Apparatus and method for data compression using signal-weighted quantizing bit allocation

• US 5,301,205 A
• Filed: 01/29/1993
• Issued: 04/05/1994
• Est. Priority Date: 01/29/1992
• Status: Expired due to Term

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Akagiri, Kenzo, Tsutsui, Kyoya
• Primary Examiner(s): Cain, David C.
• Application Number:
  US08/011,376
• Time in Patent Office: 431 Days
• Field of Search: 375/1, 380/46
• US Class Current: 375/340
• 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: H04B 1/665 using psychoacoustic properties of the ear, e.g. masking effect

Abstract

A data compression apparatus for compressing a digital input signal to provide a compressed digital output signal. In the apparatus, a circuit derives plural spectral coefficients from the digital input signal, and groups the spectral coefficients into bands. An adaptive bit allocation circuit adaptively allocates a number of spectrum-dependent quantizing bits among the bands to allocate to each band a number of spectrum-dependent quantizing bits for quantizing each of the spectral coefficients in the band. The number of spectrum-dependent quantizing bits allocated to each band by the adaptive bit allocation circuit is determined according to the band magnitude, weighted depending on the band frequency.

Claims

1. An apparatus for compressing a digital input signal to provide a compressed digital output signal, the apparatus comprising:

a means for deriving plural spectral coefficients from the digital input signal, and for grouping the spectral coefficients into bands, each band having a band magnitude and a band frequency; and

an adaptive bit allocation means for adaptively allocating a number of spectrum-dependent quantizing bits among the bands to allocate to each band a number of spectrum-dependent quantizing bits for quantizing each spectral coefficient in the band, the number of spectrum-dependent quantizing bits allocated to each band being determined according to the band magnitude, weighted depending on the band frequency.

2. The apparatus of claim 1, wherein the adaptive bit allocation means allocates to each band a number of spectrum-dependent quantizing bits determined according to the band magnitude, the band magnitude being weighted depending on the band frequency such that the number of spectrum-dependent quantizing bits allocated to lower frequency bands for a given band magnitude is increased.

3. The apparatus of claim 1, wherein:

the apparatus additionally comprises;

a means for determining, for each band, the band magnitude in response to the spectral coefficients in the band, and

a weighting means for weighting, for each band, the band magnitude depending on the band frequency to provide a weighted band magnitude; and

the adaptive bit allocation means allocates to each band the number of spectrum-dependent quantizing bits for quantizing each spectral coefficient in the band in response to the weighted band magnitude for the band.

4. The apparatus of claim 1, wherein the adaptive bit allocation means allocates to each band a number of spectrum-dependent quantizing bits determined according to the band magnitude weighted by a weighting coefficient dependent on the digital input signal and the band frequency.

5. The apparatus of claim 1, wherein:

the adaptive bit allocation means allocates to each band a number of spectrum-dependent quantizing bits determined according to the band magnitude weighted depending on the band frequency according to a selected frequency-dependent weighting pattern; and

the apparatus additionally comprises;

a means for storing plural frequency-dependent weighting patterns, and

a selecting means for selecting, in response to the digital input signal, and for providing to the adaptive bit allocation means as the selected frequency-dependent weighting pattern, an appropriate one of the plural frequency-dependent weighting patterns.

6. The apparatus of claim 5, wherein:

the digital input signal has a level; and

the selecting means is for selecting the appropriate one of the plural frequency-dependent weighting patterns in response to the level of the digital input signal.

7. The apparatus of claim 5, wherein:

the digital input signal has a spectrum; and

the selecting means is for selecting the appropriate one of the plural frequency-dependent weighting patterns in response to the spectrum of the digital input signal.

8. The apparatus of claim 5, wherein the adaptive bit allocation means allocates to each band a number of spectrum-dependent quantizing bits determined according to the band magnitude, the band magnitude being weighted depending on the band frequency according to the selected frequency-dependent weighting pattern such that the number of spectrum-dependent quantizing bits allocated to lower frequency bands for a given band magnitude is increased.

