Transmitter, encoding system and method employing use of a bit need determiner for subband coding a digital signal

• US 5,365,553 A
• Filed: 10/27/1993
• Issued: 11/15/1994
• Est. Priority Date: 11/30/1990
• Status: Expired due to Fees

• Current Assignee: US Philips Corporation (Koninklijke Philips N.V.)
• Original Assignee: US Philips Corporation (Koninklijke Philips N.V.)
• Inventors: Keesman, Gerrit J., Veldhuis, Raymond N. J., van der Waal, Robbert G., van de Kerkhof, Leon M.
• Primary Examiner(s): Bocure, Tesfaldet
• Application Number:
  US08/144,093
• Time in Patent Office: 384 Days
• Field of Search: 375/122, 375/25, 370/118, 370/105.1, 370/82, 381/29-32, 381/37, 381/36, 381/35, 360/39, 341/200
• US Class Current: 375/241
• CPC Class Codes: G06T 9/007 Transform coding, e.g. discrete cosine transform
• CPC Class Codes: H04B 1/667 using a division in frequency subbands for TV signals H04N19/63

Abstract

Transmitter, encoding system and method for subband coding a digital signal. The encoding system includes a splitter for dividing the digital signal into subband signals SB₁, . . . , SB_M ; a quantizer unit for quantizing time-equivalent q sample signal blocks of the subband signals; a bit need determiner and a bit allocator. The bit need determiner determines a bit need b_m which corresponds to the number of bits by which the q samples in a time-equivalent signal block in a subband signal SB_m should be represented, where 1≦m≦M. The bit allocator allocates n_m bits to each of the q samples of the time-equivalent signal block of subband signal SB_m on the basis of the bit need b_m and an available bit quantity B, n_m being the number of bits by which the q samples in the time-equivalent signal block of subband signal SB_m will actually be represented, where 1≦m≦M.

Claims

1. An encoding system for encoding a digital signal having a specific sampling frequency and bandwidth, comprising:

splitter means for dividing the bandwidth of the digital signal into M successive subbands, and generating, in response to the digital signal, M subband signals having reduced sampling frequencies, each of the subband signals being associated with one of the subbands;

quantizing means for quantizing time-equivalent signal blocks of the subband signals, a subband signal SB_m of the subband signals having successive signal blocks which each contain q samples of that subband signal, each sample in a signal block of subband signal SB_m having an amplitude and being quantized by n_m bits, where n_m may vary for different signal blocks of subband signal SB_m ;

bit need determining means for determining bit needs for the time-equivalent signal blocks, said bit need determining means comprising;

(a) means for estimating power within the time-equivalent signal blocks, the signal block of subband signal SB_m having a power v_m ;

(b) means for determining scale factors for the time-equivalent signal blocks, a scale factor SF_m for the signal block of subband signal SB_m being determined from a sample therein having a maximum absolute amplitude value;

(c) means for determining masking magnitudes for the time-equivalent signal blocks, the signal block of subband signal SB_m having a masking magnitude w_m which is determined in accordance with the following relationship;

##EQU8## where d_mi v_i denotes masked power in the signal block of subband signal SB_m as a result of power v_i in a time-equivalent signal block of a subband signal SB_i of the subband signals, d_mi denotes a matrix coefficient in an M×

M matrix by which the power v_i is multiplied to determine the masked power in the signal block of subband signal SB_m as a result of the time-equivalent signal block of subband signal SB_i, and w_r.m denotes a masking threshold in the signal block of subband signal SB_m ; and

(d) means for determining the following relationship for the time-equivalent signal blocks;

##EQU9## where K₁, K₂ and K₃ are constants; and

b_m is a bit need for the signal block of subband signal SB_m corresponding to the number of bits by which the q samples in that signal block should be represented, and b_m may vary for different signal blocks of the subband signal SB_m ; and

bit allocation means for allocating bits to the time-equivalent signal blocks from an available number of bits B, n_m bits being allocated to each of the q samples of the signal block of subband signal SB_m in accordance with at least the bit need, b_m, for that signal block;

wherein M, m and i are integers such that 1≦

m≦

M and 1≦

i≦

M;

q and B are integers, where q is greater than unity and B is greater than zero; and

b_m, n_m, v_m, v_i, SF_m, w_m, d_mi and w_r.m are variables, where n_m and SF_m are greater than or equal to zero.

