Frequency domain parameter sequence generating method, encoding method, decoding method, frequency domain parameter sequence generating apparatus, encoding apparatus, decoding apparatus, program, and recording medium

• US 10,332,533 B2
• Filed: 02/16/2015
• Issued: 06/25/2019
• Est. Priority Date: 04/24/2014
• Status: Active Grant

• Current Assignee: Nippon Telegraph and Telephone Corporation, University of Tokyo
• Original Assignee: Nippon Telegraph and Telephone Corporation
• Inventors: Moriya, Takehiro, Kamamoto, Yutaka, Harada, Noboru, Kameoka, Hirokazu, Sugiura, Ryosuke
• Primary Examiner(s): Riley, Marcus T
• Application Number:
  US15/302,094
• Publication Number:
  US 20170249947A1
• Time in Patent Office: 1,590 Days
• Field of Search: None
• CPC Class Codes: G10L 19/02 using spectral analysis, e.g. transform vocoders or subband vocoders
• CPC Class Codes: G10L 19/07 Line spectrum pair [LSP] vocoders
• 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 25/06 the extracted parameters being correlation coefficients
• CPC Class Codes: G10L 25/12 the extracted parameters being prediction coefficients

Abstract

The present invention reduces encoding distortion in frequency domain encoding compared to conventional techniques, and obtains LSP parameters that correspond to quantized LSP parameters for the preceding frame and are to be used in time domain encoding from coefficients equivalent to linear prediction coefficients resulting from frequency domain encoding. When p is an integer equal to or greater than 1, a linear prediction coefficient sequence which is obtained by linear prediction analysis of audio signals in a predetermined time segment is represented as a[1], a[2], . . . , a[p], and ω[1], ω[2], . . . , ω[p] are a frequency domain parameter sequence derived from the linear prediction coefficient sequence a[1], a[2], . . . , a[p], an LSP linear transformation unit (300) determines the value of each converted frequency domain parameter ˜ω[i] (i=1, 2, . . . , p) in a converted frequency domain parameter sequence ˜ω[1], ˜ω[2], . . . , ˜ω[p] using the frequency domain parameter sequence ω[1], ω[2], . . . , ω[p] as input, through linear transformation which is based on the relationship of values between ω[i] and one or more frequency domain parameters adjacent to ω[i].

Claims

1. An encoding method, implemented by an encoding apparatus having processing circuitry, comprising:

where p is an integer equal to or greater than 1, γ

is an adjustment factor which is a positive constant equal to or smaller than 1, a linear prediction coefficient sequence which is obtained by linear prediction analysis of audio signals in a predetermined time segment is represented as a[1], a[2], . . . , a[p],

generating, by the processing circuitry, an adjusted linear prediction coefficient sequence a_γ

[1], a_γ

[2], . . . , a_γ

[p] by adjusting the linear prediction coefficient sequence a[1], a[2], . . . , a[p] by calculating a_γ

[i]=a[i]×

γ

ⁱ using the adjustment factor γ

;

generating, by the processing circuitry, an adjusted LSP parameter sequence θ

_γ

[1], θ

_γ

[2], . . . , θ

_γ

[p] using the adjusted linear prediction coefficient sequence a_γ

[1], a_γ

[2], . . . , a_γ

[p];

encoding, by the processing circuitry, the adjusted LSP parameter sequence θ

_γ

[1], θ

_γ

[2], . . . , θ

_γ

[p] to generate adjusted LSP codes and an adjusted quantized LSP parameter sequence ^θ

_γ

[1], ^θ

_γ

[2], . . . , ^θ

_γ

[p] corresponding to the adjusted LSP codes;

with a frequency domain parameter sequence ω

[1], ω

[2], . . . , ω

[p] being the adjusted quantized LSP parameter sequence ^θ

_γ

[1], ^θ

_γ

[2], . . . , ^θ

_γ

[p], determining, by the processing circuitry, a converted frequency domain parameter sequence ˜

ω

[1], ˜

ω

[2], . . . , ˜

ω

[p] using the frequency domain parameter sequence ω

[1], ω

[2], . . . , ω

[p] as input to thereby generate the converted frequency domain parameter sequence ˜

