Efficient encoding/decoding of audio signals

• US 9,280,980 B2
• Filed: 02/09/2011
• Issued: 03/08/2016
• Est. Priority Date: 02/09/2011
• Status: Active Grant

• Current Assignee: Telefonaktiebolaget LM Ericsson
• Original Assignee: Telefonaktiebolaget LM Ericsson
• Inventors: Grancharov, Volodya, Norvell, Erik, Sverrisson, Sigurdur
• Primary Examiner(s): Chawan, Vijay B
• Application Number:
  US13/982,515
• Publication Number:
  US 20130317811A1
• Time in Patent Office: 1,854 Days
• Field of Search: 704/230, 704/229, 704/200.1, 704500-504, 704/205, 704/219, 704/225, 704/226, 704/224, 381/94.3, 381/94.5, 381/98
• US Class Current: 1/1
• CPC Class Codes: G10L 19/0204 using subband decomposition
• CPC Class Codes: G10L 19/0212 using orthogonal transformation
• 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 19/24 Variable rate codecs, e.g. for generating different qualities using a scalable representation such as hierarchical encoding or layered encoding
• CPC Class Codes: G10L 2019/0008 Algebraic codebooks
• CPC Class Codes: G10L 21/0388 Details of processing therefor

Abstract

A method for encoding of an audio signal comprises performing (214) of a transform of the audio signal. An energy offset is selected (216) for each of the first subbands. An energy measure of a first reference band within a low band of an encoding of a synthesis signal is obtained (212). The first high band is encoded (220) by providing quantization indices representing a respective scalar quantization of a spectrum envelope in the first subbands of the first high band relative to the energy measure of the first reference band by use of the selected energy offset. An encoder apparatus comprises means for carrying out the steps of the method. Corresponding decoder methods and apparatuses are also described.

Claims

1. A method, in an audio encoding device, for encoding of an audio signal, the method comprising:

obtaining a low band synthesis signal of an encoding of said audio signal;

obtaining a first energy measure of a first reference band within a low band (LB) in said low band synthesis signal;

performing a transform of said audio signal into a transform domain;

selecting an energy offset from a set of at least two predetermined energy offsets for each of a plurality of first subbands of a first high band (HB-1) of said audio signal in said transform domain, said first high band (HB-1) being situated at higher frequencies than said low band (LB); and

encoding said first high band (HB-1), wherein said encoding of said first high band (HB-1) comprises providing a first set of quantization indices representing a respective scalar quantization of a spectrum envelope in said plurality of first subbands of said first high band (HB-1) relative to said first energy measure, said first set of quantization indices being given with a respective said selected energy offset, and wherein said encoding of said first high band (HB-1) further comprises providing a parameter defining the used energy offset;

obtaining a second energy measure of a second reference band within said low band (LB) in said low band synthesis signal; and

encoding a second high band (HB-2) of said audio signal in said transform domain, said second high band (HB-2) being situated in frequency between said low band (LB) and said first high band (HB-1), and wherein said encoding of said second high band (HB-2) comprises providing a second set of quantization indices representing a respective scalar quantization of a spectrum envelope in a plurality of second subbands of said second high band (HB-2) relative to said second energy measure.

2. The method of claim 1, wherein said selecting an energy offset is dependent on a power distribution of said audio signal in a frequency domain.

3. The method of claim 1, wherein said selecting an energy offset is based on an open loop procedure, comprising determining a parameter characterizing a power distribution of said low band synthesis signal in a frequency domain, whereby said selecting is based on said determined parameter.

4. The method of claim 1, wherein

said encoding in turn comprises providing one first set of said quantization indices for each predetermined energy offset range;

and

said selecting an energy offset in turn comprises;

calculating a quantization error for each of said first sets of quantization indices; and

selecting the first set of said quantization indices that gives the smallest quantization error.