9. The apparatus of claim 1, additionally comprising:

a level-dependent bit allocation means for allocating a number of level-dependent quantizing bits among the bands according to a predetermined bit allocation pattern defining, for each band, a number of level-dependent quantizing bits allocated for quantizing each spectral coefficient in the band; and

a quantizing means for quantizing each spectral coefficient in a band using a number of quantizing bits equal to the sum of the number of spectrum-dependent quantizing bits allocated for quantizing each spectral coefficient in the band and the number of level-dependent quantizing bits allocated for quantizing each spectral coefficient in the band.

10. The apparatus of claim 9, wherein the adaptive bit allocation means allocates to each band a number of spectrum-dependent quantizing bits determined according to the band magnitude, the band magnitude being weighted depending on the band frequency such that the number of spectrum-dependent quantizing bits allocated to lower frequency bands for a given band magnitude is increased.

11. The apparatus of claim 9, wherein:

the apparatus additionally comprises;

a means for determining, for each band, the band magnitude in response to the spectral coefficients in the band, and

a weighting means for weighting, for each band, the band magnitude depending on the band frequency to provide a weighted band magnitude; and

the adaptive bit allocation means allocates to each band the number of spectrum-dependent quantizing bits for quantizing each spectral coefficient in the band in response to the weighted band magnitude.

12. The apparatus of claim 9, wherein the adaptive bit allocation means allocates to each band a number of spectrum-dependent quantizing bits determined according to the band magnitude weighted by a weighting coefficient dependent on the digital input signal and the band frequency.

13. The apparatus of claim 9, wherein:

the adaptive bit allocation means allocates to each band a number of spectrum-dependent quantizing bits determined according to the band magnitude weighted depending on the band frequency according to a selected frequency-dependent weighting pattern; and

the apparatus additionally comprises;

a means for storing plural frequency-dependent weighting patterns, and

a selecting means for selecting, in response to the digital input signal, and for providing to the adaptive bit allocation means as the selected frequency-dependent weighting pattern, an appropriate one of the plural frequency-dependent weighting patterns.

14. The apparatus of claim 13, wherein the adaptive bit allocation means allocates to each band a number of spectrum-dependent quantizing bits determined according to the band magnitude weighted depending on the band frequency such that the number of spectrum-dependent quantizing bits allocated to lower frequency bands for a given band magnitude is increased.

15. The apparatus of claim 9, wherein:

the level-dependent bit allocation means is for allocating level-dependent quantizing bits among the bands according to a selected bit allocation pattern; and

the apparatus additionally comprises;

a means for storing plural predetermined bit allocation patterns, and

an additional selecting means for selecting, in response to the digital input signal, and for providing to the level-dependent bit allocating means as the selected bit allocation pattern, an appropriate one of the plural predetermined bit allocation patterns.

16. The apparatus of claim 9, wherein:

a total number of available quantizing bits is available for quantizing all the spectral coefficients;

the digital input signal has a spectrum, and the spectrum has a smoothness;

the apparatus additionally comprises;

a division ratio determining means for determining a division of the total number of available quantizing bits between the number of spectrum-dependent quantizing bits and the number of level-dependent quantizing bits in response to the smoothness of the spectrum of the digital input signal, and for providing a division ratio between the number of level-dependent quantizing bits and the total number of available quantizing bits, and a complement of the division ratio,

a means for multiplying the number of spectrum-dependent quantizing bits for each band by the complement of the division ratio, and

a means for multiplying the number of level-dependent quantizing bits for each band by the division ratio; and

the quantizing means is for quantizing each spectral coefficient in a band using a number of quantizing bits equal to the sum of the multiplied number of spectrum-dependent quantizing bits allocated for quantizing each spectral coefficient in the band and the multiplied number of level-dependent quantizing bits allocated for quantizing each spectral coefficient in the band.

17. The apparatus of claim 9, wherein:

a total number of available quantizing bits is available for quantizing all the spectral coefficients;

the digital input signal has a spectrum, and the spectrum has a smoothness;

the apparatus additionally comprises a division ratio determining means for determining a division of the total number of available quantizing bits between the number of spectrum-dependent quantizing bits and the number of level-dependent quantizing bits in response to the smoothness of the spectrum of the digital input signal, and for providing a division ratio between the number of level-dependent quantizing bits and the total number of available quantizing bits, and a complement of the division ratio;

the number of spectrum-dependent quantizing bits allocated among the bands by the adaptive bit allocation means is substantially proportional to the complement of the division ratio; and

the number of level-dependent quantizing bits allocated among the bands by the fixed bit allocation means is substantially proportional to the division ratio.

