Coding of a sound field signal

• US 9,502,046 B2
• Filed: 09/20/2013
• Issued: 11/22/2016
• Est. Priority Date: 09/21/2012
• Status: Active Grant

• Current Assignee: Dolby Laboratories Licensing Corporation (Dolby Laboratories Incorporated)
• Original Assignee: Dolby International AB (Dolby Laboratories Incorporated), Dolby Laboratories Licensing Corporation (Dolby Laboratories Incorporated)
• Inventors: Purnhagen, Heiko, Samuelsson, Leif Jonas, Klejsa, Janusz, Villemoes, Lars, Hirvonen, Toni, Dickins, Glenn N., Cartwright, Richard James
• Primary Examiner(s): ALBERTALLI, BRIAN LOUIS
• Application Number:
  US14/421,132
• Publication Number:
  US 20150221313A1
• Time in Patent Office: 1,159 Days
• US Class Current: 1/1
• CPC Class Codes: G10L 19/008 Multichannel audio signal coding or decoding using interchannel correlation to reduce redundancy, e.g. joint-stereo, intensity-coding or matrixing
• CPC Class Codes: G10L 19/012 Comfort noise or silence coding
• CPC Class Codes: G10L 19/02 using spectral analysis, e.g. transform vocoders or subband vocoders
• CPC Class Codes: G10L 19/0208 Subband vocoders
• CPC Class Codes: G10L 19/032 Quantisation or dequantisation of spectral components
• CPC Class Codes: G10L 19/22 Mode decision, i.e. based on audio signal content versus external parameters
• 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 21/02 Speech enhancement, e.g. noise reduction or echo cancellation reducing echo effects in line transmission systems H04B3/20; echo suppression in hands-free telephones H04M9/08
• CPC Class Codes: G10L 21/0208 Noise filtering
• CPC Class Codes: G10L 21/0216 characterised by the method used for estimating noise
• CPC Class Codes: H04M 3/56 Arrangements for connecting several subscribers to a common circuit, i.e. affording conference facilities video conference systems H04N7/15

Abstract

A method for encoding sound field signals includes allocating coding rate by application of a uniform criterion to all subbands of all signals in a joint process. An allocation criterion may be based on a comparison, in a given subband, between a spectral envelope of the signals to be encoded and a coding noise profile, wherein the noise profile may be a sum of a noise shape and a noise offset, which noise offset is computed on the basis of the coding bit budget. The rate allocation process may be combined with an energy-compacting orthogonal transform, for which there is proposed a parameterization susceptible of efficient coding and having adjustable directivity. In a further aspect, the invention provides a corresponding decoding method.

Claims

1. An adaptive audio encoding device, comprising:

a spatial analyzer configured to receive a plurality of audio signals and to determine, based on the plurality of audio signals, frame-wise decomposition parameters;

an adaptive rotation stage configured to receive said plurality of audio signals and to output at least a first, second, and third rotated audio signal obtained by an energy-compacting orthogonal transformation, wherein quantitative properties of the transformation are determined by the decomposition parameters;

a spectral envelope analyzer configured to receive a frequency-domain representation of the rotated audio signals, which contains transform coefficients, and to output, based thereon, a spectral envelope; and

a multichannel encoder configured to receive the frequency-domain representation of the rotated audio signals and to output transform coefficients of the first rotated audio signal only for frequency subbands in a first subband collection, transform coefficients of the second rotated audio signal only for frequency subbands in a second subband collection, and transform coefficients of the third rotated audio signal only for frequency subbands in a third subband collection, wherein any subbands not included in any subband collection are to be synthesized at decoding,

wherein the multichannel encoder determines the first subband collection, the second subband collection, and the third subband collection by means of a rate allocation process based on a joint comparison of a noise profile for the rotated audio signals and the spectral envelopes of the rotated audio signals,

wherein at least one of the spatial analyser, the adaptive rotation stage, the spectral envelope analyser, and the multichannel encoder, are implemented, at least in part, by one or more hardware elements of the adaptive audio encoding device.

2. The system of claim 1, wherein:

the noise profile is based on a sum of a noise shape, which is fixed for a given frame, and a noise offset common to all rotated audio signals; and

the rate allocation process includes varying the noise offset to approximate a coding bit budget.

3. The system of claim 2, wherein the rate allocation process includes computing a rate expense for a candidate noise offset on the basis of the transform coefficients of the rotated audio signals.

4. The system of claim 2, further comprising a noise level computing section configured to determine said noise shape on the basis of the spectral envelopes of the rotated audio signals, wherein the noise shape is allowed to vary on a subband basis.

5. The system of claim 1, further comprising a rescaling section configured to rescale the spectral envelope in such manner as to make it perceptually comparable to a noise profile, which is constant with respect to frequency.

