Digital signal encoding method and apparatus and recording medium

• US 5,654,952 A
• Filed: 10/25/1995
• Issued: 08/05/1997
• Est. Priority Date: 10/28/1994
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Suzuki, Hiroshi, Takahashi, Kuniaki
• Primary Examiner(s): HUBER, PAUL W
• Application Number:
  US08/547,954
• Time in Patent Office: 650 Days
• Field of Search: 369/47, 369/48, 369/54, 369/86, 369/59, 369/60, 369/124, 360/53, 360/32, 360/48, 395/2.13, 395/2.37, 395/2.38, 381/30, 381/36, 381/37, 381/29, 381/41, 381/34, 341/50, 341/51, 341/54, 341/64
• US Class Current: 369/59.16
• CPC Class Codes: G10L 19/02 using spectral analysis, e.g. transform vocoders or subband vocoders
• CPC Class Codes: G11B 11/10504 Recording for shaping of magnetic domains G11B11/10528, for compensation of shift G11B11/1053
• CPC Class Codes: G11B 20/00007 Time or data compression or expansion audio compression based on psychoacoustics G10L19/00; data processing for reproducing audio data at different playback speeds G10L21/04; video compression H04N19/00; data compression per se H03M7/30
• CPC Class Codes: G11B 20/10527 Audio or video recording; Data buffering arrangements G11B20/12 - G11B20/18 take precedence
• CPC Class Codes: G11B 7/013 for discrete information, i.e. where each information unit is stored in a distinct discrete location , e.g. digital information formats within a data block or sector
• CPC Class Codes: H03M 7/30 Compression speech analysis-synthesis for redundancy reduction G10L19/00; for image communication H04N; Expansion; Suppression of unnecessary data, e.g. redundancy reduction

Abstract

A digital signal compressing device includes a unit for dividing an input signal into plural bands duping information compression, a unit for producing frequency-domain data of the input signal divided into plural bands, and a unit for operating on frequency-domain characteristics. The unit for operating on frequency-domain characteristics operates on the frequency-domain characteristics so that the condition of canceling the aliasing noise by the frequency division and subsequent frequency synthesis performed by a synthesis filter in association with the division at the time of information expansion will be maintained. The frequency-domain characteristics of the input signal can be acted on easily without obstructing the condition of canceling the aliasing noise without requiring a new mechanism or device other than the information compressing mechanism or device.

Claims

1. A method for compressing digital signals comprising the steps of:

dividing an input signal into at least two bands;

obtaining frequency-domain data of said at least two bands of the input signal; and

acting on at least one characteristic of said frequency domain data;

wherein said step of acting on at least one characteristic acts on frequency-domain characteristics and includes;

a multiplication coefficient calculating sub-step of calculating the size of frequency-domain spectral data, allocating the frequency characteristic operating information on said spectral data for normalization and outputting coefficients for multiplication;

an overflow check sub-step of checking whether or not a result of multiplication of said spectral data with the multiplication coefficient supplied from said multiplication coefficient calculating sub-step exceeds a pre-set upper limit;

a coefficient correcting step of normalizing the coefficients from said multiplication coefficient calculating sub-step in their entirety if, judging from an output of said overflow check sub-step, there are a large number of frequency domain spectral data for which occurs the overflow;

a storage step of correcting the coefficients so that an output of said coefficient correction sub-step will exhibit moderate changes;

an aliasing noise reducing sub-step of correcting an output of said coefficient correcting sub-step to characteristics maintaining symmetry with respect to a cut-off frequency as an axis of symmetry for canceling previous aliasing noise; and

a multiplication sub-step of multiplying the coefficients corrected by said aliasing noise reducing sub-step with the frequency-domain spectral data.

2. The method as claimed in claim 1 wherein said step of dividing an input signal into at least two bands uses a QMF filter made up at least of a low-pass filter and a high-pass filter and said step of acting on at least one characteristic affords such characteristics that the equation ##EQU10## is satisfied, an impulse response of the high-pass filter will meet the equation:

##EQU11## and impulse responses g1, g2 of synthesis filters G1, G2 will meet the equation;

##EQU12## where f1 (n) is an impulse response of the low-pass filter and F1(n) its z-transform.

