JP3968799B2 - Information signal encoding apparatus, encoding method, and information signal decoding method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an information signal encoding apparatus, encoding method, and decoding method that reduce the amount of generated data of digital information signals such as digital audio signals and digital image signals.
[0002]
[Prior art]
Predictive coding is known to reduce the amount of transmission information such as digital audio signals and digital image signals. For example, the one-dimensional DPCM forms a difference (residual) between the input sample value and the predicted value in the time direction, and the two-dimensional DPCM forms a residual between the input sample value and the predicted value in the spatial direction. Since the digital information signal has a correlation in the time direction and the spatial direction, the residual is concentrated in the vicinity of zero. Therefore, the residual signal can be quantized with a smaller number of bits than the original number of quantization bits, thereby reducing the amount of information. Further, the amount of information can be further compressed by performing variable length coding using the concentration of the distribution of residual signals. As the variable length coding, run length and Huffman coding are known.
[0003]
Furthermore, in order to reduce the quantization error, the quantization step of the value near 0 of the residual signal is made fine, and when the residual signal level is large, nonlinear quantization is also known, which makes the quantization step coarse. It has been.
[0004]
[Problems to be solved by the invention]
As one method for quantizing the residual signal, it has been proposed to collect the residual signal for each block, narrow the range in which the residual signal level exists, and reduce the number of quantization bits. . As described above, in the case of a residual signal that has been blocked, the level of the residual signal is not necessarily concentrated to 0 depending on the block, but has various level distributions. Accordingly, when the entire level range in which the blocked residual signal can exist is quantized by linear quantization, there is a problem that the quantization error increases when the number of quantization bits is reduced.
[0005]
Therefore, an object of the present invention is to provide an information signal encoding apparatus, an encoding method, and an encoding method capable of reducing the quantization error by selecting a quantization characteristic adapted to the distribution of the level of the blocked signal. And providing a decoding method.
[0006]
[Means for Solving the Problems]
  The invention according to claim 1 is an information signal encoding apparatus for encoding an input digital information signal so as to reduce the amount of generated data.
  Means for blocking the input digital information signal;
  Detecting means for detecting a minimum value and a dynamic range for each block of the blocked input digital information signal;
  Means for detecting the presence range of the level of the blocked input digital information signal based on the minimum value and the dynamic range, and generating a selection signal corresponding to the detected presence range;
  Quantization value assignment is denser in a given quantization range than in other quantization ranges, and the given quantization range is differentAmong a plurality of quantization characteristics, including the existence range, andAbove givenQuantization means for selecting a quantization characteristic having the smallest difference between the quantization range and the existence range by the selection signal, and quantizing the input digital information signal blocked by the selected quantization characteristic;
  Means for transmitting the output of the quantizing means and the selection signal;
  An information signal encoding device characterized by comprising:
  The invention according to claim 2 is an information signal encoding apparatus that applies the above-described quantization to a block residual signal.
[0007]
  The invention according to claim 3 is an information signal encoding method for encoding an input digital information signal so as to reduce the amount of generated data.
  Blocking the input digital information signal;
  Detecting a block-by-block minimum value and dynamic range of the blocked input digital information signal;
  Detecting a presence range of the level of the blocked input digital information signal based on the minimum value and the dynamic range, and generating a selection signal corresponding to the detected presence range;
  Quantization value assignment is denser in a given quantization range than in other quantization ranges, and the given quantization range is differentAmong a plurality of quantization characteristics, including the existence range, andAbove givenA quantization characteristic having the smallest difference between the quantization range and the existence range is selected by the selection signal, and the input digital information signal blocked by the selected quantization characteristic is quantized;
  Transmitting the output generated by the quantization and the selection signal;
  An information signal encoding method characterized by comprising:
  A fourth aspect of the present invention is an information signal encoding method in which the above-described quantization is applied to a block residual signal.