ω

[1], ˜

ω

[2], . . . , ˜

ω

[p] as an approximate quantized LSP parameter sequence ^θ

_app[1], ^θ

_app[2], . . . , ^θ

_app[p];

generating, by the processing circuitry, an adjusted quantized linear prediction coefficient sequence ^a_γ

[1], ^a_γ

[2], . . . , ^a_γ

[p] by converting the adjusted quantized LSP parameter sequence ^θ

_γ

[1], ^θ

_γ

[2], . . . , ^θ

_γ

[p] into linear prediction coefficients;

calculating, by the processing circuitry, a quantized smoothed power spectral envelope series ^W_γ

[1], ^W_γ

[2], . . . , ^W_γ

[N] which is a series in frequency domain corresponding to the adjusted quantized linear prediction coefficient sequence ^a_γ

[1], ^a_γ

[2], . . . , ^a_γ

[p];

generating, by the processing circuitry, frequency domain signal codes by encoding a frequency domain sample sequence X[1], X[2], . . . , X[N] corresponding to the audio signals using the quantized smoothed power spectral envelope series ^W_γ

[1], ^W_γ

[2], . . . , ^W_γ

[N];

generating, by the processing circuitry, an LSP parameter sequence θ

[1], θ

[2], . . . , θ

[p] using the linear prediction coefficient sequence a[1], a[2], . . . , a[p];

encoding, by the processing circuitry, the LSP parameter sequence θ

[1], θ

[2], . . . , θ

[p] to generate LSP codes and a quantized LSP parameter sequence ^θ

[1], ^θ

[2], . . . , ^θ

[p] corresponding to the LSP codes; and

encoding, by the processing circuitry, the audio signals to generate time domain signal codes using either the generated quantized LSP parameter sequence for a preceding time segment or the generated approximate quantized LSP parameter sequence for the preceding time segment, and the quantized LSP parameter sequence for the predetermined time segment,

wherein

the processing circuitry determines a value of each converted frequency domain parameter ˜

ω

[i] (i=1, 2, . . . , p) in the converted frequency domain parameter sequence ˜

ω

[1], ˜

ω

[2], . . . , ˜

ω

[p] through linear transformation which is based on a relationship of values between ω

[i] and one or more frequency domain parameters adjacent to ω

[i].

3. The encoding method according to claim 1 or 2, further comprising:

outputting, by the processing circuitry, either the generated frequency domain signal codes or the generated time domain signal codes, wherein

when encoding, by the processing circuitry, the audio signals to generate the time domain signal codes, the method further includes

when frequency domain signal codes have been output for the preceding time segment, encoding, by the processing circuitry, that uses the generated approximate quantized LSP parameter sequence for the preceding time segment is performed, and

when time domain signal codes have been output for the preceding time segment, encoding, by the processing circuitry, that uses the generated quantized LSP parameter sequence for the preceding time segment is performed.

6. The encoding method according to claim 1 or 2, wherein

γ

1=γ and

γ

2=1, and K is a predetermined p×

p band matrix in which diagonal elements and elements that neighbor the diagonal elements in row direction have non-zero values,

the processing circuitry generates the converted frequency domain parameter sequence ˜

ω

[1], ˜

ω

[2], . . . , ˜

ω

[p] defined by a following formula

7. The encoding method according to claim 6, wherein

the band matrix K has positive values in the diagonal elements and negative values in elements that neighbor the diagonal elements in row direction.

11. A non-transitory computer-readable recording medium having a program recorded thereon for causing a computer to carry out the steps of the encoding method according to claim 1 or 2.

2. An encoding method, implemented by an encoding apparatus having processing circuitry, comprising:

where p is an integer equal to or greater than 1, γ

is an adjustment factor which is a positive constant equal to or smaller than 1, a linear prediction coefficient sequence which is obtained by linear prediction analysis of audio signals in a predetermined time segment is represented as a[1], a[2], . . . , a[p],

generating, by the processing circuitry, an adjusted linear prediction coefficient sequence a_γ

[1], a_γ

[2], . . . , a_γ

[p] by adjusting the linear prediction coefficient sequence a[1], a[2], . . . , a[p] by calculating a_γ