5. The method of claim 1, wherein said transform encoding is a Modified Discrete Cosine Transform.

6. The method of claim 1, wherein a low frequency end of said first high band (HB-1) is 8 kHz.

7. The method of claim 1, wherein a high frequency end of said first high band is (HB-1) 11.6 kHz.

8. The method according claim 1, wherein said first high band (HB-1) comprises five first subbands.

9. The method of claim 1, wherein said low band (LB) ranges from 0-6.4 kHz.

10. The method of claim 1, wherein said first reference band comprises entire said low band (LB).

11. The method of claim 1, wherein said first reference band ranges from 0-5.9 kHz.

12. The method of claim 1, wherein said low band synthesis signal is based on an encoding by a Code-Excited Linear Prediction encoder.

13. The method of claim 1, wherein the quantization indices of said second set of quantization indices are restricted in increased energy direction.

14. The method of claim 13, wherein said energy restriction of said quantization indices is reduced with increasing frequency of said second subbands.

15. The method of claim 1, wherein said second high band (HB-2) ranges between 6.4 and 8 kHz.

16. The method of claim 1, wherein said second reference band ranges between 5.9 and 6.4 kHz.

17. The method of claim 1, wherein said second high band (HB-2) comprises three second subbands.

18. A method, in an audio decoding device, for decoding of an audio signal, the method comprising:

receiving an encoding of said audio signal, said encoding representing a first set of quantization indices of a spectrum envelope in a plurality of first subbands of a first high band (HB-1) of said audio signal, said first set of quantization indices representing energies relative to a first energy measure, said encoding further representing a parameter defining a used energy offset, wherein said encoding further represents a second set of quantization indices of a spectrum envelope in a plurality of second subbands of a second high band (HB-2) of said audio signal, said second set of quantization indices representing energies relative to a second energy measure;

obtaining a low band synthesis signal of an encoding of said audio signal;

obtaining said first energy measure as an energy measure of a first reference band within a low band (LB) in said low band synthesis signal, said first high band (HB-1) being situated at higher frequencies than said low band (LB) and said second high band (HB 2) being situated in frequency between said low band (LB) and said first high band (HB-1);

selecting an energy offset from a set of at least two predetermined energy offsets for each of said first subbands based on said parameter defining said used energy offset;

reconstructing a signal in a transform domain by determining a spectrum envelope in said first high band (HB-1) from said first set of quantization indices corresponding to said first subbands, by use of said selected energy offset and said first energy measure, for each of said first subbands of said first high band (HB-1); and

performing an inverse transform based on at least said reconstructed signal in said transform domain into said audio signal;

obtaining said second energy measure as an energy measure of a second reference band within said low band (LB) in said low band synthesis signal; and

wherein said reconstructing said signal in said transform domain further comprises determining a spectrum envelope in said second high band (HB-1) from said second set of quantization indices corresponding to said second subbands by use of said second energy measure for each of said second subbands of said second high band (HB-2).

19. The method of claim 18, wherein said transform encoding is a Modified Discrete Cosine Transform.

20. The method of claim 18, wherein a low frequency end of said first high (HB-1) band is 8 kHz.

21. The method of claim 18, wherein a high frequency end of said first high (HB-1) band is 11.6 kHz.

22. The method of claim 18, wherein said first high (HB-1) band comprises five first subbands.

23. The method of claim 18, wherein said low band (LB) ranges from 0-6.4 kHz.

24. The method of claim 18, wherein said first reference band comprises entire said low band (LB).

25. The method of claim 18, wherein said first reference band ranges from 0-5.9 kHz.

26. The method of claim 18, wherein said low band synthesis signal is based on an encoding by a Code-Excited Linear Prediction encoder.

27. The method of claim 18, wherein the quantization indices of said second set of quantization indices are restricted in increased energy direction.

28. The method of claim 27, wherein said energy restriction of said quantization indices is reduced with increasing frequency of said second subbands.