[0008]
  According to a fifth aspect of the present invention, there is provided an information signal encoding apparatus in which at least first and second hierarchical data are formed from an input digital information signal, and the first and second hierarchical data are encoded and transmitted. ,
  Means for forming second hierarchical data having a lower resolution than the first hierarchical data;
  Means for predicting the first hierarchical data from the second hierarchical data;
  Means for forming a residual signal between the predicted data and the first tier data;
  Means for blocking the residual signal;
  Detecting means for detecting a minimum value and a dynamic range for each block of the residual signal that has been blocked;
  Means for detecting an existing range of the level of the blocked residual signal based on the minimum value and the dynamic range and generating a selection signal corresponding to the detected existing range;
  Quantization value assignment is denser in a given quantization range than in other quantization ranges, and the given quantization range is differentAmong a plurality of quantization characteristics, including the existence range, andAbove givenQuantization means for quantizing the residual signal blocked by the selected quantization characteristic, the quantization characteristic having the smallest difference between the quantization range and the existence range is selected, and
  Means for transmitting the output of the quantizing means and the selection signal;
  An information signal encoding device characterized by comprising:
[0009]
  According to a sixth aspect of the present invention, there is provided an information signal encoding method in which at least first and second layer data are formed from an input digital information signal, and the first and second layer data are encoded and transmitted. ,
  Forming second hierarchical data having a lower resolution than the first hierarchical data;
  Predicting first hierarchical data from second hierarchical data;
  Forming a residual signal between the predicted data and the first tier data;
  Blocking the residual signal;
  Detecting a block-by-block minimum value and dynamic range of the blocked residual signal;
  Detecting a presence range of the level of the blocked residual signal based on the minimum value and the dynamic range, and generating a selection signal corresponding to the detected presence range;
  Quantization value assignment is denser in a given quantization range than in other quantization ranges, and the given quantization range is differentAmong a plurality of quantization characteristics, including the existence range, andAbove givenA quantization characteristic having the smallest difference between the quantization range and the existence range is selected by the selection signal, and the residual signal blocked by the selected quantization characteristic is quantized;
  Transmitting the output generated by the quantization and the selection signal;
  An information signal encoding method characterized by comprising:
[0010]
  The invention according to claim 7 is selected according to a selection signal corresponding to the existence range of the level of the residual signal detected based on the minimum value and the dynamic range of the block.Quantization value assignment is denser in a given quantization range than in other quantization ranges, and the given quantization range is differentAmong a plurality of quantization characteristics, including the existence range, andPredeterminedIn an information signal decoding method for an encoding method in which a block residual signal is quantized by a quantization characteristic having a smallest difference between a quantization range and an existence range, and a quantization output and a selection signal are transmitted.
  Receiving a selection signal and selecting an inverse quantization characteristic;
  An inverse quantization step for converting the quantized output into a representative value according to the selected inverse quantization characteristic;
  An information signal decoding method comprising the steps of: decomposing a residual signal converted into a representative value into blocks and converting the residual signal into an original order.
[0011]
The residual signal between the input digital information signal and the prediction signal is blocked. The minimum value and dynamic range of the residual signal in the block is detected. Based on the minimum value and the dynamic range, the distribution residual distribution of the block is examined. A quantization characteristic adapted to this level distribution bias is selected. Therefore, it is possible to reduce the quantization error. A quantization output and a selection signal for selecting a quantization characteristic are transmitted.
[0012]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In this embodiment, the present invention is applied to a digital image signal obtained by sampling a video signal at a predetermined sampling frequency and converting each sample into a predetermined number of quantization bits. FIG. 1 shows the overall configuration of a system according to an embodiment of the present invention.
[0013]
In FIG. 1, a digital video signal is supplied to an input terminal indicated by 121. The input signal is supplied to the subtractor 123, the output (residual signal) of the subtractor 123 is supplied to the blocking circuit 124 and the predictor 122, and the prediction signal generated by the predictor 122 is supplied to the subtractor 123. . The subtractor 123 subtracts the prediction signal from the input signal to generate a prediction residual. This residual signal is supplied to the block forming circuit 124 and converted from raster scan order to block order data. The blocked residual signal is supplied to the encoding unit 125.
[0014]
As will be described later, the encoding unit 125, when re-quantizing the residual signal with a smaller number of quantization bits than the original number of quantization bits, has a quantization characteristic adapted to the range of residual signal levels. It has the structure selected. The encoded output (quantized output and select signal) of the encoding unit 125 is supplied to the error correction code encoder 126, and a redundant code of the error correction code is added. The output of the error correction code encoder 126 is supplied to the modulation unit 127. The modulation unit 127 modulates the digital signal into a form suitable for recording, transmission, and the like. An output signal from the modulation unit 127 is supplied to the recording unit 128, and the recording signal is recorded on the information signal recording medium 129 by the recording unit 128. It is also possible to transmit data via the transmission path 130. In this case, a transmission unit is used instead of the recording unit 128. The information signal recording medium 129 is a disk-like or tape-like recording medium using magnetism, magneto-optical, phase change or the like. A semiconductor memory is also a kind of recording medium.
[0015]
Data is reproduced from the recording medium 129 by the reproduction unit 131 or data transmitted via the transmission path 130 is received. The data reproduced by the reproduction unit 131 is demodulated by the demodulator 132, and the demodulated output is supplied to the error correction code decoder 133. The decoder 133 corrects the error by using the redundant code, and corrects the remaining error that cannot be corrected to be inconspicuous.
[0016]
The output of the error correction decoder 133 is supplied to the decoding unit 134. As will be described later, the decoding unit 134 selects the inverse quantizer corresponding to the used quantizer according to the received select signal, and the selected inverse quantization, as opposed to the encoding unit 125. The inverse quantization process performed by the device is performed. A decoded residual signal is generated from the decoding unit 134. This decoded residual signal is supplied to the block decomposition circuit 135. In the block decomposition circuit 135, the block structure is returned to the order of raster scanning.