2. The encoding system as claimed in claim 1, wherein K₁ =1, K₂ =1/√

3 and K₃ is preferably equal to either 1 or zero.

3. The encoding system as claimed in claim 2, wherein said means for estimating power estimates the power v_m in the signal block of subband signal SB_m according to the following relationship:

##EQU10## where s_j is the amplitude of a jth sample in that signal block, j being an integer such that 1≦

j≦

q and s_j being a variable.

4. The encoding system as claimed in claim 1, wherein said means for estimating power estimates the power v_m in the signal block of subband signal SB_m according to the following relationship:

##EQU11## where s_j is the amplitude of a jth sample in that signal block, j being an integer such that 1≦

j≦

q and s_j being a variable.

5. The encoding systems as claimed in claim 1, wherein said bit need determining means utilizes a logarithmic representation for d_mi, v_i, w_m and w_r.m when determining the masking magnitude w_m.

6. The encoding system as claimed in claim 5, wherein K₁ =1, K₂ =1/√

3 and K₃ is preferably equal to either 1 or zero.

7. The encoding system as claimed in claim 6, wherein said means for estimating power estimates the power v_m in the signal block of subband signal SB_m according to the following relationship:

##EQU12## where s_j is the amplitude of a jth sample in that signal block, j being an integer such that 1≦

j≦

q and s_j being a variable.

8. The encoding system as claimed in claim 5, wherein said means for estimating power estimates the power v_m in the signal block of subband signal SB_m according to the following relationship:

##EQU13## where s_j is the amplitude of a jth sample in that signal block, j being an integer such that 1≦

j≦

q and s_j being a variable.

9. The encoding system as claimed in claim 5, wherein said bit need determining means further comprises means for adding and multiplying logarithmically represented values.

10. The encoding system as claimed in claim 1, further comprising signal formatting means for assembling into a frame of an output digital signal having successive frames the q samples from the time-equivalent signal blocks of the subband signals which have been quantized by said quantizing means, scale factor information being included in the frame in the form of x-bit words representing the scale factors associated with the time-equivalent signal blocks for which the q samples are included in the frame.

11. A transmitter, comprising the encoding system of claim 1.

12. The transmitter as claimed in claim 11, further comprising signal formatting means for assembling into a frame of an output digital signal having successive frames the q samples from the time-equivalent signal blocks of the subband signals which have been quantized by said quantizing means, scale factor information being included in the frame in the form of x-bit words representing the scale factors associated with the time-equivalent signal blocks for which the q samples are included in the frame.

13. The transmitter as claimed in claim 12, further comprising a recording means for recording the output digital signal in a track on a record carrier.

14. The transmitter as claimed in claim 13, wherein the record carrier is a magnetic record carrier.

18. The method as claimed in claim 11, wherein the power v_m in the signal block of subband signal SB_m is estimated in accordance with following relationship:

##EQU16## where s_j is the amplitude of a jth sample in that signal block, j being an integer such that 1≦

j≦

q and s_j being a variable.

15. A method of encoding a digital signal having a specific sampling frequency and bandwidth, comprising:

dividing the bandwidth of the digital signal into M successive subbands, and generating, in response to the digital signal, M subband signals having reduced sampling frequencies, each of the subband signals being associated with one of the subbands;

quantizing time-equivalent signal blocks of the subband signals, a subband signal SBm of the subband signals having successive signal blocks which each contain q samples of that subband signal, each sample in a signal block of subband signal SB_m having an amplitude and being quantized by n_m bits, where n_m may vary for different signal blocks of subband signal SB_m ;

wherein in order to quantize the time-equivalent signal blocks, the following steps are performed;

determining bit needs for the time-equivalent signal blocks by;

