Method and apparatus for performing bit-allocation coding for an acoustic signal of frequency region and time region correction for an acoustic signal and method and apparatus for decoding a decoded acoustic signal

• US 5,794,179 A
• Filed: 07/26/1996
• Issued: 08/11/1998
• Est. Priority Date: 07/27/1995
• Status: Expired due to Term

• Current Assignee: JVC Kenwood Corporation
• Original Assignee: Victor Company of Japan Limited (JVC Kenwood Corporation)
• Inventors: Yamabe, Takaaki
• Primary Examiner(s): Hudspeth, David R.
• Assistant Examiner(s): Storm, Donald L.
• Application Number:
  US08/686,840
• Time in Patent Office: 746 Days
• Field of Search: 704/229, 704/203, 704/204, 704/230, 704/205, 704/227, 704/212, 704/220, 704/219
• US Class Current: 704/205
• CPC Class Codes: H04B 1/667 using a division in frequency subbands for TV signals H04N19/63

Abstract

A digital acoustic signal of a time region information is divided into a plurality of sub-band groups, and the number of bits allocated to each data of the sub-band groups is reduced by a reducing number Nr in a quantizing unit according to a maximum value selected among energy values of pieces of data in each sub-band group, so that the digital acoustic signal of a frequency region information is produced. Thereafter, Nr bits are added to each data of the digital acoustic signal of the frequency region information in the inversely quantizing unit, and the data of all sub-bands are combined to produce a reproduced digital acoustic signal of the time region information. Thereafter, a residual signal indicating a difference between the digital-acoustic signal of the time region information and the reproduced digital acoustic signal of the time region information is produced, and the residual signal and the reproduced digital acoustic signal of the frequency region information are multiplexed. Therefore, when a reproduced digital acoustic signal of the time region information is produced from the reproduced digital acoustic signal of the frequency region information and is corrected according to the residual signal in a decoding operation, the digital acoustic signal can be efficiently coded and decoded while maintaining the redundancy.

Claims

1. An acoustic signal coding method, comprising the steps of:

dividing a digital acoustic signal of time region information supplied as an original signal into one or more pieces of sub-band data for each of a plurality of sub-bands to change the digital acoustic signal of the time region information to a digital acoustic signal of frequency region information;

combining the pieces of sub-band data of all sub-bands to produce a band combined signal as a reproduced digital acoustic signal of time region information;

producing a residual signal from the digital acoustic signal of the time region information and the reproduced digital acoustic signal of the time region information as time region correcting information; and

multiplexing the digital acoustic signal of the frequency region information and the time region correcting information to produce a bit stream as an output signal.

2. The acoustic signal coding method according to claim 1 in which the step of dividing a digital acoustic signal of time region information further includes the steps of:

detecting a maximum value among the pieces of sub-band data for each of the sub-bands, each piece of sub-band data being expressed at a fixed quantization bit precision;

producing bit allocating information indicating a minimum bit number required to express the maximum value among the pieces of sub-band data of one sub-band for each of the sub-bands; and

reducing the number of bits used in each piece of sub-band data by a reducing number Nr according to the bit allocating information for each sub-band to produce the digital acoustic signal of the frequency region information composed of the pieces of sub-band data, the pieces of sub-band data of the digital acoustic signal of the frequency region information being respectively expressed by an allowed number of bits indicated by the bit allocating information,for each sub-band, and

the step of combining the pieces of sub-band data of all sub-bands further includes the step of;

adding Nr bits to the pieces of sub-band data of the digital acoustic signal of the frequency region information for each sub-band before the pieces of sub-band data of all sub-bands are combined on condition that values of the Nr bits added to each piece of sub-band data are the same as that of a sign bit of the piece of sub-band data.

3. The acoustic signal coding method according to claim 1 in which the step of producing a residual signal further includes the step of:

delaying the digital acoustic signal of the time region information supplied as the original signal before the residual signal is produced to cancel the delay of the reproduced digital acoustic signal of the time region information caused by the division of the digital acoustic signal of the time region information and the combination of the pieces of sub-band data.