[0017]
The decoded residual signal is supplied to the adder circuit 136. A decoding image signal is formed by the adding circuit 136 and taken out to the output terminal 137. Further, the decoded image signal is supplied to the predictor 138, and a prediction signal is generated. The prediction signal is supplied to the addition circuit 136.
[0018]
FIG. 2 shows an example of the encoding unit 125. The blocked residual signal from the blocking circuit 124 is supplied to the input terminal 1. FIG. 3 schematically shows the formation of the residual signal. One rectangular area in FIG. 3 corresponds to one pixel. Each of a to h represents a locally decoded pixel value, and AP represents a pixel value before being encoded. A predicted value A ′ for pixel value A is generated by the predictor 122 using neighboring local decoded pixel values. For example, the predicted value A ′ is formed according to a prediction formula such as A ′ = 4c−3 (b−f), A ′ = f + c−b, or the like. The predicted values for the pixel values B, C,... Are calculated by the same prediction formula.
[0019]
The subtractor 123 subtracts the predicted value (for example, A ′) from the pixel value (for example, A) to generate a residual signal Δa. Similarly, residual signals Δb, Δc,... Are generated. In the blocking circuit 124, the generated residual signal is converted into a block structure. For example, the block forming circuit 124 forms data of blocks of residual signals Δa to Δp corresponding to the (4 × 4) block, as shown by the thick line frame in FIG. 3A. When a digital audio signal is handled, a prediction value in the time direction is formed, and a one-dimensional residual signal block is formed.
[0020]
The residual signal can be increased in concentration by blocking. When one pixel is 8-bit data, the distribution of the occurrence frequency of the residual signal of one screen is in the range of (−255 to +255) around 0, and the frequency of the residual is 0. Maximum. However, when divided into blocks, the residual level distribution has a narrower range than the original distribution. In the case of dividing into blocks, the maximum frequency of the residual does not necessarily match 0.
[0021]
This is due to the fact that the residuals in the blocks in the small space are relatively large in comparison with one screen, and the residuals are strongly correlated in the blocks. Further, the fact that the frequency of the value of 0 does not become maximum occurs when the luminance level gradually changes in the diagonal direction in the block, for example. The method of increasing the degree of concentration of the residual level distribution is an example of blocking, and other methods are also possible.
[0022]
Returning to FIG. 2, the encoding unit 125 will be described. The residual signal from the input terminal 1 is supplied to the maximum value detection circuit 2, the minimum value detection circuit 3, and the delay circuit 5. The maximum value detection circuit 2 detects the maximum value MAX for each block, and the minimum value detection circuit 3 detects the minimum value MIN for each block. The detected maximum value MAX and minimum value MIN are supplied to the subtraction circuit 4 and the select signal generation circuit 7.
[0023]
In the subtraction circuit 4, the dynamic range DR of the residual signal of the block is detected by the operation of MAX-MIN. This dynamic range DR is supplied to the select signal generation circuit 7. The select signal generation circuit 7 generates a select signal SL, and this signal SL is supplied to the distributor 6 as a control signal. The distributor 6 is supplied with the residual signal phase-matched by the delay circuit 5.
[0024]
A plurality of quantizers 8 having different quantization characteristics from each other with respect to the distributor 6₁~ 8_nIs connected. In response to the select signal SL, the distributor 6 selectively converts the residual signal into the quantizer 8.₁~ 8_nSupply to one of the. In the quantizer, the residual signal is quantized. Quantizer 8₁~ 8_nThe quantization characteristics will be described later. As an example, when each pixel is quantized with 8 bits, the value is included in the range of 0 to +255, and the value of the residual signal can be −255 to +255. Since the existence range of the level of the block residual signal is more concentrated, the number of bits of the residual signal is 8 bits as the quantization bit number of the residual signal while suppressing an increase in quantization error. A smaller number of bits can be employed. Quantizer 8₁~ 8_nQuantized output q₁~ Q_nAnd the select signal SL are supplied to the framing circuit 9. A select signal and a quantized output are output to the output terminal 10 of the framing circuit 9 as data having a predetermined frame structure.
[0025]
As described above, the residual signal generated by the predictive coding has various biases depending on the block, as shown in FIG. 4A. In FIG. 4A, the residual signal indicated by a is biased to the negative side, the residual signal indicated by b matches the value with the maximum frequency of 0, and the residual signal indicated by c is positive. It is biased to. An example of the residual signal of the block shown in FIG. 4B is (MIN = 0, MAX = + 255, DR = 255). That is, it is a residual signal having a level distribution biased to the positive side. An example of the residual signal of the block shown in FIG. 4C is (MIN = −255, MAX = 0, DR = 255). That is, it is a residual signal having a level distribution biased to the negative side.