18. The apparatus of claim 17, wherein:

the level-dependent bit allocation means is for allocating level-dependent quantizing bits among the bands according to a selected bit allocation pattern; and

the apparatus additionally comprises;

a means for storing plural predetermined bit allocation patterns, the predetermined bit allocation patterns allocating numbers of level-dependent quantizing bits corresponding to different values of the division ratio, and

an additional selecting means for selecting, in response to the digital input signal, and in response to the division ratio, and for providing to the level-dependent bit allocating means as the selected bit allocation pattern, an appropriate one of the plural predetermined bit allocation patterns.

19. The apparatus of claim 1, wherein the means for deriving plural spectral coefficients comprises:

a frequency dividing means for receiving the digital input signal and for dividing the digital input signal into plural frequency ranges, and for providing a frequency range signal in each frequency range;

a time dividing means for dividing in time a frequency range signal into blocks; and

an orthogonal transform means for transforming a block of the frequency range signal to provide the plural spectral coefficients.

20. A method for compressing a digital input signal to provide a compressed digital output signal, the method comprising the steps of:

deriving plural spectral coefficients from the digital input signal, and grouping the spectral coefficients into bands, each band having a band magnitude and a band frequency; and

adaptively allocating a number of spectrum-dependent quantizing bits among the bands to allocate to each band a number of spectrum-dependent quantizing bits for quantizing each spectral coefficient in the band, the number of spectrum-dependent quantizing bits allocated to each band being determined according to the band magnitude, weighted depending on the band frequency.

21. The method of claim 20, wherein, in the allocating step, the band magnitude is weighted depending on the band frequency such that the number of spectrum-dependent quantizing bits allocated to lower frequency bands for a given band magnitude is increased.

22. The method of claim 20, wherein:

the method additionally comprises the steps of;

determining, for each band, the band magnitude in response to the spectral coefficients in the band, and

weighting, for each band, the band magnitude depending on the band frequency to provide a weighted band magnitude; and

in the allocating step, the number of spectrum-dependent quantizing bits for quantizing each spectral coefficient in the band is allocated to each band in response to the weighted band magnitude for the band.

23. The method of claim 20, wherein, in the allocating step, the band magnitude for each band is weighted by a weighting coefficient dependent on the digital input signal and the band frequency.

24. The method of claim 20, wherein:

in the allocating step, the band magnitude for each band is weighted depending on the band frequency according to a selected frequency-dependent weighting pattern; and

the method additionally comprises the steps of;

providing plural frequency-dependent weighting patterns, and

selecting in response to the digital input signal, and providing as the selected frequency-dependent weighting pattern, an appropriate one of the plural frequency-dependent weighting patterns.

25. The method of claim 24, for compressing a digital input signal having a level, wherein, in the selecting step, the appropriate one of the plural frequency-dependent weighting patterns is selected in response to the level of the digital input signal.

26. The method of claim 24, for compressing a digital input signal having a spectrum, wherein, in the selecting step, the appropriate one of the plural frequency-dependent weighting patterns is selected in response to the spectrum of the digital input signal.

27. The method of claim 24, wherein, in the allocating step, the band magnitude is weighted such that the number of spectrum-dependent quantizing bits allocated to lower frequency bands for a given band magnitude is increased.

28. The method of claim 20, additionally comprising the steps of:

allocating a number of level-dependent quantizing bits among the bands according to a predetermined bit allocation pattern defining, for each band, a number of level-dependent quantizing bits allocated for quantizing each spectral coefficient in the band; and

quantizing each spectral coefficient in a band using a number of quantizing bits equal to the sum of the number of spectrum-dependent quantizing bits allocated for quantizing each spectral coefficient in the band and the number of level-dependent quantizing bits allocated for quantizing each spectral coefficient in the band.

29. The method of claim 28, wherein in the step of allocating a number of spectrum-dependent quantizing bits among the bands, the band magnitude is weighted such that the number of spectrum-dependent quantizing bits allocated to lower frequency bands for a given band magnitude is increased.

