Apparatus for processing a signal and method thereof for determining an LPC coding degree based on reduction of a value of LPC residual

• US 9,613,630 B2
• Filed: 11/12/2010
• Issued: 04/04/2017
• Est. Priority Date: 11/12/2009
• Status: Expired due to Fees

• Current Assignee: LG Electronics, Inc. (LG Corporation)
• Original Assignee: LG Electronics, Inc. (LG Corporation)
• Inventors: Yoon, Sung Yong, Choi, Tack Sung, Lee, Hyun Kook
• Primary Examiner(s): Kazeminezhad, Farzad
• Application Number:
  US13/508,410
• Publication Number:
  US 20120226496A1
• Time in Patent Office: 2,335 Days
• Field of Search: 704201
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 19/0212 using orthogonal transformation
• CPC Class Codes: G10L 19/06 Determination or coding of the spectral characteristics, e.g. of the short-term prediction coefficients
• CPC Class Codes: G10L 19/107 Sparse pulse excitation, e.g. by using algebraic codebook
• CPC Class Codes: G10L 19/20 using sound class specific coding, hybrid encoders or object based coding
• CPC Class Codes: G10L 19/22 Mode decision, i.e. based on audio signal content versus external parameters

Abstract

An apparatus for processing a signal and method thereof are disclosed. The present invention includes receiving coding mode information indicating a speech coding scheme or an audio coding scheme, linear prediction coding degree information indicating a linear prediction coding degree, and the signal including at least one of a speech signal and an audio signal; decoding the signal according to the speech coding scheme or the audio coding scheme based on the coding mode information; decoding linear prediction coding coefficients of the signal based on the linear prediction coding degree information; and generating an output signal by applying the decoded linear prediction coding coefficients to the decoded signal. In this case, the linear prediction coding degree information is determined based on a variation of a value of an LPC residual generated from performing the linear prediction coding on the signal. It further determines whether or not the said residual decreases by an increase in the linear prediction coding degree, and it determines an appropriate LPC coding degree in each case.

Claims

1. A method of processing a signal, comprising:

receiving coding mode information indicating a speech coding scheme or an audio coding scheme, linear prediction coding degree information indicating a linear prediction coding degree, and the signal including at least one of a speech signal or an audio signal;

decoding the signal according to the speech coding scheme or the audio coding scheme based on the coding mode information;

decoding linear prediction coding coefficients of the signal based on the linear prediction coding degree information; and

generating an output signal by applying the decoded linear prediction coding coefficients to the decoded signal,

wherein the linear prediction coding degree is determined based on a variation of a value of an LPC residual generated from performing the linear prediction coding on the signal by;

determining whether the value of the LPC residual generated from performing the linear prediction coding is reduced by increasing the linear prediction coding degree;

based on a determination that the value of the LPC residual is reduced by increasing the linear prediction coding degree, determining the linear prediction coding degree to be a first degree; and

based on a determination that the value of the LPC residual is not reduced by increasing the linear prediction coding degree, determining the linear prediction coding degree to be a second degree lower than the first degree,

wherein the first degree is greater than zero degree.

2. The method of claim 1, wherein if the signal is the speech signal having a lot of voiced sound, the linear prediction coding degree information indicates a degree higher than that of the speech signal having a lot of unvoiced sound.

3. The method of claim 1, wherein if the signal is the audio signal having a strong tonal component, the linear prediction coding degree information indicates a degree higher than that of the audio signal having a weak tonal component.

4. The method of claim 1, wherein if the signal is an audio-like signal, the linear prediction coding degree information indicates a degree lower than that of a speech-like signal.

5. The method of claim 1, wherein a frame length of the audio signal is an integer multiple of a frame length of the speech signal.

6. The method of claim 1, further comprising:

if the decoded signal is the speech or audio signal and a signal of a previous frame is different from the decoded signal, preventing aliasing by compensating the decoded signal using the signal of the previous frame.