[0026]
In general, if the code has n bits, 2ⁿOr 2ⁿThe range between MAX and MIN is divided into a range of −1. At the time of quantization, the same code is assigned to the residual signal belonging to the divided range. When the code is restored, it is converted to a representative value in the middle of the range. FIG. 5A shows the characteristics of conventional linear quantization. Here, n = 3, and the range between MAX and MIN is divided into seven ranges. On the vertical axis (output level), x indicates a representative value restored by inverse quantization. The conventional linear quantization is such that the quantization distortion is uniform for the seven level ranges.
[0027]
However, in the case of the residual signal shown in FIG. 4B, since the residual signal actually exists in the positive range, three quantized values are assigned to the positive and negative sides, respectively. In the case of linear quantization, assignment of the quantization value to the negative side is useless. Further, in the case of the residual signal shown in FIG. 4C, the allocation of the quantized value to the positive side is wasted. Therefore, in one embodiment of the present invention, non-linear quantization can be performed in consideration of the deviation of the distribution of the residual signal. For example, when the existence range of the residual signal is biased to the positive side, as shown in FIG. 5B, a nonlinear quantum such that five quantized values are assigned to the positive side and one quantized value is assigned to the negative side. To do. On the contrary, when the residual signal existence range is biased toward the negative side, nonlinear quantization is performed so that more quantization values are assigned to the negative side.
[0028]
A plurality of quantizers 8 connected to the distributor 6₁~ 8_nAs described above, each has a quantization characteristic considering the existence range of the residual signal. Quantizer 8 if necessary₁~ 8_nOne of them has the same linear quantization characteristic as the conventional one, and the other has a nonlinear quantization characteristic. The select signal generation circuit 7 can check whether the presence range of the residual signal is centered on 0 or biased to one of positive and negative from the dynamic range DR and the minimum value MIN. More specifically, the minimum value MIN and the dynamic range DR are respectively compared with threshold values, and a multi-bit select signal SL including information on bias and existence range is formed. The distributor 6 is controlled by the select signal SL, and the residual signal of the block is supplied to the quantizer having an appropriate quantization characteristic.
[0029]
FIG. 6 shows another example of the quantization characteristic selected based on the dynamic range DR and the minimum value MIN. The vertical axis in FIG. 6 indicates the dynamic range DR (0 to +511), and the horizontal axis indicates the minimum value MIN (−255 to +255). The residual signal may exist below the diagonal line connecting DR = + 511 and MIN = + 255. In FIG. 6, by drawing a line that bisects the dynamic range DR and a line that equally divides the minimum value MIN into four, the residual signal that may exist is converted into six quantization characteristics Q1 to Q6 and the conventional one not shown. Can be quantized by the linear quantization characteristic of
[0030]
The quantization characteristic Q1 quantizes a residual signal having a distribution biased to the positive side. The quantization characteristic Q2 quantizes the residual signal having a distribution slightly biased to the positive side. The quantization characteristics Q3 and Q4 quantize the residual signal having a distribution slightly biased to the negative side. The quantization characteristics Q5 and Q6 quantize the residual signal having a distribution biased to 0 and negative. Also, in these quantization characteristics Q1 to Q6, it is not necessary to share the number of quantization bits, and the quantization characteristics Q₁, Q₂, Q_Three, Q_FiveThe number of quantization bits of n₁And other quantization characteristics Q_Four, Q₆The number of bits in n₂You may make it differ. In this case, for example, n₂> N₁It is said. In the case where the residual signal exists over a plurality of ranges, linear quantization processing similar to the conventional one is performed.
[0031]
If a plurality of different quantization characteristics can be set so as to cover the existence range of the residual signal, quantization characteristics other than those shown in FIG. 6 may be set. For example, a characteristic for quantizing a range including 0 of the residual signal may be provided separately.
[0032]
An example of the decoding unit 134 of FIG. 2 will be described with reference to FIG. Reproduced or received data from the error correction decoder 133 is supplied to the input terminal 21 of the decoding unit 134. The frame decomposition circuit 22 selects the select signal SL and the quantized output q (q₁~ Q_nOne of them). The separated select signal SL is held in the register 23, and the separated quantized output is supplied to the distributor 24.
[0033]
Inverse quantizer 25 for distributor 24₁~ 25_nIs connected. Similar to the distributor 6 of the encoding unit 125, the distributor 24 is controlled by the select signal SL, and the quantized output from the frame decomposition circuit 22 is supplied to the corresponding inverse quantizer. Inverse quantizer 25₁~ 25_nIs the quantizer 8 of the encoding unit 125.₁~ 8_nConversely, the quantized value is converted into a representative value. The restored representative value is taken out to the output terminal 26. The decoded residual signal taken out to the output terminal 26 is supplied to the block decomposition circuit 125 (see FIG. 1).