30. The method of claim 28, wherein:

the method additionally comprises the steps of;

determining, for each band, the band magnitude in response to the spectral coefficients in the band, and

weighting, for each band, the band magnitude depending on the band frequency to provide a weighted band magnitude; and

in the step of allocating a number of spectrum-dependent quantizing bits among the bands, the number of spectrum-dependent quantizing bits for quantizing each spectral coefficient in the band is allocated to each band in response to the weighted band magnitude.

31. The method of claim 28, wherein, in the step of allocating a number of spectrum-dependent quantizing bits among the bands, the band magnitude for each band is weighted by a weighting coefficient dependent on the digital input signal and the band frequency.

32. The method of claim 28, wherein:

in the step of allocating a number of spectrum-dependent quantizing bits among the bands, the band magnitude is weighted depending on the band frequency according to a selected frequency-dependent weighting pattern; and

the method additionally comprises the steps of;

providing plural frequency-dependent weighting patterns, and

selecting in response to the digital input signal, and providing as the selected frequency-dependent weighting pattern, an appropriate one of the plural frequency-dependent weighting patterns.

33. The method of claim 32, wherein, in the step of allocating a number of spectrum-dependent quantizing bits among the bands, the band magnitude is weighted such that the number of spectrum-dependent quantizing bits allocated to lower frequency bands for a given band magnitude is increased.

34. The method of claim 28, wherein:

in the step of allocating a number of level-dependent quantizing bits among the bands, level-dependent quantizing bits are allocated among the bands according to a selected bit allocation pattern; and

the method additionally comprises the steps of;

providing plural predetermined bit allocation patterns, and

selecting in response to the digital input signal, and providing as the selected bit allocation pattern, an appropriate one of the plural predetermined bit allocation patterns.

35. The method of claim 28, for compressing a digital input signal having a spectrum, the spectrum having a smoothness, and using a total number of available quantizing bits for quantizing all the spectral coefficients, wherein:

the method additionally comprises;

determining a division of the total number of available quantizing bits between the number of spectrum-dependent quantizing bits and the number of level-dependent quantizing bits in response to the smoothness of the spectrum of the digital input signal,

providing in response to the determined division a division ratio between the number of level-dependent quantizing bits and the total number of available quantizing bits, and a complement of the division ratio,

multiplying the number of spectrum-dependent quantizing bits for each band by the complement of the division ratio, and

multiplying the number of level-dependent quantizing bits for each band by the division ratio; and

in the quantizing step, each spectral coefficient in a band is quantized using a number of quantizing bits equal to the sum of the multiplied number of spectrum-dependent quantizing bits allocated for quantizing each spectral coefficient in the band and the multiplied number of level-dependent quantizing bits allocated for quantizing each spectral coefficient in the band.

36. The method of claim 28, for compressing a digital input signal having a spectrum, the spectrum having a smoothness, and using a total number of available quantizing bits for quantizing all the spectral coefficients, wherein:

the method additionally comprises;

determining a division of the total number of available quantizing bits between the number of spectrum-dependent quantizing bits and the number of level-dependent quantizing bits in response to the smoothness of the spectrum of the digital input signal, and

providing in response to the determined division a division ratio between the number of level-dependent quantizing bits and the total number of available quantizing bits, and a complement of the division ratio;

in the step of allocating a number of spectrum-dependent quantizing bits among the bands, the number of spectrum-dependent quantizing bits allocated among the bands is substantially proportional to the complement of the division ratio; and

in the step of allocating a number of level-dependent quantizing bits among the bands, the number of level-dependent quantizing bits allocated among the bands is substantially proportional to the division ratio.

37. The method of claim 36, wherein:

in the step of allocating a number of level-dependent quantizing bits among the bands, level-dependent quantizing bits are allocated among the bands according to a selected bit allocation pattern; and

the method additionally comprises the steps of;

providing plural predetermined bit allocation patterns, the predetermined bit allocation patterns allocating numbers of level-dependent quantizing bits corresponding to different values of the division ratio, and

selecting in response to the digital input signal, and in response to the division ratio, and providing as the selected bit allocation pattern, an appropriate one of the plural predetermined bit allocation patterns.