4. An The acoustic signal coding method according to claim 1 in which the step of combining the pieces of sub-band data of all sub-bands further includes the steps of:

detecting a maximum value among the pieces of sub-band data of the band combined signal, each piece of sub-band data being expressed at a fixed quantization bit precision;

producing bit allocating information indicating a minimum bit required to express the maximum value among the pieces of sub-band data of one sub-band; and

reducing the number of bits used in each piece of sub-band data by a reducing number Nr according to the bit allocating information to produce the reproduced digital acoustic signal of the time region information composed of the pieces of sub-band data, the pieces of sub-band data of the digital acoustic signal of the frequency region information being respectively expressed by an allowed number of bits indicated by the bit allocating information.

5. An acoustic signal coding and decoding method, comprising the steps of:

dividing a digital acoustic signal of time region information supplied as an original signal into one or more pieces of sub-band data for each of a plurality of sub-bands to change the digital acoustic signal of the time region information to a digital acoustic signal of frequency region information;

combining the pieces of sub-band data of all sub-bands to produce a band combined signal as a first reproduced digital acoustic signal of the time region information;

producing a residual signal from the first reproduced digital acoustic signal of the time region information and the digital acoustic signal of the time region information as time region correcting information;

multiplexing the digital acoustic signal of the frequency region information and the time region correcting information to a bit stream;

demultiplexing the bit stream to separate the digital acoustic signal of the frequency region information from the time region correcting information;

combining the pieces of sub-band data of all sub-bands included in the digital acoustic signal of the frequency region information to produce a second reproduced digital acoustic signal of the time region information; and

correcting the second reproduced digital acoustic signal of the time region information according to the time region correcting information demultiplexed.

6. An The acoustic signal coding and decoding method according to claim 5 in which the step of dividing a digital acoustic signal of time region information further includes the steps of:

detecting a maximum value among the pieces of sub-band data for each of the sub-bands, each niece of sub-band data being expressed at a fixed quantization bit precision;

producing bit allocating information indicating a minimum bit number required to express the maximum value among the pieces of sub-band data of one sub-band for each of the sub-bands; and

reducing the number of bits used in each piece of sub-band data by a reducing number Nr according to the bit allocating information for each sub-band to produce the digital acoustic signal of the frequency region information composed of the pieces of sub-band data, the pieces of sub-band data of the digital acoustic signal of the frequency region information being respectively expressed by an allowed number of bits indicated by the bit allocating information for each sub-band,

the step of combining the pieces of sub-band data of all sub-bands to produce a band combined signal further includes the step of;

adding Nr bits to the pieces of sub-band data of the digital acoustic signal of the frequency region information for each sub-band before the pieces of sub-band data of all sub-bands are combined on condition that values of the Nr bits added to each piece of sub-band data are the same as that of a sign bit of the piece of sub-band data, and

the step of combining the pieces of sub-band data of all sub-bands included in the digital acoustic signal further includes the step of;

adding Nr bits having the same value as that of a sign bit of each piece of sub-band data to each piece of sub-band data of the digital acoustic signal of the frequency region information for each sub-band before the pieces of sub-band data of all sub-bands are combined on condition that values of the Nr bits added to each piece of sub-band data are the same as that of a sign bit of the piece of sub-band data.

7. The acoustic signal coding and decoding method according to claim 5 in which the step of producing a residual signal further includes the step of:

delaying the digital acoustic signal of the time region information supplied as the original signal before the residual signal is produced to cancel the delay of the first reproduced digital acoustic signal of the time region information caused by the division of the digital acoustic signal of the time region information and the combination of the pieces of sub-band data for the production of the first reproduced digital acoustic signal and to cancel the delay of the second reproduced digital acoustic signal of the time region information caused by the division of the digital acoustic signal of the time region information and the combination of the pieces of sub-band data for the production of the second reproduced digital acoustic signal.

8. An acoustic signal coding apparatus comprising:

a band dividing filter for dividing a digital acoustic signal of time region information supplied as an original signal into one or more pieces of sub-band data for each of a plurality of sub-bands to change the digital acoustic signal of the time region information to a digital acoustic signal of frequency region information;

a maximum value selecting unit for selecting a maximum value among the pieces of sub-band data divided by the band dividing filter for each of the sub-bands, each piece of sub-band data being expressed at a fixed quantization bit precision;

a bit allocating information producing unit for producing bit allocating information indicating a minimum bit number required to express the maximum value selected by the maximum value selecting unit among the pieces of sub-band data of one sub-band for each of the sub-bands;