[0034]
Next, another embodiment in which the present invention is applied to hierarchical coding will be described. The hierarchical encoding device described here can prevent an increase in the number of encoding target pixels by performing prediction between hierarchies and using a simple arithmetic expression for the inter-layer data.
[0035]
The hierarchical encoding method will be described with reference to FIG. FIG. 8 shows, as an example, a schematic diagram between four hierarchies having the first hierarchy as the lowest hierarchy (original picture) and the fourth hierarchy as the highest hierarchy. For example, when the average value of the lower layer data of four pixels corresponding spatially is adopted as the upper layer data generation method, the number of transmission pixels can be prevented from increasing.
[0036]
That is, the upper layer data is M and the lower layer pixel value is x.₀ , X₁ , X₂ , X_Three Then,
M = 1/4 · (x₀ + X₁ + X₂ + X_Three)
As a result, data M is formed. And, for example, x of the data M and the four data_ThreeThe other three data are transmitted. On the receiving or playback side,
x_Three = 4 · M- (x₀ + X₁ + X₂ )
Non-transmission pixel x by a simple arithmetic expression_Three Can be easily restored. In FIG. 8, hatched rectangles indicate non-transmission pixels.
[0037]
FIG. 9 shows a configuration of hierarchical encoding of, for example, five layers using the above-described averaging. Assume that the first hierarchy is the resolution level of the input image. The block size of this first layer is (1 × 1). The second layer data is generated by averaging four pixels of the first layer data. In this example, the first hierarchy data X₁ (0) to X₁ Based on the average value of (3), the second hierarchical data X₂ (0) is generated. X₂ Second layer data X adjacent to (0)₂ (1) to X₂ Similarly, (3) is generated by averaging four pixels of the first layer data. The block size of the second hierarchy is (1/2 × 1/2).
[0038]
Further, the third layer data is generated by averaging four pixels of the second layer data corresponding spatially. The block size of the third hierarchy is (1/4 × 1/4). Similarly, the fourth layer data is generated from the third layer data. The block size of the fourth layer is (1/8 × 1/8). Finally, the fifth hierarchy data X which is the highest hierarchy_Five (0) is the fourth layer data X_Four (0) to X_Four It is generated by the average value of (3). The block size of the fifth layer is (1/16 × 1/16).
[0039]
By applying the class classification adaptive prediction to the upper layer data for the hierarchical structure data in which the increase in the number of encoding target pixels is prevented, the lower layer data is predicted, and the difference between the lower layer data and the predicted value ( That is, it is possible to reduce the amount of transmission data by generating a residual signal. FIG. 10 shows such a coding unit. The first hierarchy data d0 is supplied as input image data d0 to the averaging circuit 42 and the subtractor 46 via the input terminal 41. The first layer data is image data of the original resolution.
[0040]
The input pixel data d0 is subjected to ¼ averaging processing in the averaging circuit 42 to generate hierarchical data d1. This hierarchical data d1 corresponds to the second hierarchical data shown in FIG. The generated hierarchical data d1 is supplied to the averaging circuit 43 and the subtractor 47.
[0041]
The averaging circuit 43 performs the same processing as the averaging circuit 42 on the hierarchical data d1, and generates hierarchical data d2. This hierarchical data d2 corresponds to the third hierarchical data. The generated hierarchical data d2 is supplied to the averaging circuit 44 and the subtracter 48. Similarly, the averaging circuit 44 performs ¼ averaging processing on the hierarchical data d2 to generate hierarchical data d3. This hierarchical data d3 corresponds to the fourth hierarchical data. The generated hierarchical data d3 is supplied to the averaging circuit 45 and the subtracter 49. Further, the averaging circuit 45 similarly performs a 1/4 averaging process on the hierarchical data d3 to generate hierarchical data d4. This hierarchical data d4 corresponds to the fifth hierarchical data. The generated hierarchical data d4 is supplied to the quantizer 54.
[0042]
And about these five hierarchical data, prediction is performed between hierarchies. First, the quantization processing for compression performed in the fifth layer is performed by the quantizer 54. The output data d21 of the quantizer 54 is supplied to the variable length code encoder 71 and also to the inverse quantizer 58. The output of the encoder 71 is taken out to the output terminal 76 as fifth-layer data. The output data d16 of the inverse quantizer 58 supplied with the encoded data d21 is supplied to the class classification adaptive prediction circuit 62.
[0043]
In the class classification adaptive prediction circuit 62, prediction processing is performed using the data d16, a predicted value d12 of the fourth layer data is generated, and the predicted value d12 is supplied to the subtractor 49. In the subtractor 49, a difference value between the hierarchical data d3 supplied from the averaging circuit 44 and the predicted value d12 is obtained, and the difference value d8 is supplied to the quantizer 53.