(a) estimating power within the time-equivalent signal blocks, the signal block of subband signal SB_m having a power v_m ;

(b) determining scale factors for the time-equivalent signal blocks, a scale factor SF_m for the signal block of subband signal SB_m being determined from a sample therein which has a maximum absolute amplitude value;

(c) determining masking magnitudes for the time-equivalent signal blocks, the signal block of subband signal SB_m having a masking magnitude w_m which is determined in accordance with the following relationship;

##EQU14## where d_mi v_i denotes masked power in the signal block of subband signal SB_m as a result of power v_i in a time-equivalent signal block of a subband signal SB_i of the subband signals, d_mi denotes a matrix coefficient in an M×

M matrix by which the power v_i is multiplied to determine the masked power in the signal block of subband signal SB_m as a result of the time-equivalent signal block of subband signal SB_i, and w_r.m denotes a masking threshold in the signal block of subband signal SB_m ; and

(d) determining the following relationship for the time-equivalent signal blocks of the subband signals;

##EQU15## where K₁, K₂ and K₃ are constants; and

b_m is a bit need for the signal block of subband signal SB_m corresponding to the number of bits by which the q samples in that signal block should be represented, and b_m may vary for different signal blocks of subband signal SB_m ; and

allocating bits to the time-equivalent signal blocks from an available number of bits B, n_m bits being allocated each of the q samples of the signal block of subband signal SB_m in accordance with the bit need b_m for that signal block;

wherein M, m and i are integers such that 1≦

m≦

M and 1≦

i≦

M;

q and B are integers, where q is greater than unit and B is greater than zero; and

b_m, n_m, v_m, v_i, SF_m, w_m, d_mi and w_r.m are variables, where n_m and SF_m are greater than or equal to zero.

16. The method as claimed in claim 15, wherein in determining the masking magnitude w_m a logarithmic representation is used for d_mi, v_i, w_m and w_r.m.

17. The method as claimed in claim 15, wherein K₁ =1, K₂ =1/√

3 and K₃ is preferably equal to either 1 or zero.

19. The method as claimed in claim 15, further comprising the step of assembling into a frame of an output digital signal having successive frame the q samples from the time-equivalent signal blocks of the subband signals which have been quantized, scale factor information being included in the frame in the form of x-bit words representing the scale factors associated with the time-equivalent signal blocks for which the q samples are included in the frame.

20. The method as claimed in claim 19, further comprising the recording the output digital signal in a track on a record carrier.

21. The method as claimed in claim 20, wherein the record carrier is a magnetic record carrier.

22. An encoding system for encoding a digital signal, comprising:

means for dividing the digital signal into a plurality of subband signals, each of the subband signals having a plurality of signal blocks, each containing q samples of that subband signal, where q is a positive integer, which are successive in time, each of the signal blocks of a subband signal being time-equivalent with a corresponding signal block of each of the other subband signals, corresponding signal blocks of the subband signals constituting time-equivalent signal blocks;

means for quantizing each of the q samples of each of the time-equivalent signal blocks with n_m bits, where n_m is a variable greater than or equal to zero which may vary for the time-equivalent signal blocks and/or different signal blocks within the same subband signal and m is a positive integer denoting which one of the subband signals a signal block comes from;

bit determining means for determining a bit need b_m for each of the time-equivalent signal blocks, where b_m is a variable which may vary for the time-equivalent signal blocks and/or different signal blocks within the same subband signal, the bit need b_m for each of the time-equivalent signal blocks corresponding to the number of bits by which the q samples in that signal block should be represented and being determined on the basis of a scale factor for that signal block, a linear combination of each masked power in that signal block resulting from power in each of the time-equivalent signals blocks and a masking threshold for that signal block; and

means for allocating, from an available number of bits B, where B is a positive integer, the n_m bits to each of the q samples of each of the time-equivalent signal blocks in accordance with the bit need b_m for each of the time-equivalent signal blocks.