a quantizing unit for reducing the number of bits used in each piece of sub-band data divided by the band dividing filter by a reducing number Nr according to the bit allocating information produced by the bit allocating information producing unit for each sub-band to produce the digital acoustic signal of the frequency region information composed of the pieces of sub-band data, the pieces of sub-band data of the digital acoustic signal of the frequency region information being respectively expressed by an allowed number of bits indicated by the bit allocating information for each sub-band;

an inversely quantizing unit for adding Nr bits to each piece of sub-band data of one sub-band, in which the number of bits is reduced by the reducing number Nr by the quantizing unit, for each sub-band on condition that values of the Nr bits added to each piece of sub-band data are the same as that of a sign bit of the niece of sub-band data;

a band combining filter for combining the pieces of sub-band data of all sub-bands produced by the inversely quantizing unit to produce a band combined signal as a reproduced digital acoustic signal of time region information;

a residual signal producing unit for producing a residual signal from the digital acoustic signal of the time region information and the reproduced digital acoustic signal of the time region information produced by the band combining filter and outputting the residual signal as time region correcting information; and

a multiplexing unit for multiplexing the digital acoustic signal of the frequency region information produced by the quantizing unit and the time region correcting information produced by the residual signal producing unit to produce a bit stream as an output signal.

9. The acoustic signal coding apparatus according to claim 8, further comprising:

a delaying unit for delaying the digital acoustic signal of the time region information supplied as the original signal to cancel the delay of the reproduced digital acoustic signal of the time region information caused by the division of the digital acoustic signal of the time region information performed by the band dividing filter and the combination of the pieces of sub-band data performed by the band combining filter.

10. An acoustic signal coding and decoding apparatus comprising:

a band dividing filter for dividing a digital acoustic signal of time region information supplied as an original signal into one or more pieces of sub-band data for each of a plurality of sub-bands to change the digital acoustic signal of the time region information to a digital acoustic signal of frequency region information;

a maximum value selecting unit for selecting a maximum value among the pieces of sub-band data divided by the band dividing filter for each of the sub-bands, each piece of sub-band data being expressed at a fixed quantization bit precision;

a bit allocating information producing unit for producing bit allocating information indicating a minimum bit number required to express the maximum value selected by the maximum value selecting unit among the pieces of sub-band data of one sub-band for each of the sub-bands;

a quantizing unit for reducing the number of bits used in each piece of sub-band data divided by the band dividing filter by a reducing number Nr according to the bit allocating information produced by the bit allocating information producing unit for each sub-band to produce the digital acoustic signal of the frequency region information composed of the pieces of sub-band data, the pieces of sub-band data of the digital acoustic signal of the frequency region information being respectively expressed by an allowed number of bits indicated by the bit allocating information for each sub-band;

a first inversely quantizing unit for adding Nr bits to the pieces of sub-band data of one sub-band, in which the number of bits is reduced by the reducing number Nr by the quantizing unit, for each sub-band on condition that values of the Nr bits added to each piece of sub-band data are the same as that of a sign bit of the piece of sub-band data;

a first band combining filter for combining the pieces of sub-band data of all sub-bands produced by the first inversely quantizing unit to produce a band combined signal as a first reproduced digital acoustic signal of time region information;

a residual signal producing unit for producing a residual signal from the digital acoustic signal of the time region information and the first reproduced digital acoustic signal of the time region information produced by the first band combining filter and outputting the residual signal as time region correcting information;

a multiplexing unit for multiplexing the digital acoustic signal of the frequency region information produced by the quantizing unit and the time region correcting information produced by the residual signal producing unit to produce a bit stream as an output signal;

a demultiplexing unit for demultiplexing the bit stream produced by the multiplexing unit to separate the digital acoustic signal of the frequency region information from the time region correcting information;

a second inversely quantizing unit for adding Nr bits to the pieces of sub-band data of the digital acoustic signal of the frequency region information demultiplexed by the demultiplexing unit for each sub-band on condition that a value of the Nr bits added to each piece of sub-band data is the same as that of a sign bit of the piece of sub-band data;

a second band combining filter for combining the pieces of sub-band data of all sub-bands in the digital acoustic signal of the frequency region information produced by the second inversely quantizing unit to produce a second reproduced digital acoustic signal of the time region information; and

a residual correcting unit for correcting the second reproduced digital acoustic signal of the time region information produced by the second band combining filter according to the time region correcting information demultiplexed by the demultiplexing unit.