[0044]
In the quantizer 53 to which the difference value d8 is supplied, requantization is performed so as to reduce the number of quantization bits as in the quantizer 54. The output data of the quantizer 53 is supplied to the arithmetic unit 66 and the inverse quantizer 57. In this calculator 66, a process of thinning out one pixel from four pixels is performed. Data d20 output from the arithmetic unit 66 is encoded by the variable-length code encoder 70, and the output of the encoder 70 is taken out as output data of the fourth layer to the output terminal 75.
[0045]
The fourth layer data d12 predicted by the class classification adaptive prediction circuit 62 and the output data (decoded residual signal) d15 of the inverse quantizer 57 are supplied to the class classification adaptive prediction circuit 61. In the class classification adaptive prediction circuit 61, the data d15 is added to the data d12 to form local decoded data of the fourth layer, and prediction processing is performed using the local decoded data. A predicted value d11 is generated, and the predicted value d11 is supplied to the subtracter 48. In the subtracter 48, a difference value between the data d2 supplied from the averaging circuit 43 and the predicted value d11 is obtained, and the difference value d7 is supplied to the quantizer 52.
[0046]
The output data of the quantizer 52 supplied with the difference value d7 is supplied to the arithmetic unit 65 and the inverse quantizer 56. In this calculator 65, a process of thinning out one pixel from four pixels is performed. The third layer data d19 output from the arithmetic unit 65 is supplied to the variable length code encoder 69, and the output of the encoder 69 is taken out to the output terminal 74 as third layer data.
[0047]
The third layer data d11 predicted by the class classification adaptive prediction circuit 61 and the output data d14 of the inverse quantizer 56 supplied with the encoded data from the quantizer 52 are supplied to the class classification adaptive prediction circuit 60. In the class classification adaptive prediction circuit 60, the data d14 is added to the data d11 to form third-layer local decoded data, and the local decoded data is used to perform prediction processing. A predicted value d10 is generated, and the predicted value d10 is supplied to the subtractor 47. In the subtractor 47, a difference value between the data d1 supplied from the averaging circuit 42 and the predicted value d10 is obtained, and the difference value d6 is supplied to the quantizer 51.
[0048]
The output data of the quantizer 51 is supplied to the arithmetic unit 64 and the inverse quantizer 55. The calculator 64 performs a process of thinning out one pixel from four pixels. The second layer data d18 output from the computing unit 64 is supplied to the variable length code encoder 68, and the output of the encoder 68 is taken out to the output terminal 73 as second layer data.
[0049]
The second hierarchy data d10 predicted by the class classification adaptive prediction circuit 60 and the output data d13 of the inverse quantizer 55 supplied with the encoded data from the quantizer 51 are supplied to the class classification adaptive prediction circuit 59. In the class classification adaptive prediction circuit 59, the data d13 is added to the data d10 to form second-level local decoded data, and prediction processing is performed using the local decoded data. A predicted value d9 is generated, and the predicted value d9 is supplied to the subtractor 46. In the subtractor 46, a difference value between the input pixel data d0 supplied from the input terminal 41 and the predicted value d9 is obtained, and the difference value d5 is supplied to the quantizer 50.
[0050]
The output data of the quantizer 50 supplied with the difference value d5 is supplied to the calculator 63. The calculator 63 performs a process of thinning out one pixel from four pixels. The first layer data d17 output from the computing unit 63 is supplied to the variable length code encoder 67, and the output of the encoder 67 is taken out to the output terminal 72 as the first layer data.
[0051]
Each of the class classification adaptive prediction circuits 59, 60, 61, and 62 predicts a lower layer pixel to be predicted based on a level distribution of a plurality of spatially neighboring pixels (included in the upper layer). It is. FIG. 12 shows an example of a class classification adaptive prediction circuit. Local decoded data from the input terminal 141 is supplied to the peripheral code value forming unit 142. The peripheral code value forming unit 142 includes a plurality of pieces of data x located in the vicinity of the lower layer pixel to be predicted.₁, X₂, ..., x_nAre synchronized. The peripheral code value is supplied to the class classification unit 143 and the delay unit 145. The class classification unit 143 calculates the peripheral code value x₁~ X_nThe class code corresponding to the level distribution pattern is output. As the class code, the peripheral code value itself may be used. However, since the number of classes becomes enormous, a code obtained by compressing each bit number of the peripheral code to 1 bit by ADRC or the like is used. The class code generated from the class classification unit 143 is supplied to the prediction coefficient memory 144 as an address signal.
[0052]
The prediction coefficient memory 144 stores the prediction coefficient w acquired in advance by learning.₁~ W_nIs stored for each address. That is, using a teacher signal (for example, fourth layer data) and an input signal (fifth layer data formed by averaging from the fourth layer data), a plurality of input signal data and coefficients A predicted value is obtained by linear linear combination of the above, and a coefficient that minimizes the square sum of errors between the predicted value and the true value of the teacher signal is obtained for each class by the method of least squares. Prediction coefficients w1 to w read from the prediction coefficient memory 144 corresponding to the class code._nAnd the peripheral code value x from the delay unit 145₁~ X_nAre supplied to the prediction calculation unit 146.