29. The method of claim 18, wherein said second high band (HB-2) ranges between 6.4 and 8 kHz.

30. The method of claim 18, wherein said second reference band ranges between 5.9 and 6.4 kHz.

31. The method of claim 18, wherein said second high band (HB-2) comprises three second subbands.

32. An encoder apparatus for encoding of an audio signal, comprising:

a transform encoder configured to perform a transform of said audio signal into a transform domain;

a selector configured to select an energy offset from a set of at least two predetermined energy offsets for each of a plurality of first subbands of a first high band (HB-1) of said audio signal in said transform domain;

a synthesizer configured to obtain a low band synthesis signal of an encoding of said audio signal;

an energy reference block, connected to said synthesize and configured to obtain a first energy measure of a first reference band within a low band (LB) in said low band synthesis signal, said first high band HB-1) being situated at higher frequencies than said low band (LB);

an encoder block, connected to said selector and said energy reference block and configured to encode said first high band (HB-1) so as to provide a first set of quantization indices representing a respective scalar quantization of a spectrum envelope in said plurality of first subbands of said first high band (HB-1) relative to said first energy measure, said first set of quantization indices being given with a respective said selected energy offset, and so as to provide a parameter defining the used energy offset;

wherein said energy reference block is further configured to obtain a second energy measure of a second reference band within said low band (LB) of said low band synthesis signal;

wherein said encoder block is further configured to encode a second high band (HB-2) of said audio signal in said transform domain, said second high band (HB-2) being situated in frequency between said low band (LB) and said first high band (HB-1), wherein said encoder block is configured to encode the second high band (HB-2) so as to provide a second set of quantization indices representing a respective scalar quantization of a spectrum envelope in a plurality of second subbands of said second high band (HB-2) relative to said second energy measure.

33. The encoder apparatus of claim 32, wherein said selector is configured to select an energy offset in dependence on a power distribution of said audio signal in a frequency domain.

34. The encoder apparatus of claim 32, wherein said selector is configured to determine a parameter characterizing a power distribution of said low band synthesis signal in a frequency domain and to select an energy offset based on said determined parameter.

35. The encoder apparatus of claim 32, wherein

said encoder block is configured to provide one first set of said quantization indices for each predetermined energy offset range;

and

said selector is configured to receive said first sets of quantization indices for all predetermined energy offset ranges and comprises a calculation block configured to calculate a quantization error for each of said first sets of quantization indices and a selection block configured to select the first set of said quantization indices giving the smallest quantization error.

36. The encoder apparatus of claim 32, wherein said transform encoder is a Modified Discrete Cosine Transform encoder.

37. A network node comprising the encoder apparatus of claim 32.

38. A decoder apparatus for decoding of an audio signal, the decoder apparatus comprising:

an input block configured to receive an encoding of said audio signal, said encoding representing a first set of quantization indices of a spectrum envelope in a plurality of first subbands of a first high band (HB-1) of said audio signal, said first set of quantization indices representing energies relative to a first energy measure, said encoding further representing a parameter defining a used energy offset, said encoding further representing a second set of quantization indices of a spectrum envelope in a plurality of second subbands of a second high band (HB-2) of said audio signal, said second set of quantization indices representing energies relative to a second energy measure;

a synthesizer configured to obtain a low band synthesis signal of an encoding of said audio signal;

an energy reference block, connected to said synthesizer and configured to obtain said first energy measure as an energy measure of a first reference band within a low band (LB) in said low band synthesis signal, said first high band (HB-1) being situated at higher frequencies than said low band (LB) and said second high band (HB-2) being situated in frequency between said low band (LB) and said first high band (HB-1);

a selector, connected to said input block and configured to select an energy offset from a set of at least two predetermined energy offsets for each of said first subbands based on said parameter defining said used energy offset;

a reconstruction block, connected to said input block, said selector, and said energy reference block, and configured to reconstruct a signal in a transform domain by determining a spectrum envelope in said first high band (HB-1) from said first set of quantization indices corresponding to said first subbands, by use of said selected energy offset and said first energy measure, for each of said first subbands of said first high band (HB-1); and

an inverse transform decoder, connected to said reconstruction block and configured to perform an inverse transform based on at least said reconstructed signal in said transform domain into said audio signal;

wherein said energy reference block is further configured to obtain said second energy measure as an energy measure of a second reference band within said low band (LB) of said low band synthesis signal; and

wherein said reconstruction block is further configured to determine a spectrum envelope in said second high band (HB-1) from said second set of quantization indices corresponding to said second subbands by use of said second energy measure for each of said second subbands of said second high band (HB-2).

39. The decoder apparatus of claim 38, wherein said inverse transform decoder is a Modified Discrete Cosine inverse Transform decoder.

40. A network node comprising the decoder apparatus of claim 38.