Processing in the encoded domain of an audio signal encoded by ADPCM coding

Patent metadata

Field	Value
Publication number	US-9990932-B2
Application number	US-201214008435-A
Country	US
Kind code	B2
Filing date	Mar 27, 2012
Priority date	Mar 29, 2011
Publication date	Jun 5, 2018
Grant date	Jun 5, 2018

Abstract

A method for processing an encoded audio signal in a binary stream by MICDA predictive coding. The method includes the following steps: determining a signal assessed from quantification indices of the binary stream; determining unencoded parameters representative of the audio signal from the assessed signal; and processing the encoded audio signal using the determined parameters. Also provided is a device implementing the method.

First claim

The invention claimed is:

1. A method for processing a coded audio signal, coded as a bitstream by a predictive coding of ADPCM type, wherein the method comprises the following acts performed by an audio signal processing device:
receiving a bitstream of a coded audio signal from a coding device of ADPCM type that has coded the audio signal into the bitstream;
generating by the audio signal processing device an estimated signal on the basis of quantization indices of the bitstream, the generating being carried out for a current sample by:
obtaining an adaptation parameter associated with the quantization index for a previous sample; and
calculating the estimated signal for the current sample on the basis of the adaptation parameter obtained, of the signal estimated for the previous sample and of a predefined forget factor;

generating by the audio signal processing device uncoded parameters representative of said audio signal, on the basis of said estimated signal;
processing said coded audio signal by the processing device, comprising modifying at least one zone of the coded audio signal using the uncoded parameters to produce a processed bitstream, without decoding the coded audio signal; and
outputting the processed bit stream to a decoding device.

2. The method as claimed in claim 1 , wherein the adaptation parameter is defined in a prior act, in such a way that its value complies with the property of monotonic increase with the quantized amplitude values corresponding to the quantization indices.

3. The method as claimed in claim 1 , wherein the adaptation parameter is obtained in real time by simple binary shifting and subtraction on the basis of the quantization index.

4. The method as claimed in claim 2 , wherein the estimated signal is generated according to the equation:
se ( n )= V*se ( n− 1)+ M ( I ( n− 1))

where se (n) represents the signal estimated for the current sample, se(n−1) the estimated signal of the previous sample, M(I(n−1)) the adaptation parameter as a function of the quantization index of the previous sample and V a forget factor with a variable less than 1.

5. The method as claimed in claim 1 , wherein the uncoded parameters are generated on the basis of the signal estimated by an analysis step and form part of parameters belonging to the group consisting of:
the energy of the audio signal;
the periodicity or otherwise of the audio signal;
the classification of the audio signal;
the length of the fundamental period of the audio signal;
the temporal position of the glottal pulses;
the vocal activity.

6. The method as claimed in claim 1 , wherein the processing of the coded audio signal comprises an analysis of a periodicity parameter obtained from the estimated signal for discriminating voice of speakers generating the audio signal and in the case of detection of an undesirable voices the modifying comprises reducing the undesirable voice to silence in the coded domain.

7. The method as claimed in claim 1 , wherein the processing of the coded audio signal comprises a detection of vocal inactivity and the modifying comprises replacement of a corresponding zone of the bitstream by a bitstream obtained by coding of a comfort noise.

8. A device for processing a coded audio signal, coded as a bitstream by a predictive coding of ADPCM type, wherein the device comprises:
an input receiving a bitstream of a coded audio signal from a coding device of ADPCM type that has coded the audio signal into the bitstream;
a module configured to generate an estimated signal on the basis of quantization indices of the bitstream, by implementing the following acts, for a current sample:
obtaining an adaptation parameter associated with the quantization index for the previous sample; and
calculation of the estimated signal for the current sample on the basis of the adaptation parameter obtained, of the signal estimated for the previous sample and of a predefined forget factor;

an analysis module configured to generate uncoded parameters representative of said audio signal, on the basis of said estimated signal; and
a module configured to process said coded audio signal, comprising modifying at least one zone of the coded audio signal using the uncoded parameters to produce a processed bit stream without decoding the coded audio signal; and
an output providing the processed bitstream to a decoding device.

9. A communication gateway comprising the device for processing as claimed in claim 8 .

10. A storage device comprising a computer program stored thereon and comprising code instructions for implementation of acts of a method for processing a coded audio signal, coded as a bitstream by a predictive coding of ADPCM type, when these instructions are executed by a processor, wherein the method comprises the following acts:
receiving a bitstream of a coded audio signal from a coding device of ADPCM type that has coded the audio signal into the bitstream;
generating an estimated signal on the basis of quantization indices of the bitstream, the generating being carried out for a current sample by:
obtaining an adaptation parameter associated with the quantization index for a previous sample; and
calculating the estimated signal for the current sample on the basis of the adaptation parameter obtained, of the signal estimated for the previous sample and of a predefined forget factor;

generating uncoded parameters representative of said audio signal, on the basis of said estimated signal;
processing said coded audio signal by the processing device, comprising modifying at least one zone of the coded audio signal using the uncoded parameters to produce a processed bitstream without decoding the coded audio signal; and
outputting the processed bit stream to a decoding device.

Assignees

• Cormier Adrien
• Kovesi Balazs
• Lamblin Claude
• Orange

Inventors

• Cormier Adrien
• Kovesi Balazs
• Lamblin Claude

Classifications

• G10L25/93

  Discriminating between voiced and unvoiced parts of speech signals (G10L25/90 takes precedence) · CPC title

• G10L25/90

  Pitch determination of speech signals · CPC title

• H03M3/042

  with adaptable step size, e.g. adaptive differential pulse code modulation [ADPCM] · CPC title

• G10L19/04 Primary

  using predictive techniques · CPC title