38. The method of claim 20, wherein the step of deriving plural spectral coefficients from the digital input signal comprises the steps of:

dividing the digital input signal into plural frequency ranges, and providing a frequency range signal in each frequency range;

dividing a frequency range signal in time into blocks; and

orthogonally transforming a block of the frequency range signal to provide the plural spectral coefficients.

39. A medium whereon is stored a compressed digital signal derived from a digital input signal by a data compressing method comprising the steps of:

deriving plural spectral coefficients from the digital input signal, and grouping the spectral coefficients into bands, each band having a band magnitude and a band frequency;

adaptively allocating a number of spectrum-dependent quantizing bits among the bands to allocate to each band a number of spectrum-dependent quantizing bits for quantizing each spectral coefficient in the band, the number of spectrum-dependent quantizing bits allocated to each band being determined according to the band magnitude, weighted depending on the band frequency;

quantizing each spectral coefficient in each band using a number of quantizing bits including the number of spectrum-dependent quantizing bits allocated for quantizing each spectral coefficient in the band; and

including the quantized spectral coefficients in the compressed digital signal.

40. The medium of claim 39, wherein, in the allocating step of the data compressing method, the band magnitude is weighted depending on the band frequency such that the number of spectrum-dependent quantizing bits allocated to lower frequency bands for a given band magnitude is increased.

41. The medium of claim 39, wherein, in the data compressing method:

the method additionally comprises the steps of;

determining, for each band, the band magnitude in response to the spectral coefficients in the band, and

weighting, for each band, the band magnitude depending on the band frequency to provide a weighted band magnitude; and

in the allocating step, the number of spectrum-dependent quantizing bits for quantizing each spectral coefficient in the band is allocated to each band in response to the weighted band magnitude for the band.

42. The medium of claim 39, wherein, in the allocating step of the data compressing method, the band magnitude for each band is weighted by a weighting coefficient dependent on the digital input signal and the band frequency.

43. The medium of claim 39, wherein, in the data compressing method:

in the allocating step, the band magnitude for each band is weighted depending on the band frequency according to a selected frequency-dependent weighting pattern; and

the method additionally comprises the steps of;

providing plural frequency-dependent weighting patterns, and

selecting in response to the digital input signal, and providing as the selected frequency-dependent weighting pattern, an appropriate one of the plural frequency-dependent weighting patterns.

44. The medium of claim 39, wherein, in the data compressing method:

the method additionally comprises the step of allocating a number of level-dependent quantizing bits among the bands according to a predetermined bit allocation pattern defining, for each band, a number of level-dependent quantizing bits allocated for quantizing each spectral coefficient in the band; and

in the quantizing step, each spectral coefficient in each band is quantized using a number of quantizing bits equal to the sum of the number of spectrum-dependent quantizing bits allocated for quantizing each spectral coefficient in the band and the number of level-dependent quantizing bits allocated for quantizing each spectral coefficient in the band.

45. The medium of claim 44, wherein in the step in the data compressing method of allocating a number of spectrum-dependent quantizing bits among the bands, the band magnitude is weighted such that the number of spectrum-dependent quantizing bits allocated to lower frequency bands for a given band magnitude is increased.

46. The medium of claim 44, wherein, in the data compressing method:

the method additionally comprises the steps of;

determining, for each band, the band magnitude in response to the spectral coefficients in the band, and

weighting, for each band, the band magnitude depending on the band frequency to provide a weighted band magnitude; and

in the step of allocating a number of spectrum-dependent quantizing bits among the bands, the number of spectrum-dependent quantizing bits for quantizing each spectral coefficient in the band is allocated to each band in response to the weighted band magnitude.

47. The medium of claim 44, wherein, in the step in the data compressing method of allocating a number of spectrum-dependent quantizing bits among the bands, the band magnitude for each band is weighted by a weighting coefficient dependent on the digital input signal and the band frequency.