7. An apparatus for processing a signal, comprising:

a multiplexer receiving an LPC bitstream including linear prediction coding degree information indicating a linear prediction coding degree and linear prediction coding coefficients, the signal including at least one of a speech signal or an audio signal, and coding mode information indicating a speech coding scheme or an audio coding scheme;

an ACELP decoding unit decoding the signal according to the speech coding scheme if the coding mode information indicates the speech coding scheme;

a TCX decoding unit decoding the signal according to the audio coding scheme if the coding mode information indicates the audio coding scheme; and

an LPC synthesis unit decoding the linear prediction coding coefficients of the signal based on the linear prediction coding degree information, and generating an output signal by applying the decoded linear prediction coding coefficients to the decoded signal,

wherein the linear prediction coding degree information is determined based on a variation of a value of an LPC residual generated from performing the linear prediction coding on the signal by;

determining whether the value of the LPC residual generated from performing the linear prediction coding is reduced by increasing the linear prediction coding degree;

based on a determination that the value of LPC residual is reduced by increasing the linear prediction coding degree, determining the linear prediction coding degree to be a first degree; and

based on a determination that the value of LPC residual is not reduced by increasing the linear prediction coding degree, determining the linear prediction coding degree to be a second degree lower than the first degree,

wherein the first degree is greater than zero degree.

8. The apparatus of claim 7, wherein the LPC synthesis unit comprises:

an LPC degree information decoding unit decoding the linear prediction coding degree information from the LPC bitstream;

an LPC coefficients extracting unit extracting the linear prediction coding coefficients from the LPC bitstream based on the linear prediction coding degree information; and

a signal synthesis unit generating the output signal by applying the linear prediction coding coefficients to the decoded signal.

9. The apparatus of claim 8, if the linear prediction coding degree information indicates that the linear prediction coding coefficients are not decoded, wherein the decoded signal bypasses the signal synthesis unit and is then outputted.

10. The apparatus of claim 7, wherein if the signal is the speech signal having a lot of voiced sound, the linear prediction coding degree information indicates a degree higher than that of the speech signal having a lot of unvoiced sound.

11. The apparatus of claim 7, wherein if the signal is the audio signal having a strong tonal component, the linear prediction coding degree information indicates a degree higher than that of the audio signal having a week tonal component.

12. The apparatus of claim 7, further comprising:

if a coding scheme of the decoded signal adopts the coding scheme different from that of the decoded signal of a previous frame, a signal compensating unit preventing aliasing by compensating the decoded signal using the decoded signal of the previous frame.

13. The apparatus of claim 7, wherein a frame length of the audio signal is an integer multiple of a frame length of the speech signal.

14. A method of processing a signal comprising:

determining linear prediction coding degree information indicating a degree of linear prediction coding coefficients according to property of an input signal;

determining the linear prediction coding coefficients based on the linear prediction coding degree information;

generating an LPC residual using the linear prediction coding coefficients and the input signal; and

encoding the LPC residual using either an audio coding scheme or a speech coding scheme,

wherein the linear prediction coding degree information is determined based on a variation of a value of an LPC residual generated from performing the linear prediction coding on the input signal by;

determining whether the value of the LPC residual generated from performing the linear prediction coding is reduced by increasing the linear prediction coding degree;

based on a determination that the value of LPC residual is reduced by increasing the linear prediction coding degree, determining the linear prediction coding degree to be a first degree; and

based on a determination that the value of LPC residual is not reduced by increasing the linear prediction coding degree, determining the linear prediction coding degree to be a second degree lower than the first degree,

wherein the first degree is greater than zero degree.

15. An apparatus for processing a signal, comprising:

an LPC analysis unit determining linear prediction coding degree information indicating a variable degree of linear prediction coding coefficients according to property of an input signal, the LPC analysis unit determining the linear prediction coding coefficients based on the linear prediction coding degree information, the LPC analysis unit generating an LPC residual using the linear prediction coding coefficients and the input signal;

an ACELP encoding unit encoding the LPC residual using an audio coding scheme;

a TCX encoding unit encoding the LPC residual using a speech coding scheme;

and a multiplexer generating a bitstream including the linear prediction coding degree information, the linear prediction coding coefficients and the encoded LPC residual encoded by the ACELP encoding unit or the TCX encoding unit,

wherein the linear prediction coding degree information is determined based on a variation of a value of an LPC residual generated from performing the linear prediction coding on the input signal by;

determining whether the value of the LPC residual generated from performing the linear prediction coding is reduced by increasing the linear prediction coding degree;

based on a determination that the value of LPC residual is reduced by increasing the linear prediction coding degree, determining the linear prediction coding degree to be a first degree; and

based on a determination that the value of LPC residual is not reduced by increasing the linear prediction coding degree, determining the linear prediction coding degree to be a second degree lower than the first degree,

wherein the first degree is greater than zero degree.