[i]=a[i]×

γ

ⁱ using the adjustment factor γ

;

generating, by the processing circuitry, an adjusted LSP parameter sequence θ

_γ

[1], θ

_γ

[2], . . . , θ

_γ

[p] using the adjusted linear prediction coefficient sequence a_γ

[1], a_γ

[2], . . . , a_γ

[p];

encoding, by the processing circuitry, the adjusted LSP parameter sequence θ

_γ

[1], θ

_γ

[2], . . . , θ

_γ

[p] to generate adjusted LSP codes and an adjusted quantized LSP parameter sequence ^θ

_γ

[1], ^θ

_γ

[2], . . . , ^θ

_γ

[p] corresponding to the adjusted LSP codes;

with a frequency domain parameter sequence ω

[1], ω

[2], . . . , ω

[p] being the adjusted quantized LSP parameter sequence ^θ

_γ

[1], ^θ

_γ

[2], . . . , ^θ

_γ

[p], determining, by the processing circuitry, a converted frequency domain parameter sequence ˜

ω

[1], ˜

ω

[2], . . . , ˜

ω

[p] using the frequency domain parameter sequence ω

[1], ω

[2], . . . , ω

[p] as input to thereby generate the converted frequency domain parameter sequence ˜

ω

[1], ˜

ω

[2], . . . , ˜

ω

[p] as an approximate quantized LSP parameter sequence ^θ

_app[1], ^θ

_app[2], . . . , ^θ

_app[p];

calculating, by the processing circuitry, a quantized smoothed power spectral envelope series ^W_γ

[1], ^W_γ

[2], . . . , ^W_γ

[N] based on the adjusted quantized LSP parameter sequence ^θ

_γ

[1], ^θ

_γ

[2], . . . , ^θ

_γ

[p];

generating, by the processing circuitry, frequency domain signal codes by encoding a frequency domain sample sequence X[1], X[2], X[N] corresponding to the audio signals using the quantized smoothed power spectral envelope series ^W_γ

[1], ^W_γ

[2], . . . , ^W_γ

[N];

generating, by the processing circuitry, an LSP parameter sequence θ

[1], θ

[2], . . . , θ

[p] using the linear prediction coefficient sequence a[1], a[2], . . . , a[p];

encoding, by the processing circuitry, the LSP parameter sequence θ

[1], θ

[2], . . . , θ

[p] to generate LSP codes and a quantized LSP parameter sequence ^θ

[1], ^θ

[2], . . . , ^θ

[p] corresponding to the LSP codes; and

encoding, by the processing circuitry, the audio signals to generate time domain signal codes using either the generated quantized LSP parameter sequence obtained in the LSP encoding step for a preceding time segment or an approximate quantized LSP parameter sequence obtained in the LSP linear transformation step for the preceding time segment, and the quantized LSP parameter sequence for the predetermined time segment,

wherein

the processing circuitry determines a value of each converted frequency domain parameter ˜

ω

[i] (i=1, 2, . . . , p) in the converted frequency domain parameter sequence ˜

ω

[1], ˜

ω

[2], . . . , ˜

ω

[p] through linear transformation which is based on a relationship of values between ω

[i] and one or more frequency domain parameters adjacent to ω

[i].

4. An encoding apparatus comprising:

where p is an integer equal to or greater than 1, γ

is an adjustment factor which is a positive constant equal to or smaller than 1, a linear prediction coefficient sequence which is obtained by linear prediction analysis of audio signals in a predetermined time segment is represented as a[1], a[2], . . . , a[p],

processing circuitry configured to implement

a linear prediction coefficient adjusting unit that generates an adjusted linear prediction coefficient sequence a_γ

[1], a_γ

[2], . . . , a_γ

[p] by adjusting the linear prediction coefficient sequence a[1], a[2], . . . , a[p] by calculating a_γ

[i]=a[i]×

γ

ⁱ using the adjustment factor γ

;

an adjusted LSP generating unit that generates an adjusted LSP parameter sequence θ

_γ

[1], θ

_γ

[2], . . . , θ

_γ

[p] using the adjusted linear prediction coefficient sequence a_γ

[1], a_γ

[2], . . . , a_γ

[p];

an adjusted LSP encoding unit that encodes the adjusted LSP parameter sequence θ

_γ

[1], θ

_γ

[2], . . . , θ

_γ

[p] to generate adjusted LSP codes and an adjusted quantized LSP parameter sequence ^θ