11. The acoustic signal coding and decoding apparatus according to claim 10, further comprising:

a delaying unit for delaying the digital acoustic signal of the time region information supplied as the original signal to cancel the delay of the first reproduced digital acoustic signal of the time region information caused by the division of the digital acoustic signal of the time region information performed by the band dividing filter and the combination of the pieces of sub-band data performed by the first band combining filter and to cancel the delay of the second reproduced digital acoustic signal of the time region information caused by the division of the digital acoustic signal of the time region information performed by the band dividing filter and the combination of the pieces of sub-band data performed by the second band combining filter.

12. An acoustic signal coding method, comprising the steps of:

receiving a digital acoustic signal of time region information as an original signal;

dividing the digital acoustic signal of time region information into one or more pieces of sub-band data for each of a plurality of sub-bands to change the digital acoustic signal of the time region information to a digital acoustic signal of frequency region information;

combining the pieces of sub-band data of all sub-bands to produce a band combined signal as a reproduced digital acoustic signal of time region information;

producing a residual signal composed of pieces of residual data from the digital acoustic signal of the time region information received as the original signal and the reproduced digital acoustic signal of the time region information as time region correcting information;

producing a piece of pulse data according to the residual signal on condition that a residual value of one piece of having an absolute value higher than a maximum value, which is expressed by a prescribed number of bits determined according to an average of absolute values of the pieces of residual data of the residual signal, is made lower than the maximum value in cases where the residual signal is again produced from a new digital acoustic signal of time region information obtained by adding the pulse data to the digital acoustic signal of time region information;

adding the pulse data to the digital acoustic signal of time region information to produce an adjusted digital acoustic signal of frequency region information from the digital acoustic signal of time region information;

producing an adjusted reproduced digital acoustic signal of time region information from the adjusted digital acoustic signal of frequency region information;

producing an adjusted residual signal from the adjusted reproduced digital acoustic signal of time region information as adjusted time region correcting information; and

multiplexing the digital acoustic signal of the frequency region information and the adjusted time region correcting information to produce a bit stream as an output signal.

13. The acoustic signal coding method according to claim 12, in which the step of producing a piece of pulse data comprises the steps of:

calculating the average value of the residual values of the residual signal;

determining a target bit allocation number Nt according to the average value;

detecting a piece of particular residual data of the residual signal having an absolute value higher than a maximum value expressed by Nt bits from the pieces of residual data of the residual signal;

determining a particular sample of the digital acoustic signal of time region information corresponding to the particular residual data of the residual signal;

determining a pulse value of the pulse data according to the target bit allocation number Nt and the absolute value of the particular residual data to reduce a value of the particular residual data of the residual signal in cases where the residual signal is produced from the digital acoustic signal of time region information in which the pulse data having the pulse value is added to the particular sample;

producing the pulse data having the pulse value, and the step of adding the pulse data includes the step of adding the pulse data having the pulse value to the particular sample of the digital acoustic signal of time region information.

14. The acoustic signal coding method according to claim 13, in which the step of determining a target bit allocation number Nt comprises the steps of:

setting an upper limit of values, to which the target bit allocation number Nt is applied, to a value obtained by subtracting an allowance value from a maximum value expressed by Nt bits; and

determining the target bit allocation number Nt applied for the average value on condition that the average value is equal to or lower than the upper limit of the target bit allocation number Nt and the average value is higher than an upper limit of another target bit allocation number Nt-1.

15. The acoustic signal coding method according to claim 13, in which the step of determining a pulse value comprises the steps of:

calculating a difference between 3/4 of a maximum value expressed by Nt bits and the absolute value of the particular residual data;

setting the difference to an absolute value of the pulse value; and

setting the pulse value to a sign inverse to that of the particular sample of the digital acoustic signal of time region information.

16. The acoustic signal coding method according to claim 12, in which the step of dividing the digital acoustic signal of time region information, the step of combining the pieces of sub-band data of all sub-bands, the step of producing a residual signal, the step of producing a piece of pulse data and the step of adding the pulse data are repeated until absolute values of all pieces of residual data of the adjusted residual signal become lower than a prescribed value.

