METHOD FOR ENCODING AND DECODING AN AUDIO SIGNAL AND APPARATUS FOR SAME

• US 20110320196A1
• Filed: 01/27/2010
• Published: 12/29/2011
• Est. Priority Date: 01/28/2009
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Choo, Ki Hyun, Sung, Ho Sang, Kim, Jung-Hoe, Oh, Eun Mi
• Granted Patent: US 8,918,324 B2
• Time in Patent Office: Days
• US Class Current: 704/229
• CPC Class Codes: G10L 19/12 the excitation function being a code excitation, e.g. in code excited linear prediction [CELP] vocoders
• CPC Class Codes: G10L 19/20 using sound class specific coding, hybrid encoders or object based coding

Abstract

A method for coding and decoding an audio signal or speech signal and an apparatus adopting the method are provided.

Claims

1. A coding method for an audio signal or speech signal, comprising:

switching an input signal to any one of a linear prediction domain (LPD) coding mode and a frequency domain coding mode according to a predetermined standard;

selecting LP coding modes of a plurality of frames included in one super frame when the input signal is switched to the LPD coding mode; and

controlling to include algebraic code excited linear prediction (ACELP) bit allocation information in a bit stream only when at least one of the LP coding modes is ACELP.

2. A decoding method for an audio signal or speech signal, comprising:

analyzing a linear prediction (LP) coding mode (lpd mode) included in a bit stream; and

performing transform coded excitation (TCX) decoding with respect to a super frame when a plurality of frames included in the super frame are coded by TCX as a result of the analysis, and performing algebraic code excited linear prediction (ACELP) decoding by further reading ACELP bit allocation information when at least a part of the super frame is coded by ACELP.

3. A decoding method for an audio signal or speech signal, comprising:

switching to a frequency domain decoding mode or a linear prediction domain (LPD) decoding mode by analyzing core mode information (core_mode) included in a bit stream;

reading context reset information (arith_reset_flag) with respect to a frame in the frequency domain decoding mode, thereby performing frequency domain decoding when the context reset information is set; and

analyzing an LP decoding mode (lpd_mode) of a super frame in the LPD decoding mode, and performing algebraic code excited linear prediction (ACELP) decoding when an object frame included in the super frame is coded by ACELP.

4. The decoding method of claim 3, further comprising:

reading the context reset information (arith_reset_flag) of the object frame when the LP coding mode is transform coded excitation (TCX); and

performing TCX decoding when the context reset information is set.

5. The decoding method of claim 3, wherein the frequency domain decoding is frequency domain decoding of unified speech and audio codec (USAC).

6. A coding method for an audio signal or speech signal, comprising:

switching an input signal to any one of a linear prediction domain (LPD) coding mode and a frequency domain coding mode according to a predetermined standard; and

coding the input signal to the LPD coding mode or the frequency domain coding mode such that random access availability information (random access) related to a plurality of frames of a coded super frame is included.

7. The coding method of claim 6, wherein the random access availability information is allocated to each of the plurality of frames.

8. The coding method of claim 6, wherein the random access availability information is included in a payload.

9. The coding method of claim 6, wherein context reset information (arith_reset_flag) is set to 1 in a frame which is randomly accessible.

10. The coding method of claim 6, wherein window type information (window_sequence) of a frame is 3 bits when the frame is randomly accessible.

11. A decoding method for an audio signal or speech signal, comprising:

determining whether a frame included in a super frame is randomly accessible by analyzing random access availability information (random access) included in a bit stream;

continuing determination of random accessibility of a next frame when the frame is not randomly accessible, switching to a frequency domain decoding mode or a linear prediction domain (LPD) decoding mode by analyzing core mode information (core_mode) when the frame is randomly accessible, and performing frequency domain decoding or LPD decoding.

12. A decoding method for an audio signal or speech signal, comprising:

analyzing linear prediction (LP) coding mode (lpd_mode) included in a bit stream; and

performing ACELP decoding by reading algebraic code excited linear prediction (ACELP) bit allocation information when a frame is coded by ACELP for the first time among at least one frame included in a super frame as a result of the analysis.

13. A decoding method for an audio signal or speech signal, comprising:

analyzing a linear prediction (LP) coding mode (lpd_mode) included in a bit stream; and

performing algebraic code excited linear prediction (ACELP) decoding of ACELP bit allocation information in a second ACELP frame, using ACELP bit allocation information read during decoding of a first ACELP frame, in a case where at least one frame included in a super frame is coded by ACELP as a result of the analysis.

14. A decoding method for an audio signal or speech signal, comprising:

analyzing a linear prediction (LP) coding mode (lpd_mode) included in a bit stream; and

performing algebraic code excited linear prediction (ACELP) decoding of ACELP bit allocation information in a second ACELP frame, using ACELP bit allocation information read during decoding of a first ACELP frame, in a case where at least one frame included in a super frame is coded by ACELP as a result of the analysis.

15. A decoding method for an audio signal or speech signal, comprising:

analyzing new linear prediction (LP) coding mode (new_lpd_mode) reconstructed in an order of a coding method;

determining whether to read an algebraic code excited linear prediction (ACELP) coding mode according to a value of the new LP coding mode;

reading the ACELP coding mode;

performing decoding according to the ACELP coding mode and the new LP coding mode.

16. A decoding method for an audio signal or speech signal, comprising:

analyzing a reconstructed temporal core mode (temporal_core_mode) by allocating a subordinate new linear prediction (LP) coding mode (new_lpd_mode) having a high availability in an algebraic code excited linear prediction (ACELP) coding mode;

determining the ACELP coding mode and the subordinate new LP coding mode (new_lpd_mode) according to the temporal core mode; and

performing decoding according to the ACELP coding mode and the subordinate new LP coding mode (new_lpd_mode).

17. A decoding method for an audio signal or speech signal, comprising:

analyzing a frame mode (frame_mode) for allocating subordinate groups of a new linear prediction (LP) coding mode (new_lpd_mode);

determining an algebraic code excited linear prediction (ACELP) coding mode corresponding to the frame mode (frame_mode) and a subordinate new LP coding mode (new_lpd_mode);

determining the ACELP coding mode and the subordinate new LP coding mode (new_lpd_mode) using the read ACELP coding mode and the subordinate new LP coding mode (new_lpd_mode); and

performing decoding according to the ACELP coding mode and the subordinate new LP coding mode (new_lpd_mode).