6. The system of claim 1, wherein the spatial analyzer comprises a decomposition parameter encoder and is configured to supply the decomposition parameters to the adaptive rotation stage in quantized form.

7. The system of claim 1, wherein the spatial analyzer is configured to determine the decomposition parameters on the basis of an analysis frequency subrange of said plurality of audio signals.

8. The system of claim 1, further comprising a time-to-frequency transform stage arranged upstream of the spatial analyzer, said time-to-frequency transform stage being configured to receive a time-domain representation of at least one signal and to output, based thereon, a frequency-domain representation of the at least one signal.

9. The system of claim 8, wherein the time-to-frequency transform stage is arranged immediately upstream of the spatial analyzer.

10. The system of claim 8, wherein the adaptive rotation stage is configured to apply said orthogonal decomposition to a decomposition frequency subrange of said plurality of audio signals,

said system further comprising a combining section configured to form the rotated audio signals by concatenating the result of the orthogonal transformation in the decomposition subrange and said plurality of audio signals outside the decomposition subrange.

11. The system of claim 1, further comprising a time-invariant pre-conditioning stage configured to output said plurality of audio signals based on an equal number of input audio signals.

12. The system of claim 11, wherein the input audio signals are obtainable by three angularly distributed directive transducers and the pre-conditioning stage is configured to output a linear combination of the input audio signals with coefficients proportional to the elements of the following matrix:

13. An adaptive audio encoding method, comprising:

determining frame-wise decomposition parameters on the basis of a plurality of audio signals;

rotating the audio signals into at least a first, second, and third rotated audio signal using an energy-compacting orthogonal transformation, wherein quantitative properties of the transformation are determined by the decomposition parameters;

computing a spectral envelope based on a frequency-domain representation of the rotated audio signals, which frequency-domain representation contains transform coefficients;

determining a first, second, and third subband collection associated with a respective one of the first, second, and third rotated audio signals by means of a rate allocation process based on a joint comparison of a noise profile for all rotated audio signals and spectral envelopes of all rotated audio signals; and

outputting transform coefficients of the first rotated audio signal only for frequency subbands in the first subband collection, transform coefficients of the second rotated audio signal only for frequency subbands in the second subband collection, and transform coefficients of the third rotated audio signal only for frequency subbands in the third subband collection, wherein any subbands not included in any subband collection are to be synthesized at decoding.

14. An adaptive audio decoding device, comprising:

a multichannel decoder configured to receive transform coefficients relating to a plurality of rotated audio signals and to output, based thereon, restored rotated audio signals;

an adaptive rotation inversion stage configured to receive decomposition parameters and the restored rotated audio signals and to output, based thereon, at least three de-rotated audio signals, wherein the adaptive rotation inversion stage computes an inverse of a decomposition matrix on the basis of the decomposition parameters and forms a linear combination of the restored rotated audio signals with coefficients proportional to the elements of said inverse,

wherein at least one of the multichannel decoder and the adaptive rotation inversion stage, are implemented, at least in part, by one or more hardware elements of the adaptive audio decoding device.

15. The system of claim 14, wherein the multichannel decoder is configured to further receive a spectral envelope and a noise offset, and is adapted to detect frequency subbands for which transform coefficients are absent and to synthesize any such subbands in the restored rotated audio signals in accordance with a noise profile derived from the noise offset and the spectral envelope.

16. The system of claim 14, further comprising a time-invariant post-processing stage configured to provide, on the basis of the restored rotated audio signals, an equal number of restored output audio signals.

17. The system of claim 16, configured to produce restored output signals to be rendered by three angularly distributed directive audio sources, wherein the post-processing stage is configured to output signals obtained as a linear combination of the restored rotated signals with coefficients proportional to the elements of the inverse of the following matrix:

18. The system of claim 14, further comprising a frequency-to-time transform stage arranged downstream of the multichannel decoder, said frequency-to-time transform stage being configured to receive a frequency-domain representation of at least one signal and to output, based thereon, a time-domain representation of the at least one signal.

19. The system of claim 18, wherein the frequency-to-time transform stage is arranged downstream of the adaptive rotation inversion stage, which is configured to form said linear combination of only a decomposition frequency subrange of the restored rotated audio signals,

said system further comprising a combining section configured to concatenate the result of the linear combination in the decomposition subrange and the restored rotated audio signals outside the decomposition subrange.

20. The system of claim 18, wherein the frequency-to-time transform stage is arranged downstream of the adaptive rotation inversion stage, which is configured to form said linear combination independently for at least two distinct decomposition frequency subranges based on corresponding values of the decomposition parameters,

said system further comprising a combining section configured to concatenate the result of the linear combination in the decomposition frequency subranges.