48. The medium of claim 44, wherein, in the data compressing method:

in the step of allocating a number of spectrum-dependent quantizing bits among the bands, the band magnitude is weighted depending on the band frequency according to a selected frequency-dependent weighting pattern; and

the method additionally comprises the steps of;

providing plural frequency-dependent weighting patterns, and

selecting in response to the digital input signal, and providing as the selected frequency-dependent weighting pattern, an appropriate one of the plural frequency-dependent weighting patterns.

49. The medium of claim 44, wherein, in the data compressing method:

in the step of allocating a number of level-dependent quantizing bits among the bands, level-dependent quantizing bits are allocated among the bands according to a selected bit allocation pattern; and

the method additionally comprises the steps of;

providing plural predetermined bit allocation patterns, and

selecting in response to the digital input signal, and providing as the selected bit allocation pattern, an appropriate one of the plural predetermined bit allocation patterns.

50. The medium of claim 44, wherein the data compressing method is for compressing a digital input signal having a spectrum, the spectrum having a smoothness, and the method uses a total number of available quantizing bits for quantizing all the spectral coefficients, wherein:

the method additionally comprises;

determining a division of the total number of available quantizing bits between the number of spectrum-dependent quantizing bits and the number of level-dependent quantizing bits in response to the smoothness of the spectrum of the digital input signal,

providing in response to the determined division a division ratio between the number of level-dependent quantizing bits and the total number of available quantizing bits, and a complement of the division ratio,

multiplying the number of spectrum-dependent quantizing bits for each band by the complement of the division ratio, and

multiplying the number of level-dependent quantizing bits for each band by the division ratio; and

in the quantizing step, each spectral coefficient in a band is quantized using a number of quantizing bits equal to the sum of the multiplied number of spectrum-dependent quantizing bits allocated for quantizing each spectral coefficient in the band and the multiplied number of level-dependent quantizing bits allocated for quantizing each spectral coefficient in the band.

51. The medium of claim 39, wherein, in the data compressing method, the step of deriving plural spectral coefficients from the digital input signal comprises the steps of:

dividing the digital input signal into plural frequency ranges, and providing a frequency range signal in each frequency range;

dividing a frequency range signal in time into blocks; and

orthogonally transforming a block of the frequency range signal to provide the plural spectral coefficients.

52. The medium of claim 39, wherein the medium is selected from a group including an optical disc, a semiconductor memory, and a magnetic tape.

53. An expander for expanding a compressed digital signal derived from a digital input signal by a data compressing method comprising the steps of:

deriving plural spectral coefficients from the digital input signal, and grouping the spectral coefficients into bands, each band having a band magnitude and a band frequency,

adaptively allocating a number of spectrum-dependent quantizing bits among the bands to allocate to each band a number of spectrum-dependent quantizing bits for quantizing each spectral coefficient in the band, the number of spectrum-dependent quantizing bits allocated to each band being determined according to the band magnitude, weighted depending on the band frequency,

quantizing each spectral coefficient in each band using a number of quantizing bits including the number of spectrum-dependent quantizing bits allocated for quantizing each spectral coefficient in the band, and

including in the compressed digital signal, for each band, the quantized spectral coefficients and data indicating the number of quantizing bits,

the expander including;

a means for separating from the compressed digital signal, for each band, the quantized spectral coefficients and the data indicating the number of quantizing bits;

a means for dequantizing, for each band, the quantized spectral coefficients in response to the data indicating the number of quantizing bits; and

a means for deriving the output signal from the dequantized spectral coefficients.

54. The expander of claim 53, wherein, in the allocating step of the data compressing method, the band magnitude is weighted depending on the band frequency such that the number of spectrum-dependent quantizing bits allocated to lower frequency bands for a given band magnitude is increased.

55. The expander of claim 53, wherein, in the data compressing method:

the method additionally comprises the steps of;

determining, for each band, the band magnitude in response to the spectral coefficients in the band, and

weighting, for each band, the band magnitude depending on the band frequency to provide a weighted band magnitude; and

in the allocating step, the number of spectrum-dependent quantizing bits for quantizing each spectral coefficient in the band is allocated to each band in response to the weighted band magnitude for the band.

56. The expander of claim 53, wherein, in the allocating step of the data compressing method, the band magnitude for each band is weighted by a weighting coefficient dependent on the digital input signal and the band frequency.