[0053]
In the prediction calculation unit 146, the predicted value y is calculated by the following linear linear combination formula.
y = w₁x₁+ W₂x₂+ ... + w_nx_n
The predicted value y obtained by the prediction calculation unit 146 is taken out to the output terminal 147. Note that the peripheral code value used for classification and the peripheral code value used for prediction calculation may be different.
[0054]
A configuration similar to that of the encoding unit 125 (see FIG. 2) in the above-described embodiment is also provided on the encoder side of the hierarchical encoding. That is, like the encoding unit 125, each of the quantizers 50, 51, 52, 53, and 54 has a configuration in which the quantization characteristics are controlled in accordance with the existence range of the residual signal.
[0055]
Next, FIG. 11 shows a configuration example on the decoder side of the hierarchical encoding corresponding to the encoder described above. Each hierarchical data generated on the encoder side is supplied to input terminals 81, 82, 83, 84, 85 as d30 to d34, respectively. The variable length code decoders 86, 87, 88, 89, 90 decode the variable length code. Inverse decoders 91, 92, 93, 94, and 95 are connected to these decoders, respectively.
[0056]
First, the fifth hierarchy input data d34 is subjected to decoding processing corresponding to the quantization performed by the encoder in the inverse quantizer 95, and becomes image data d39, which is supplied to the class classification adaptive prediction circuit 107 and the arithmetic unit 103. Is done. Further, the image data d39 is taken out from the output terminal 112 as the image output of the fifth hierarchy.
[0057]
In the class classification adaptive prediction circuit 107, class classification adaptive prediction is performed on the image data of the fourth layer, and the predicted value d47 of the fourth layer data is generated. The adder 99 adds the data d38 (that is, the difference value) from the inverse quantizer 94 and the predicted value d47. The image data d43 is supplied from the adder 99 to the calculator 103, and the calculator 103 performs the above-described calculation to obtain the value of the non-transmission pixel, and the image data d39 supplied from the inverse quantizer 95 and All pixel values in the fourth layer are restored from the image data d43. In this calculator 103, all the restored pixel values are supplied as image data d51 to the class classification adaptive prediction circuit 106 and the calculator 102. The image data d51 is taken out from the output terminal 111 as the output of the fourth hierarchy.
[0058]
In the class classification adaptive prediction circuit 106, class classification adaptive prediction is performed on the image data of the third hierarchy, and the predicted value d46 of the third hierarchy data is generated. The adder 98 adds the data d37 from the inverse quantizer 93 and the predicted value d46. The image data d42 is supplied from the adder 98 to the computing unit 102, the value of the non-transmission pixel is obtained by the computing unit 102, and all the pixel values of the third hierarchy are obtained from the image data d51 and the image data d42 supplied from the computing unit 103. Is restored. In this computing unit 102, all the restored pixel values are supplied as image data d50 to the class classification adaptive prediction circuit 105 and the computing unit 101. The image data d50 is taken out from the output terminal 110 as the output of the third hierarchy.
[0059]
Also, the class classification adaptive prediction circuit 105 performs class classification adaptive prediction on the second layer image data, and generates a predicted value d45 of the second layer data. The adder 97 adds the data d36 from the inverse quantizer 92 and the predicted value d45. The image data d41 is supplied from the adder 97 to the computing unit 101, the value of the non-transmission pixel is obtained by the computing unit 101, and all the pixel values of the second hierarchy are obtained from the image data d50 and the image data d41 supplied from the computing unit 102. Is restored. In this computing unit 101, all the restored pixel values are supplied as image data d49 to the class classification adaptive prediction circuit 104 and the computing unit 100. Further, the image data d49 is taken out from the output terminal 109 as the output of the second hierarchy.
[0060]
Further, the class classification adaptive prediction circuit 104 performs class classification adaptive prediction on the first layer image data, and generates a predicted value d44 of the first layer data. The adder 96 adds the data d35 from the inverse quantizer 91 and the predicted value d44. The image data d40 is supplied from the adder 96 to the computing unit 100, the value of the non-transmission pixel is obtained by the computing unit 100, and all the pixel values of the first layer are obtained from the image data d49 and the image data d40 supplied from the computing unit 101. Is restored. In this computing unit 100, all restored pixel values are taken out from the output terminal 108 as image data d48 as the output of the first layer. In this way, it is possible to improve coding efficiency by introducing class classification adaptive prediction in hierarchical coding in which an increase in the number of pixels to be coded is prevented.