23. The encoding system as claimed in claim 22, wherein said bit need determining means comprises:

estimation means for estimating the power within each of the time-equivalent signal blocks;

first determining means for determining the scale factor for each of the time-equivalent signal blocks;

second determining means for determining a masking magnitude for each of the time-equivalent signal blocks, the masking magnitude for a signal block being determined based on the masked power in the signal block as a result of the power in each of the time-equivalent signals blocks and the masking threshold for the signal block; and

third determining means for determining the bit need b_m for each of the time-equivalent signal blocks.

24. The encoding system as claimed in claim 23, wherein said estimation means estimates the power in a signal block in accordance with the following relationship:

##EQU17## where v_m is a variable denoting the power in a signal block and s_j is variable denoting the amplitude of a jth sample in the signal block, j being an integer such that 1≦

j≦

q.

25. The encoding systems as claimed in claim 23, wherein said first determining means determines the scale factor for a signal block from a sample therein having a maximum absolute amplitude value.

26. The encoding systems as claimed in claim 23, wherein said second determining means determines the masking magnitude for a signal block in accordance with the following relationship:

##EQU18## where w_m is a variable denoting the masking magnitude for the signal block, M is a positive integer equal to the number of subband signals, each of which are denoted by i, which is a positive integer such that 1≦

i≦

M, d_mi v_i is the masked power in the signal block as a result of power v_i, where v_i is variable, in one of the time-equivalent signal blocks, which time-equivalent signal block is from subband signal i, d_mi is variable denoting a matrix coefficient in an M×

M matrix by which the power v_i is multiplied to determine the masked power in the signal block as a result of the time-equivalent signal block from subband signal i, and w_r.m is a variable denoting the masking threshold in the signal block.

27. The encoding system as claimed in claim 23, wherein said third determining means determines the bits need b_m for a signal block in accordance with the following relationship:

##EQU19## where K₁, K₂ and K₃ are constants, SF_m is a variable denoting the scale factor for the signal block and wm is a variable denoting the masking magnitude for the signal block.

28. The encoding system as claimed in claim 22, further comprising signal formatting means for assembling into a frame of an output digital signal having successive frames the q samples from the time-equivalent signal blocks which have been quantized, scale factor information being included in the frame in the form of x-bit words representing the scale factors associated with the time-equivalent signal blocks for which the q samples are included in the frame.

29. A transmitter, comprising the encoding system of claim 28.

30. The transmitter as claimed in claim 29, further comprising a recording means for recording the output signal in a track of a record carrier.

31. A transmitter, comprising the encoding system of claim 22.

32. The encoding systems as claimed in claim 22, wherein the subband signals have reduced sampling frequencies as compared to the digital signal.

33. A method for encoding a digital signal, comprising:

dividing the digital signal into a plurality of subband signals, each of the subband signals having a plurality of signal blocks, each containing q samples of that subband signal, where q is a positive integer, which are successive in time, each of the signal blocks of a subband signal being time-equivalent with a corresponding signal block of each of the other subband signals, corresponding signal blocks of the subband signals constituting time-equivalent signal blocks; and

quantizing each of the q samples of each of the time-equivalent signal blocks with n_m bits, where n_m is a variable greater than or equal to zero which may vary for the time-equivalent signal blocks and/or different signal blocks within the same subband signal and m is a positive integer denoting which one of the subband signals a signal block comes from;

wherein in order to quantize each of the time-equivalent signal blocks, the following additional steps are preformed;

determining a bit need b_m for each of the time-equivalent signal blocks, where b_m is a variable which may vary for the time-equivalent signal blocks and/or different signal blocks within the same subband signal, the bit need b_m for each of the time-equivalent signal blocks corresponding to the number of bits by which the q samples in that signal block should be represented and being determined on the basis of a scale factor for that signal block, a linear combination of each masked power in that signal block resulting from power in each of the time-equivalent signals blocks and a masking threshold for that signal block; and

allocating, from an available number of bits B, where B is a positive integer, the n_m bits to each of the q samples of each of the time-equivalent signal blocks in accordance with the bit need b_m for each of the time-equivalent signal blocks.