_γ

[1], ^θ

_γ

[2], . . . , ^θ

_γ

[p] corresponding to the adjusted LSP codes;

an LSP linear transformation unit that, with a frequency domain parameter sequence ω

[1], ω

[2], . . . , ω

[p] being the adjusted quantized LSP parameter sequence ^θ

_γ

[1], ^θ

_γ

[2], . . . , ^θ

_γ

[p], executes a parameter sequence converting unit that determines a converted frequency domain parameter sequence ˜

ω

[1], ˜

ω

[2], . . . , ˜

ω

[p] using the frequency domain parameter sequence ω

[1], ω

[2], . . . , ω

[p] as input to thereby generate the converted frequency domain parameter sequence ˜

ω

[1], ˜

ω

[2], . . . , ˜

ω

[p] as an approximate quantized LSP parameter sequence ^θ

_app[1], ^θ

_app[2], ^θ

_app[p];

a quantized linear prediction coefficient sequence generating unit that generates an adjusted quantized linear prediction coefficient sequence ^a_γ

[1], ^a_γ

[2], . . . , ^a_γ

[p] by converting the adjusted quantized LSP parameter sequence ^θ

_γ

[1], ^θ

_γ

[2], . . . , ^θ

_γ

[p] into linear prediction coefficients;

a quantized smoothed power spectral envelope series calculating unit that calculates a quantized smoothed power spectral envelope series ^W_γ

[1], ^W_γ

[2], . . . , ^W_γ

[N] which is a series in frequency domain corresponding to the adjusted quantized linear prediction coefficient sequence ^a_γ

[1], ^a_γ

[2], . . . , ^a_γ

[p];

a frequency domain encoding unit that generates frequency domain signal codes by encoding a frequency domain sample sequence X[1], X[2], . . . , X[N] corresponding to the audio signals using the quantized smoothed power spectral envelope series ^W_γ

[1], ^W_γ

[2], . . . , W_γ

[N];

an LSP generating unit that generates an LSP parameter sequence θ

[1], θ

[2], . . . , θ

[p] using the linear prediction coefficient sequence a[1], a[2], . . . , a[p];

an LSP encoding unit that encodes the LSP parameter sequence θ

[1], θ

[2], . . . , θ

[p] to generate LSP codes and a quantized LSP parameter sequence ^θ

[1], ^θ

[2], . . . , ^θ

[p] corresponding to the LSP codes; and

a time domain encoding unit that encodes the audio signals to generate time domain signal codes using either the quantized LSP parameter sequence obtained in the LSP encoding unit for a preceding time segment or the approximate quantized LSP parameter sequence obtained in the LSP linear transformation unit for the preceding time segment, and the quantized LSP parameter sequence for the predetermined time segment,

wherein

the parameter sequence conversion unit determines a value of each converted frequency domain parameter ˜

ω

[i] (i=1, 2, . . . , p) in the converted frequency domain parameter sequence ˜

ω

[1], ˜

ω

[2], . . . , ˜

ω

[p] through linear transformation which is based on a relationship of values between a [i] and one or more frequency domain parameters adjacent to ω

[i].

8. The encoding apparatus according to claim 4 or 5, the processing circuitry being further configured to implement:

an output unit that outputs either the frequency domain signal codes generated in the frequency domain encoding unit or the time domain signal codes generated in the time domain encoding unit, wherein

the time domain encoding unit,

when frequency domain signal codes have been output in the output unit for the preceding time segment, encodes that uses the approximate quantized LSP parameter sequence obtained in the LSP linear transformation unit for the preceding time segment is performed, and

when time domain signal codes have been output in the output unit for the preceding time segment, encodes that uses the quantized LSP parameter sequence obtained in the LSP generation unit for the preceding time segment is performed.

9. The encoding apparatus according to claim 4 or 5, wherein

γ

1=γ and

γ

2=1, and K is a predetermined p×

p band matrix in which diagonal elements and elements that neighbor the diagonal elements in row direction have non-zero values,

the parameter sequence conversion unit generates the converted frequency domain parameter sequence ˜

ω

[1], ˜

ω

[2], . . . , ˜

ω

[p] defined by a following formula

10. The encoding apparatus according to claim 9, wherein

the band matrix K has positive values in the diagonal elements and negative values in elements that neighbor the diagonal elements in row direction.

