and which obtains a speech signal based on the excitation signal and the synthesis filter, said method comprising:

determining whether the speech signal to be decoded is a narrowband signal or a wideband signal;

enhancing a spectrum envelope of the speech signal obtained based on the excitation signal and the synthesis filter, by a post-filter; and

switching parameter sets, which modifies characteristics of the post-filter, according to whether the speech signal is wideband signal or narrowband signal.

10. A speech decoding method which generates an excitation signal and a synthesis filter from coded data and which obtains a speech signal based on the excitation signal and the synthesis filter, said method comprising:

determining whether the speech signal to be decoded is a narrowband signal or a wideband signal;

enhancing a spectrum envelope of the speech signal obtained based on the excitation signal and the synthesis filter, by a post-filter; and

determining parameter sets for controlling a degree of emphasis by which the spectrum envelope is emphasized, a first parameter set used if the speech signal is the wideband signal and a second parameter used if the speech signal is the narrowband signal being determined such that the first parameter set provides a lower degree of emphasis than the second parameter set.

11. A speech decoding method which generates an excitation signal and a synthesis filter from coded data and which uses a decoding process including (i) a lower-band generation process in which a lower-band speech signal is generated from the excitation signal and the synthesis filter, and (ii) a higher-band generation process in which a higher-band signal applied to the lower-band speech signal is generated, said method comprising:

acquiring identification information used for determining whether the speech signal to be decoded is a narrowband signal or a wideband signal; and

controlling the decoding process such that, when the identification information represents a narrowband signal, the higher-band generation process is stopped, a higher-band signal generated in the higher-band generation process is modified to indicate zero or a small value, or the higher-band signal generated in the higher-band generation process is prevented from being applied to the lower-band speech signal generated in the lower-band generation process.

12. A speech decoding apparatus which employs:

a unit configured to generate an excitation signal from coded data, a unit configured to generate a synthesis filter, and a unit configured to decode a speech signal from the excitation signal and the synthesis filter, said apparatus comprising;

a unit configured to determine whether the speech signal to be decoded is a narrowband signal or a wideband signal;

a unit configured to obtain the speech signal by having the speech signal, obtained based on the excitation signal and the synthesis filter filtered through a post-filter which enhances a spectrum envelope of the speech signal; and

a unit configured to switch parameter sets, used for modifying characteristics of the post-filter, according to whether the speech signal is a wideband signal or a narrowband signal.

13. A speech decoding apparatus which employs:

a unit configured to generate an excitation signal from coded data, a unit configured to generate a synthesis filter, and a unit configured to decode a speech signal from the excitation signal and the synthesis filter, said apparatus comprising;

a determination unit configured to determine whether the speech signal to be decoded is a narrowband signal or a wideband signal;

a unit configured to obtain the speech signal by having the speech signal, obtained based on the excitation signal and the synthesis filter filtered through a post-filter which enhances a spectrum envelope of the speech signal; and

a unit configured to determine parameter sets for controlling a degree of emphasis by which the spectrum envelope used by the post-filter is emphasized, a first parameter set used if the speech signal is the wideband signal and a second parameter used if the speech signal is the narrowband signal being determined such that the first parameter set provides a lower degree of emphasis than the second parameter set.