Method and apparatus for split-band encoding and split-band decoding of audio information using adaptive bit allocation to adjacent subbands

• US 5,752,225 A
• Filed: 06/07/1995
• Issued: 05/12/1998
• Est. Priority Date: 01/27/1989
• Status: Expired due to Fees

• Current Assignee: Dolby Laboratories Licensing Corporation (Dolby Laboratories Incorporated)
• Original Assignee: Dolby Laboratories Licensing Corporation (Dolby Laboratories Incorporated)
• Inventors: Fielder, Louis Dunn
• Primary Examiner(s): Knepper, David D.
• Application Number:
  US08/481,638
• Time in Patent Office: 1,070 Days
• Field of Search: 395/2.1, 395/2.2, 395/2.38, 395/2.12, 395/2.13, 395/2.39, 704/201, 704/211, 704/229, 704/230, 704/203, 704/204
• US Class Current: 704/229
• CPC Class Codes: G06T 9/005 Statistical coding, e.g. Huffman, run length coding
• CPC Class Codes: G10L 19/0212 using orthogonal transformation
• CPC Class Codes: G10L 19/16 Vocoder architecture
• CPC Class Codes: H04B 1/665 using psychoacoustic properties of the ear, e.g. masking effect

Abstract

A split-band encoder and a split-band decoder use adaptive bit allocation to achieve high-quality audio coding at low bit rates. Bits are adaptively allocated to one or two subbands adjacent to a given subband based on signal levels in the given subband.

Claims

1. An encoder for the high-quality digital encoding of wideband audio information, comprising

subband means for receiving said wideband audio information and for generating subband information for a plurality of subbands in response thereto,

quantizing means for forming digital words by adaptively allocating bits to and quantizing said subband information, wherein bits are adaptively allocated to digital words representing a respective subband and one or two adjacent subbands on the basis of a spectral level in said respective subband, and

formatting means for assembling digital information including said digital words into a coded signal having a format suitable for transmission or storage.

2. An encoder according to claim 1 wherein said subband means comprises means for forming time-domain signal sample blocks in response to said wideband audio information and means for generating analysis-window weighted time-domain signal sample blocks by weighting each of said time-domain signal sample blocks with an analysis window, and wherein said subband information comprises blocks of frequency-domain transform coefficients grouped into subbands, said transform coefficients generated by applying a discrete transform to said analysis-window weighted time-domain signal sample blocks.

3. An encoder according to claim 2 wherein said quantizing means further comprises

means for estimating relative spectral level of said subband information in each subband of said plurality of subbands,

means for allocating at most a maximum number of bits to digital words representing a first group of subbands possessing the greatest relative spectral levels and stopping when a certain number of bits has been allocated to each of the digital words representing said first group of subbands, and

means for allocating bits to digital words representing subbands adjacent to subbands in which each of the digital words have been allocated said certain number of bits.

4. An encoder according to claim 3 wherein said certain number of bits is equal to said maximum number of bits.

5. An encoder according to claim 1 wherein said quantizing means further comprises

means for estimating relative spectral level of said subband information in each subband of said plurality of subbands,

means for allocating at most a maximum number of bits to digital words representing a first group of subbands possessing the greatest relative spectral levels and stopping when a certain number of bits has been allocated to each of the digital words representing said first group of subbands, and

means for allocating bits to digital words representing subbands adjacent to subbands in which each of the digital words have been allocated said certain number of bits.

6. An encoder according to claim 5 wherein said certain number of bits is equal to said maximum number of bits.

7. A decoder for the high-quality reproduction of wideband audio information from a coded signal comprising digital words representing subband information for a plurality of subbands of said wideband audio information, said decoder comprising

deformatting means for receiving said coded signal and for deriving said digital words therefrom,

linearization means for generating said subband information by adaptively allocating bits to and dequantizing said digital words, wherein bits are adaptively allocated to digital words representing a respective subband and one or two adjacent subbands on the basis of a spectral level in said respective subband, and

inverse subband means for generating a representation of said wideband audio information in response to said subband information.