3. The method as claimed in claim 1, wherein the cut-off frequency is fs/4.

4. The method as claimed in claim 1, wherein said step of acting on at least one characteristic achieves pre-set characteristics by changing the condition of canceling the frequency-domain characteristics and the aliasing noise.

5. The method as claimed in claim 4, wherein the cut-off frequency is fs/4.

6. The method as claimed in claim 1, wherein said step of acting on at least one characteristic places priority on operating on the frequency-domain characteristics rather than on the condition of canceling the aliasing noise in the high range.

7. The method as claimed in claim 6, wherein the cut-off frequency is fs/4.

8. A recording medium having recorded thereon signals formed by the digital signal compressing method as claimed in any one of claims 1 to 2.

9. A method for compressing digital signals comprising the steps of:

dividing an input signal into at least two bands;

obtaining frequency-domain data of said at least two bands of the input signal; and

acting on at least one characteristic of said frequency domain data;

wherein said step of acting on at least one characteristic acts on frequency-domain characteristics for maintaining the condition of canceling aliasing noise that results from the frequency division;

wherein said step of acting on at least one characteristic includes;

a multiplication coefficient calculating sub-step of calculating the size of frequency-domain spectral data, allocating the frequency characteristic operating information on said spectral data for normalization and outputting coefficients for multiplication;

an overflow check sub-step of checking whether or not a result of multiplication of said spectral data with the multiplication coefficient supplied from said multiplication coefficient calculating sub-step exceeds a pre-set upper limit;

a coefficient correcting step of normalizing the coefficients from said multiplication coefficient calculating sub-step in their entirety if, judging from an output of said overflow check sub-step, there are a large number of frequency domain spectral data for which occurs the overflow;

a storage step of correcting the coefficients so that an output of said coefficient correction sub-step will exhibit moderate changes;

an aliasing noise reducing sub-step of correcting an output of said coefficient correcting sub-step to characteristics maintaining symmetry with respect to a cut-off frequency fs/4 as an axis of symmetry for canceling previous aliasing noise; and

a multiplication sub-step of multiplying the coefficients corrected by said aliasing noise reducing sub-step with the frequency-domain spectral data and

wherein said step of dividing an input signal into at least two bands uses a QMF filter made up at least of a low-pass filter and a high-pass filter and said step of acting on at least one characteristic affords such characteristics that the equation ##EQU13## is satisfied, an impulse response of the high-pass filter will meet the equation;

##EQU14## and impulse responses g1, g2 of synthesis filters G1, G2 will meet the equation;

##EQU15## where f1 (n) is an impulse response of the low-pass filter and F1(n) its z-transform.

10. The method as claimed in claim 9 wherein said step of acting on at least one characteristic achieves pre-set characteristics by changing the condition of canceling the frequency-domain characteristics and the aliasing noise.

11. The method as claimed in claim 9 wherein said step of acting on at least one characteristic places priority on operating on the frequency-domain characteristics rather than on the condition of canceling the aliasing noise in the high range.

12. A recording medium having recorded thereon signals formed by the digital signal compressing method as claimed in any one of claims 9 to 11.

13. A device for compressing digital signals comprising:

means for dividing an input signal into at least two bands;

means for obtaining frequency-domain data of said at least two bands of the input signal; and

means for acting on at least one characteristic of said frequency-domain data;

wherein said means for acting on at least one characteristic acts on frequency-domain characteristics for maintaining the condition of canceling the aliasing noise that results from the frequency division

wherein said means for acting on at least one characteristic includes;

a multiplication coefficient calculating circuit for calculating the size of frequency-domain spectral data, allocating the frequency characteristic operating information on said spectral data for normalization and outputting coefficients for multiplication;

an overflow check circuit for checking whether or not a result of multiplication of said spectral data with the multiplication coefficient supplied from said multiplication coefficient calculating means exceeds a pre-set upper limit;