34. The method as claimed in claim 33, wherein determining the bit need b_m for each of the time-equivalent signal blocks includes:

estimating the power within each of the time-equivalent signal blocks;

determining the scale factor for each of the time-equivalent signal blocks; and

determining a masking magnitude for each of the time-equivalent signal blocks, the masking magnitude for a signal block being determined based on the masked power in the signal block as a result of the power in each of the time-equivalent signals blocks and the masking threshold for the signal block.

35. The method as claimed in claim 34, wherein the power in a signal block is estimated in accordance with the following relationship:

##EQU20## where v_m is a variable denoting the power in a signal block and s_j is variable denoting the amplitude of a jth sample in the signal block, j being an integer such that 1≦

j≦

q.

36. The method as claimed in claim 34, wherein the scale factor for a signal block is determined from a sample therein having a maximum absolute amplitude value.

37. The method as claimed in claim 34, wherein the masking magnitude for a signal block is determined in accordance with the following relationship:

##EQU21## where w_m is a variable denoting the masking magnitude for the signal block, M is a positive integer equal to the number of subband signals, each of which are denoted by i, which is a positive integer such that 1≦

i≦

M, d_mi v_i is the masked power in the signal block as a result of power v_i, where v_i is variable, in one of the time-equivalent signal blocks, which time-equivalent signal block is from subband signal i, d_mi is variable denoting a matrix coefficient in an M×

M matrix by which the power v_i is multiplied to determine the masked power in the signal block as a result of the time-equivalent signal block from subband signal i, and w_r.m is a variable denoting the masking threshold in the signal block.

38. The method as claimed in claim 34, further comprising assembling into a frame of an output digital signal having successive frames the q samples from the time-equivalent signal blocks which have been quantized, scale factor information being included in the frame in the form of x-bit words representing the scale factors associated with the time-equivalent signal blocks for which the q samples are included in the frame.

39. The method as claimed in claim 33, wherein the bit need b_m for a signal block is determined in accordance with the following relationship:

##EQU22## where K₁, K₂ and K₃ are constants, SF_m is a variable denoting the scale factor for the signal block and wm is a variable denoting a masking magnitude for the signal block, the masking magnitude being a function of each masked power in the signal block resulting from power in each of the time-equivalent signals blocks and the masking threshold for the signal block.

40. A bit need determining device for determining bits needs for time-equivalent signal blocks of subband signals, the device comprising:

means for estimating power within each of the time-equivalent signal blocks;

means for determining a scale factor for each of the time-equivalent signal blocks;

means for determining a masking magnitude for each of the time-equivalent signal blocks, the masking magnitude for a time-equivalent signal block being determined based on a linear combination of each masked power in the time-equivalent signal block resulting from the power in each of the time-equivalent signals blocks and a masking threshold for the time-equivalent signal block; and

means for determining a bit need b_m for each of the time equivalent signal blocks, the bit need b_m for a time-equivalent signal block being determined based upon the scale factor and the masking magnitude for the time-equivalent signal block.

41. A bit need determining device for determining bit needs for time-equivalent signal blocks of M subband signals, each of the time-equivalent signal blocks having q samples, where q is a positive integer, the device comprising:

means for estimating power within the time-equivalent signal blocks, the power within a time-equivalent signal block being denoted v_m, where v_m is a variable, and m is a positive integer, such that 1≦

m≦

M, denoting which one of subband signals the time-equivalent signal block comes from;

means for determining scale factors for the time-equivalent signal blocks, a scale factor SF_m, where SF_m is a variable greater than or equal to zero, for a time-equivalent signal block being determined from a sample therein having a maximum absolute amplitude value;

means for determining masking magnitudes for the time-equivalent signal blocks, a time-equivalent signal block having a masking magnitude w_m, where w_m is a variable, which is determined in accordance with the following relationship;

##EQU23## where i is a positive integer, such that 1≦

i≦

M, denoting one of the subband signals, d_mi v_i denotes masked power in the time-equivalent signal block as a result of power v_i, where v_i is variable, in one of the time-equivalent signal blocks, which is from subband signal i, d_mi is variable denoting a matrix coefficient in an M×

M matrix by which the power v_i is multiplied to determine the masked power in the time-equivalent signal block as a result of the one of the time-equivalent signal blocks from subband signal i, and w_r.m is a variable denoting the masking threshold in the signal block; and

means for determining the bit need b_m for the time-equivalent signal block, the bit need bm for a time-equivalent signal block being determined in accordance with the following relationship;

##EQU24## where K₁, K₂ and K₃ are constants.