8. A decoder according to claim 7 wherein said inverse subband means generates said representation of wideband audio information by applying an inverse discrete transform to said subband information.

9. A decoder according to claim 7 wherein said inverse subband means comprises

means for generating time-domain signal sample blocks by applying an inverse discrete transform to said subband information,

means for weighting said time-domain signal sample blocks by a synthesis window, and

means for cancelling window weighting effects to generate said representation of wideband audio information.

10. A decoder according to claim 9 wherein said linearization means comprises

means for estimating relative spectral level of said subband information in each subband of said plurality of subbands,

means for allocating at most a maximum number of bits to digital words representing a first group of subbands possessing the greatest relative spectral levels and stopping when a certain number of bits has been allocated to each of the digital words representing said first group of subbands, and

means for allocating bits to digital words representing subbands adjacent to subbands in which each of the digital words has been allocated said certain number of bits.

11. A decoder according to claim 10 wherein said certain number of bits is equal to said maximum number of bits.

12. A decoder according to claim 8 wherein said linearization means comprises

means for estimating relative spectral level of said subband information in each subband of said plurality of subbands,

means for allocating at most a maximum number of bits to digital words representing a first group of subbands possessing the greatest relative spectral levels and stopping when a certain number of bits has been allocated to each of the digital words representing said first group of subbands, and

means for allocating bits to digital words representing subbands adjacent to subbands in which each of the digital words has been allocated said certain number of bits.

13. A decoder according to claim 12 wherein said certain number of bits is equal to said maximum number of bits.

14. A decoder according to claim 7 wherein said linearization means comprises

means for estimating relative spectral level of said subband information in each subband of said plurality of subbands,

means for allocating at most a maximum number of bits to digital words representing a first group of subbands possessing the greatest relative spectral levels and stopping when a certain number of bits has been allocated to each of the digital words representing said first group of subbands, and

means for allocating bits to digital words representing subbands adjacent to subbands in which each of the digital words has been allocated said certain number of bits.

15. A decoder according to claim 14 wherein said certain number of bits is equal to said maximum number of bits.

16. A system for the high quality processing of wideband audio information including an encoder and a decoder, said encoder comprising

subband means for receiving said wideband audio information and for generating first subband information for a plurality of subbands in response thereto,

quantizing means for forming digital words by adaptively allocating bits to and quantizing said first subband information, wherein bits are adaptively allocated to digital words representing a respective subband and one or two adjacent subbands on the basis of a spectral level in said respective subband, and

formatting means for assembling digital information including said digital words into a coded signal having a format suitable for transmission or storage,

and said decoder comprising

deformatting means for receiving said coded signal and for deriving said digital words therefrom,

linearization means for generating second subband information by adaptively allocating bits to and dequantizing said digital words, and

inverse subband means for generating a representation of said wideband audio information in response to said second subband information.

17. A system according to claim 16 wherein

said subband means comprises means for forming time-domain signal sample blocks in response to said wideband audio information and means for generating analysis-window weighted time-domain signal sample blocks by weighting each of said time-domain signal sample blocks with an analysis window, and wherein said first subband information comprises blocks of frequency-domain transform coefficients grouped into subbands, said transform coefficients generated by applying a discrete transform to said analysis-window weighted time-domain signal sample blocks, and

wherein said inverse subband means comprises means for generating time-domain signal sample blocks by applying an inverse discrete transform to said second subband information, and means for cancelling window weighting effects to generate said representation of wideband audio information.

18. A system according to claim 16 wherein said quantizing means further comprises

means for estimating relative spectral level of said first subband information in each subband of said plurality of subbands,

means for allocating at most a maximum number of bits to first digital words representing a first group of subbands possessing the greatest relative spectral levels and stopping when a certain number of bits has been allocated to each of the first digital words representing said first group of subbands, and

means for allocating bits to first digital words representing subbands adjacent to subbands in which each of the first digital words have been allocated said certain number of bits.

19. An encoder according to claim 18 wherein said certain number of bits is equal to said maximum number of bits.