a coefficient correcting circuit for normalizing the coefficients from said multiplication coefficient calculating means in their entirety if, judging from an output of said overflow check sub-step, there are a large number of frequency domain spectral data for which occurs the overflow;

a memory for correcting the coefficients so that an output of said coefficient correction means will exhibit moderate changes;

an aliasing noise reducing circuit for correcting an output of said coefficient correcting means to characteristics maintaining symmetry with respect to a cut-off frequency as an axis of symmetry for canceling previous aliasing noise; and

a multiplier for multiplying the coefficients corrected by said aliasing noise reducing means with the frequency-domain spectral data.

14. The device as claimed in claim 13 wherein said means for dividing an input signal into at least two bands uses a QMF filter made up at least of a low-pass filter and a high-pass filter and said means for acting on at least one characteristic affords such characteristics that the equation ##EQU16## is satisfied, an impulse response of the high-pass filter will meet the equation:

##EQU17## and impulse responses g1, g2 of synthesis filters G1, G2 will meet the equation;

##EQU18## where f1(n) is an impulse response of the low-pass filter and F1(n) its z-transform.

15. The device as claimed in claim 13, wherein said cut-off frequency is fs/4.

16. The device as claimed in claim 13, wherein said means for acting on at least one characteristic achieves pre-set characteristics by changing the condition of canceling the frequency-domain characteristics and the aliasing noise.

17. The device as claimed in claim 16, wherein said cut-off frequency is fs/4.

18. The device as claimed in claim 13, wherein said means for acting on at least one characteristic places priority on operating on the frequency-domain characteristics rather than on the condition of canceling the aliasing noise in the high range.

19. The device as claimed in claim 18, wherein said cut-off frequency is fs/4.

20. A device for compressing digital signals comprising:

means for dividing an input signal into at least two bands;

means for obtaining frequency-domain data of said at least two bands of the input signal; and

means for acting on at least one characteristic of said frequency-domain data;

wherein said means for acting on at least one characteristic acts on frequency-domain characteristics for maintaining the condition of canceling aliasing noise by the frequency division,

wherein said means for dividing an input signal into at least two bands uses a QMF filter made up at least of a low-pass filter and a high-pass filter and said means for acting on at least one characteristic affords such characteristics that the equation ##EQU19## is satisfied, an impulse response of the high-pass filter will meet the equation;

##EQU20## and impulse responses g1, g2 of synthesis filters G1, G2 will meet the equation;

##EQU21## where f1(n) is an impulse response of the low-pass filter and F1(n) its z-transform; and

wherein said means for acting on at least one characteristic includes;

a multiplication coefficient calculating circuit for calculating the size of frequency-domain spectral data, allocating the frequency characteristic operating information on said spectral data for normalization and outputting coefficients for multiplication;

an overflow check circuit for checking whether or not a result of multiplication of said spectral data with the multiplication coefficient supplied from said multiplication coefficient calculating means exceeds a pre-set upper limit;

a coefficient correcting circuit for normalizing the coefficients from said multiplication coefficient calculating means in their entirety if, judging from an output of said overflow check sub-step, there are a large number of frequency domain spectral data for which occurs the overflow;

a memory for correcting the coefficients so that an output of said coefficient correction means will exhibit moderate changes;

an aliasing noise reducing circuit for correcting an output of said coefficient correcting means to characteristics maintaining symmetry with respect to a cut-off frequency as an axis of symmetry for canceling previous aliasing noise; and

a multiplier for multiplying the coefficients corrected by said aliasing noise reducing means with the frequency-domain spectral data.

21. The device as claimed in claim 20 wherein said means for acting on at least one characteristic achieves pre-set characteristics by changing the condition of canceling the frequency-domain characteristics and the aliasing noise.

22. The device as claimed in claim 21 wherein said means for acting on at least one characteristic places priority on operating on the frequency-domain characteristics rather than on the condition of canceling the aliasing noise in the high range.