HYBRID INSTANTANEOUS/DIFFERENTIAL PITCH PERIOD CODING

• US 20110029304A1
• Filed: 07/30/2010
• Published: 02/03/2011
• Est. Priority Date: 08/03/2009
• Status: Active Grant

• Current Assignee: Avago Technologies International Sales Pte Limited (Broadcom, Inc.)
• Original Assignee: Broadcom Corporation (Broadcom, Inc.)
• Inventors: Kang, Hong-Goo, Chen, Juin-Hwey
• Granted Patent: US 9,269,366 B2
• Time in Patent Office: Days
• US Class Current: 704/208
• CPC Class Codes: G10L 19/08 Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters
• CPC Class Codes: G10L 19/22 Mode decision, i.e. based on audio signal content versus external parameters

Abstract

A hybrid instantaneous/differential encoding technique is described herein that may be used to reduce the bit rate required to encode a pitch period associated with a segment of a speech signal in a manner that will result in relatively little or no degradation of a decoded speech signal generated using the encoded pitch period. The hybrid instantaneous/differential encoding technique is advantageously applicable to any speech codec that encodes a pitch period associated with a segment of a speech signal.

Claims

1. A method for encoding an audio signal comprising a series of temporally-ordered segments, comprising:

determining if instantaneous coding or differential coding should be applied to encode a pitch period associated with a current segment of the audio signal by determining if a magnitude of the difference between the pitch period associated with the current segment and the pitch period associated with the previous segment exceeds a threshold;

responsive to determining that instantaneous coding should be applied, outputting a quantized representation of the pitch period associated with the current segment as part of an encoded representation of the current segment; and

responsive to determining that differential coding should be applied, encoding a difference between the pitch period associated with the current segment and a pitch period associated with a previous segment in the series of segments and outputting the encoded difference rather than the quantized representation of the pitch period as part of the encoded representation of the current segment.

2. A method for encoding an audio signal comprising a series of temporally-ordered segments, comprising:

determining if instantaneous coding or differential coding should be applied to encode a pitch period associated with a current segment of the audio signal by determining if the current segment represents a first segment of a voiced speech region of the audio signal;

responsive to determining that instantaneous coding should be applied, outputting a quantized representation of the pitch period associated with the current segment as part of an encoded representation of the current segment; and

responsive to determining that differential coding should be applied, encoding a difference between the pitch period associated with the current segment and a pitch period associated with a previous segment in the series of segments and outputting the encoded difference rather than the quantized representation of the pitch period as part of the encoded representation of the current segment.

3. The method of claim 2, wherein determining if the current segment represents a first segment of a voiced speech region of the audio signal comprises:

determining if each of the previous segment and the current segment represents voiced speech; and

responsive to determining that the previous segment does not represent voiced speech and that the current segment represents voiced speech, determining that the current segment represents a first segment of a voiced speech region of the audio signal.

4. The method of claim 3, wherein determining if each of the previous segment and the current segment represents voiced speech comprises determining if each of the previous segment and the current segment represents voiced speech based on signal characteristics associated with each of the segments.

5. The method of claim 3, wherein determining if each of the previous segment and the current segment represents voiced speech comprises determining if each of the previous segment and the current segment represents voiced speech based on an encoding mode identifier associated with each of the segments.

6. The method of claim 3, wherein determining if each of the previous segment and the current segment represents voiced speech comprises determining if each of the previous segment and the current segment represents one of silence, unvoiced speech, stationary voiced speech, or non-stationary voiced speech.

7. The method of claim 2, wherein encoding the difference between the pitch period associated with the current segment and the pitch period associated with the previous segment comprises:

quantizing the difference using a number of bits that is less than a number of bits used to generate the quantized representation of the pitch period.

8. The method of claim 2, wherein encoding the difference between the pitch period associated with the current segment and the pitch period associated with the previous segment comprises:

applying entropy coding to encode the difference.

9. The method of claim 8, wherein applying entropy coding to encode the difference comprises:

applying Huffman coding to encode the difference.

10. The method of claim 9, wherein applying Huffman coding to encode the difference comprises:

selecting one of a plurality of different Huffman codes to represent the difference, wherein each of the plurality of different Huffman codes is of a different length and consists of one or more zeroes followed by a one.

11. The method of claim 2, further comprising:

determining the pitch period associated with the previous segment and the current segment using a pitch period extraction algorithm that operates to smooth a pitch contour associated with the audio signal.

12. The method of claim 11, wherein using a pitch period extraction algorithm that operates to smooth a pitch contour associated with the audio signal comprises:

performing a first-pass pitch period extraction process that extracts first-pass pitch periods associated with the audio signal, the first-pass pitch periods collectively representing a first-pass pitch contour of the audio signal;

storing the first-pass pitch periods;

performing a second-pass pitch period extraction process that utilizes the stored first-pass pitch periods and the audio signal to obtain second-pass pitch periods associated with the audio signal, the second-pass pitch periods collectively representing a smoothed version of the first-pass pitch contour.