5. An encoding apparatus comprising:

where p is an integer equal to or greater than 1, γ

is an adjustment factor which is a positive constant equal to or smaller than 1, a linear prediction coefficient sequence which is obtained by linear prediction analysis of audio signals in a predetermined time segment is represented as a[1], a[2], . . . , a[p],

processing circuitry configured to implement

a linear prediction coefficient adjusting unit that generates an adjusted linear prediction coefficient sequence a_γ

[1], a_γ

[2], . . . , a_γ

[p] by adjusting the linear prediction coefficient sequence a[1], a[2], a[p] by calculating a_γ

[i]=a[i]×

γ

ⁱ using the adjustment factor γ

;

an adjusted LSP generating unit that generates an adjusted LSP parameter sequence θ

_γ

[1], θ

_γ

[2], . . . , θ

_γ

[p] using the adjusted linear prediction coefficient sequence a_γ

[1], a_γ

[2], . . . , a_γ

[p],

an adjusted LSP encoding unit that encodes the adjusted LSP parameter sequence θ

₁[1], θ

_γ

[2], . . . , θ

_γ

[p] to generate adjusted LSP codes and an adjusted quantized LSP parameter sequence ^θ

_γ

[1], ^θ

_γ

[2], . . . , ^θ

_γ

[p] which is determined by quantization of values in the adjusted LSP parameter sequence corresponding to the adjusted LSP codes;

an LSP linear transformation unit that, with a frequency domain parameter sequence ω

[1], ω

[2], . . . , ω

[p] being the adjusted quantized LSP parameter sequence ^θ

_γ

[1], ^θ

_γ

[2], . . . , ^θ

_γ

[p], executes a parameter sequence converting unit that determines a converted frequency domain parameter sequence ˜

ω

[1], ˜

ω

[2], . . . , ˜

ω

[p] using the frequency domain parameter sequence ω

[1], ω

[2], . . . , ω

[p] as input to thereby generate the converted frequency domain parameter sequence ˜

ω

[1], ˜

ω

[2], . . . , ˜

ω

[p] as an approximate quantized LSP parameter sequence ^θ

_app[1], ^θ

_app[2], . . . , ^θ

_app[p];

a quantized smoothed power spectral envelope series calculating unit that calculates a quantized smoothed power spectral envelope series ^W_γ

[1], ^W_γ

[2], . . . , ^W_γ

[N] based on the adjusted quantized LSP parameter sequence ^θ

_γ

[1], ^θ

_γ

[2], . . . , ^θ

_γ

[p];

a frequency domain encoding unit that generates frequency domain signal codes by encoding a frequency domain sample sequence X[1], X[2], . . . , X[N] corresponding to the audio signals using the quantized smoothed power spectral envelope series ^W_γ

[1], ^W_γ

[2], . . . , ^W_γ

[N];

an LSP generating unit that generates an LSP parameter sequence θ

[1], θ

[2], . . . , θ

[p] using the linear prediction coefficient sequence a[1], a[2], . . . , a[p];

an LSP encoding unit that encodes the LSP parameter sequence θ

[1], θ

[2], . . . , θ

[p] to generate LSP codes and a quantized LSP parameter sequence ^θ

[1], ^θ

[2], . . . , ^θ

[p] corresponding to the LSP codes; and

a time domain encoding unit that encodes the audio signals to generate time domain signal codes using either the quantized LSP parameter sequence obtained in the LSP encoding unit for a preceding time segment or the approximate quantized LSP parameter sequence obtained in the LSP linear transformation unit for the preceding time segment, and the quantized LSP parameter sequence for the predetermined time segment,

wherein

the parameter sequence conversion unit determines a value of each converted frequency domain parameter ˜

ω

[i] (i=1, 2, . . . , p) in the converted frequency domain parameter sequence ˜

ω

[1], ˜

ω

[2], . . . , ˜

ω

[p] through linear transformation which is based on a relationship of values between ω

[i] and one or more frequency domain parameters adjacent to ω

[i].