17. The acoustic signal coding method according to claim 12, in which the step of producing a piece of pulse data comprises the steps of:

detecting one or more pieces of particular residual data having absolute values higher than a prescribed value from the pieces of residual data of the residual signal; and

producing one or more pieces of pulse data corresponding to the pieces of particular residual data to make residual values of the particular residual data of the residual signal lower than the prescribed value in cases where the pieces of pulse data are added to the digital acoustic signal of time region information to again produce the residual signal.

18. An acoustic signal coding apparatus, comprising:

a band dividing filter for dividing a digital acoustic signal of time region information supplied as an original signal into one or more pieces of sub-band data for each of a plurality of sub-bands to change the digital acoustic signal of the time region information to a digital acoustic signal of frequency region information;

a maximum value selecting unit for selecting a maximum value among the pieces of sub-band data divided by the band dividing filter for each of the sub-bands;

a bit allocating information producing unit for producing bit allocating information indicating a minimum bit number required to express the maximum value selected by the maximum value selecting unit among the pieces of sub-band data of one sub-band according to the maximum value selected by the maximum value selecting unit for each of the sub-bands;

a quantizing unit for reducing the number of bits used in each piece of sub-band data divided by the band dividing filter by a reducing number Nr according to the bit allocating information produced by the bit allocating information producing unit for each sub-band to produce the digital acoustic signal of the frequency region information composed of the pieces of sub-band data, the pieces of sub-band data of the digital acoustic signal of the frequency region information being respectively expressed by an allowed number of bits indicated by the bit allocating information for each sub-band;

an inversely quantizing unit for adding Nr bits to the pieces of sub-band data of one sub-band, in which the number of bits is reduced by the reducing number Nr by the quantizing unit, for each sub-band on condition that values of the Nr bits added to each piece of sub-band data are the same as that of a sign bit of the piece of sub-band data;

a band combining filter for combining the pieces of sub-band data of all sub-bands produced by the inversely quantizing unit to produce a band combined signal as a reproduced digital acoustic signal of time region information;

a residual signal producing unit for producing a residual signal composed of pieces of residual data from the digital acoustic signal of the time region information and the reproduced digital acoustic signal of the time region information produced by the band combining filter, each piece of residual data having a residual value;

a pulse data producing unit for producing a piece of pulse data according to the residual signal produced by the residual signal producing unit on condition that a residual value of one piece of residual data having an absolute value higher than a maximum value, which is expressed by a prescribed number of bits determined according to an average of absolute values of the pieces of residual data of the residual signal, is made lower than the prescribed value in cases where the residual signal is again produced from a new digital acoustic signal of time region information obtained by adding the pulse data to the digital acoustic signal of time region information;

an acoustic signal adjusting unit for adding the pulse data produced by the pulse data producing unit to the digital acoustic signal of time region information to produce an adjusted digital acoustic signal of time region information and produce an adjusted residual signal from the adjusted reproduced digital acoustic signal of time region information as adjusted time region correcting information in the residual signal producing unit through the band dividing filter, the quantizing unit, the inversely quantizing unit and the band combining filter; and

a multiplexing unit for multiplexing the digital acoustic signal of the frequency region information produced by the quantizing unit and the adjusted time region correcting information produced by the residual signal producing unit to produce a bit stream as an output signal.

19. The acoustic signal coding apparatus according to claim 18 in which the pulse data producing unit comprises:

a residual signal average value calculating unit for calculating the average value of the residual values of the residual signal;

a target bit allocation number determining unit for determining a target bit allocation number Nt according to the average value calculated by the residual signal average value calculating unit;

a residual data detecting unit for detecting a piece of particular residual data of the residual signal having an absolute value higher than a maximum value expressed by Nt bits determined by the target bit allocation number determining unit from the pieces of residual data of the residual signal;

pulse data value determining unit for determining a particular sample of the digital acoustic signal of time region information corresponding to the particular residual data of the residual signal detected by the residual data detecting unit, determining a pulse value of the pulse data according to the target bit allocation number Nt determined by the target bit allocation number determining unit and the absolute value of the particular residual data detected by the residual data detecting unit to reduce a value of the particular residual data of the residual signal in cases where the residual signal is produced from the digital acoustic signal of time region information in which the pulse data having the pulse value is added to the particular sample and producing the pulse data having the pulse value.