13. The method of claim 12, wherein performing the second-pass pitch period extraction process to obtain second-pass pitch periods associated with the audio signal comprises enforcing a constraint upon the size of a difference between a pitch period associated with two adjacent segments of the audio signal.

14. The method of claim 2, wherein:

further responsive to determining that instantaneous coding should be applied to encode the pitch period associated with the current segment of the audio signal, determining that differential coding should be applied to encode a pitch period associated with a subsequent segment in the series of segments.

15. A method for decoding an audio signal comprising a series of temporally-ordered segments, comprising:

receiving an encoded representation of a current segment of the audio signal;

determining if a pitch period associated with the current segment has been encoded in accordance with an instantaneous coding process or a differential coding process based on one or more bits included in the encoded representation of the current segment;

responsive to determining that the pitch period associated with the current segment has been encoded in accordance with an instantaneous coding process, obtaining the pitch period associated with the current segment by de-quantizing a quantized representation of the pitch period associated with the current segment that is included in the encoded representation of the current segment; and

responsive to determining that the pitch period associated with the current segment has been encoded in accordance with a differential coding process, obtaining the pitch period associated with the current segment by decoding an encoded representation of a difference that is included in the encoded representation of the current segment and by adding the difference to a pitch period associated with a previous segment in the series of segments.

16. The method of claim 15, wherein determining if the pitch period associated with the current segment has been encoded in accordance with the instantaneous coding process or the differential coding process based on one or more bits included in the encoded representation of the current segment comprises:

determining the value of a binary flag that is included in the encoded representation of the current segment.

17. The method of claim 15, wherein determining if the pitch period associated with the current segment has been encoded in accordance with the instantaneous coding process or the differential coding process based on one or more bits included in the encoded representation of the current segment comprises:

determining if the current segment represents a first segment of a voiced speech region of the audio signal based on at least one or more bits included in the encoded representation of the current segment;

responsive to determining that the current segment represents a first segment of a voiced speech region of the audio signal, determining that the pitch period associated with the current segment has been encoded in accordance with the instantaneous coding process; and

responsive to determining that the current segment does not represent a first segment of a voiced speech region of the audio signal, determining that the pitch period associated with the current segment has been encoded in accordance with the differential coding process.

18. The method of claim 17, wherein determining if the current segment represents the first segment of a voiced speech region of the audio signal based on at least one or more bits included in the encoded representation of the current segment comprises:

determining if the previous segment represents voiced speech based on one or more bits included in an encoded representation of the previous segment;

determining if the current segment represents voiced speech based on one or more bits included in the encoded representation of the current segment; and

determining that the current segment represents the first segment of a voiced speech region of the audio signal if it is determined that the previous segment does not represent voiced speech and that the current segment represents voiced speech.

19. The method of claim 15, wherein decoding the encoded representation of the difference comprises:

de-quantizing a quantized representation of the difference, the quantized representation of the difference being represented by a number of bits that is less than a number of bits used to represent the quantized representation of the pitch period.

20. The method of claim 15, wherein decoding the encoded representation of the difference comprises:

applying entropy decoding to decode the encoded representation of the difference.

21. The method of claim 20, wherein applying entropy decoding to decode the encoded representation of the difference comprises:

applying Huffman decoding to decode the encoded representation of the difference.

22. The method of claim 21, wherein applying Huffman decoding to decode the encoded representation of the difference comprises:

obtaining a difference associated with one of a plurality of Huffman codes that comprises the encoded representation of the difference, wherein each of the plurality of different Huffman codes is of a different length and consists of one or more zeroes followed by a one.

23. The method of claim 15, wherein:

further responsive to determining that the pitch period associated with the current segment has been encoded in accordance with the instantaneous coding process, determining that a pitch period associated with a subsequent segment in the series of segments has been encoded in accordance with a differential coding process.

24. A system, comprising:

an encoder that generates an encoded representation of each of a series of temporally-ordered segments that comprise an audio signal by selectively applying either instantaneous coding or differential encoding to encode a pitch period associated with each segment based on whether the segment represents a first segment of a voiced speech region of the audio signal, wherein applying instantaneous coding comprises generating a quantized representation of a pitch period associated with a segment and wherein applying differential encoding comprises generating an encoded representation of a difference between a pitch period associated with a segment and a pitch period associated with a previous segment in the series of segments; and

a decoder that decodes the encoded representation of each of the series of temporally-ordered segments by selectively decoding, for each encoded representation, either a quantized representation of a pitch period or an encoded representation of a difference between a pitch period associated with a segment and a pitch period associated with a previous segment in the series of segments.