57. The expander of claim 53, wherein, in the data compressing method:

in the allocating step, the band magnitude for each band is weighted depending on the band frequency according to a selected frequency-dependent weighting pattern; and

the method additionally comprises the steps of;

providing plural frequency-dependent weighting patterns, and

selecting in response to the digital input signal, and providing as the selected frequency-dependent weighting pattern, an appropriate one of the plural frequency-dependent weighting patterns.

58. The expander of claim 53, wherein, in the data compressing method:

the method additionally comprises the step of allocating a number of level-dependent quantizing bits among the bands according to a predetermined bit allocation pattern defining, for each band, a number of level-dependent quantizing bits allocated for quantizing each spectral coefficient in the band; and

in the quantizing step, each spectral coefficient in each band is quantized using a number of quantizing bits equal to the sum of the number of spectrum-dependent quantizing bits allocated for quantizing each spectral coefficient in the band and the number of level-dependent quantizing bits allocated for quantizing each spectral coefficient in the band.

59. The expander of claim 58, wherein in the step in the data compressing method of allocating a number of spectrum-dependent quantizing bits among the bands, the band magnitude is weighted such that the number of spectrum-dependent quantizing bits allocated to lower frequency bands for a given band magnitude is increased.

60. The expander of claim 58, wherein, in the data compressing method:

the method additionally comprises the steps of;

determining, for each band, the band magnitude in response to the spectral coefficients in the band, and

weighting, for each band, the band magnitude depending on the band frequency to provide a weighted band magnitude; and

in the step of allocating a number of spectrum-dependent quantizing bits among the bands, the number of spectrum-dependent quantizing bits for quantizing each spectral coefficient in the band is allocated to each band in response to the weighted band magnitude.

61. The expander of claim 58, wherein, in the step in the data compressing method of allocating a number of spectrum-dependent quantizing bits among the bands, the band magnitude for each band is weighted by a weighting coefficient dependent on the digital input signal and the band frequency.

62. The expander of claim 58, wherein, in the data compressing method:

in the step of allocating a number of spectrum-dependent quantizing bits among the bands, the band magnitude is weighted depending on the band frequency according to a selected frequency-dependent weighting pattern; and

the method additionally comprises the steps of;

providing plural frequency-dependent weighting patterns, and

selecting in response to the digital input signal, and providing as the selected frequency-dependent weighting pattern, an appropriate one of the plural frequency-dependent weighting patterns.

63. The expander of claim 58, wherein, in the data compressing method:

in the step of allocating a number of level-dependent quantizing bits among the bands, level-dependent quantizing bits are allocated among the bands according to a selected bit allocation pattern; and

the method additionally comprises the steps of;

providing plural predetermined bit allocation patterns, and

selecting in response to the digital input signal, and providing as the selected bit allocation pattern, an appropriate one of the plural predetermined bit allocation patterns.

64. The expander of claim 58, wherein the data compressing method is for compressing a digital input signal having a spectrum, the spectrum having a smoothness, and the method uses a total number of available quantizing bits for quantizing all the spectral coefficients, wherein:

the method additionally comprises;

determining a division of the total number of available quantizing bits between the number of spectrum-dependent quantizing bits and the number of level-dependent quantizing bits in response to the smoothness of the spectrum of the digital input signal,

providing in response to the determined division a division ratio between the number of level-dependent quantizing bits and the total number of available quantizing bits, and a complement of the division ratio,

multiplying the number of spectrum-dependent quantizing bits for each band by the complement of the division ratio, and

multiplying the number of level-dependent quantizing bits for each band by to the division ratio; and

in the quantizing step, each spectral coefficient in a band is quantized using a number of quantizing bits equal to the sum of the multiplied number of spectrum-dependent quantizing bits allocated for quantizing each spectral coefficient in the band and the multiplied number of level-dependent quantizing bits allocated for quantizing each spectral coefficient in the band.

65. The expander of claim 53, wherein, in the data compressing method, the step of deriving plural spectral coefficients from the digital input signal comprises the steps of:

dividing the digital input signal into plural frequency ranges, and providing a frequency range signal in each frequency range;

dividing a frequency range signal in time into blocks; and

orthogonally transforming a block of the frequency range signal to provide the plural spectral coefficients.