42. The bit need determining device as claimed in claim 41, wherein K₁ =1, K₂ =1/√

3 and K₃ is preferably equal to either 1 or zero.

43. The bit need determining device as claimed in claim 41, wherein said means for estimating power estimates the power v_m in the time-equivalent signal block according to the following relationship:

##EQU25## where S_j is variable denoting the amplitude of a jth sample in the time-equivalent signal block, j being an integer such that 1≦

j≦

q.

44. The bit need determining device as claimed in claim 41, wherein said bit need determining means utilizes a logarithmic representation for the values of d_mi, v_i, w_m and w_r.m when determining the masking magnitude for the time-equivalent signal block.

45. The bit need determining device as claimed in claim 44, further comprising means for adding and multiplying the logarithmically represented values.

46. A transmitter, comprising the bit need determining device claimed in claim 41.

47. A method for determining bits needs for time-equivalent signal blocks of subband signals, the device comprising:

estimating power within each of the time-equivalent signal blocks;

determining a scale factor for each of the time-equivalent signal blocks;

determining a masking magnitude for each of the time-equivalent signal blocks, the masking magnitude for a time-equivalent signal block being determined based on each masked power in the time-equivalent signal block resulting from the power in each of the time-equivalent signals blocks and a masking threshold for the time-equivalent signal block; and

determining a bit need b_m for each of the time equivalent signal blocks, the bit need b_m for a time-equivalent signal block being determined based upon the scale factor and the masking magnitude for the time-equivalent signal block.

48. A method for determining bit needs for time-equivalent signal blocks of M subband signals, each of the time-equivalent signal blocks having q samples, where q is a positive integer, the device comprising:

estimating power within the time-equivalent signal blocks, the power within a time-equivalent signal block being denoted v_m, where v_m is a variable, and m is a positive integer, such that 1≦

m≦

M, denoting which one of subband signals the time-equivalent signal block comes from;

determining scale factors for the time-equivalent signal blocks, a scale factor SF_m, where SF_m is a variable greater than or equal to zero, for a time-equivalent signal block being determined from a sample therein having a maximum absolute amplitude value;

determining masking magnitudes for the time-equivalent signal blocks, a time-equivalent signal block having a masking magnitude w_m, where w_m is a variable, which is determined in accordance with the following relationship;

##EQU26## where i is a positive integer, such that 1≦

i≦

M, denoting one of the subband signals, d_mi v_i denotes masked power in the time-equivalent signal block as a result of power v_i, where v_i is variable, in one of the time-equivalent signal blocks, which is from subband signal i, d_mi is variable denoting a matrix coefficient in an M×

M matrix by which the power v_i is multiplied to determine the masked power in the time-equivalent signal block as a result of the one of the time-equivalent signal blocks from subband signal i, and w_r.m is a variable denoting the masking threshold in the signal block; and

determining the bit need b_m for the time-equivalent signal block, the bit need bm for a time-equivalent signal block being determined in accordance with the following relationship;

##EQU27## where K₁, K₂ and K₃ are constants.

49. The method as claimed in claim 48, wherein K₁ =1, K₂ =1/√

3 and K₃ is preferably equal to either 1 or zero.

50. The method as claimed in claim 48, wherein the power v_m in the time-equivalent signal block is estimated according to the following relationship:

##EQU28## where s_j is variable denoting the amplitude of a jth sample in the time-equivalent signal block, j being an integer such that 1≦

j≦

q.

51. The method as claimed in claim 48, wherein a logarithmic representation for the values of d_mi, v_i, w_m and w_r.m are utilized in determining the masking magnitude for the time-equivalent signal block.