[0061]
The inverse quantizers 91, 92, 93, 94, and 95 have the same configuration as that of the decoding unit 134 of the above-described embodiment. Accordingly, in another embodiment in which the present invention is applied to the above-described hierarchical coding, the residual signal that is blocked is quantized while suppressing an increase in quantization distortion, as in the above-described embodiment. be able to.
[0062]
The present invention can also be applied to quantization of a residual signal generated by predictive encoding other than the predictive encoding described above. The present invention can be applied not only to residual signals but also to quantization of image signals. Furthermore, the present invention can also be applied to a system having a buffering configuration for controlling the amount of generated data by controlling the quantization step width.
[0063]
【The invention's effect】
According to the present invention, when data whose existence range is narrowed by blocking is quantized, the quantization characteristic adapted to the existence range is selected, so that the quantization distortion can be reduced.
[Brief description of the drawings]
FIG. 1 is a block diagram of an example of a recording / reproducing or transmission system to which the present invention can be applied.
FIG. 2 is a block diagram of an embodiment of the present invention.
FIG. 3 is a schematic diagram used for explaining generation of a residual signal and its blocking in one embodiment of the present invention.
FIG. 4 is a schematic diagram used to describe a residual signal in one embodiment of the present invention.
FIG. 5 is a schematic diagram used for explaining quantization characteristics in one embodiment of the present invention;
FIG. 6 is a schematic diagram used for explaining another example of quantization characteristics;
FIG. 7 is a block diagram of a decoding unit in one embodiment of the present invention.
FIG. 8 is a schematic diagram used to describe an example of hierarchical encoding.
FIG. 9 is a schematic diagram used to describe an example of hierarchical encoding.
FIG. 10 is a block diagram showing an example of a configuration on the encoding side of another embodiment in which the present invention is applied to hierarchical encoding.
FIG. 11 is a block diagram showing an example of a configuration on the decoding side according to another embodiment of the present invention.
FIG. 12 is a block diagram showing an example of a class classification adaptive prediction circuit in another embodiment of the present invention.
[Explanation of symbols]
2 ... maximum value detection circuit, 3 ... minimum value detection circuit, 6 ... distributor, 7 ... select signal generation circuit, 8₁~ 8_n... Quantizer

Claims (7)

In an information signal encoding apparatus for encoding an input digital information signal so as to reduce the amount of generated data,
Means for blocking the input digital information signal;
Detecting means for detecting a minimum value and a dynamic range for each block of the input digital information signal that has been blocked;
Means for detecting an existing range of the level of the input digital information signal that has been blocked based on the minimum value and the dynamic range, and generating a selection signal corresponding to the detected existing range;
The quantization value assignment is denser than the quantization value assignment in the other quantization ranges in a predetermined quantization range, and includes the existence range among a plurality of quantization characteristics different in the predetermined quantization range , In addition, a quantization characteristic having the smallest difference between the predetermined quantization range and the existence range is selected by the selection signal, and the input digital information signal blocked by the selected quantization characteristic is quantized. Quantization means;
An information signal encoding apparatus comprising: means for transmitting the output of the quantization means and the selection signal. In an information signal encoding apparatus for encoding an input digital information signal so as to reduce the amount of generated data,
Means for generating a residual signal between sample values of the input digital information signal;
Means for blocking the residual signal;
Detecting means for detecting a minimum value and a dynamic range for each block of the blocked residual signal;
Means for detecting a presence range of a level of the blocked residual signal based on the minimum value and the dynamic range, and generating a selection signal corresponding to the detected presence range;
The quantization value assignment is denser than the quantization value assignment in the other quantization ranges in a predetermined quantization range, and includes the existence range among a plurality of quantization characteristics different in the predetermined quantization range , In addition, a quantization characteristic having the smallest difference between the predetermined quantization range and the existence range is selected by the selection signal, and a quantizer that quantizes the block residual signal with the selected quantization characteristic. And
An information signal encoding apparatus comprising: means for transmitting the output of the quantization means and the selection signal. In an information signal encoding method for encoding an input digital information signal so as to reduce the amount of generated data,
Blocking the input digital information signal;
Detecting a minimum value and a dynamic range for each block of the blocked input digital information signal;
Detecting a presence range of the level of the blocked input digital information signal based on the minimum value and the dynamic range, and generating a selection signal corresponding to the detected presence range;
The quantization value assignment is denser than the quantization value assignment in the other quantization ranges in a predetermined quantization range, and includes the existence range among a plurality of quantization characteristics different in the predetermined quantization range , In addition, a quantization characteristic having the smallest difference between the predetermined quantization range and the existence range is selected by the selection signal, and the input digital information signal blocked by the selected quantization characteristic is quantized. Steps,
An information signal encoding method comprising: transmitting an output generated by the quantization and the selection signal. In an information signal encoding method for encoding an input digital information signal so as to reduce the amount of generated data,
Generating a residual signal between sample values of the input digital information signal;
Blocking the residual signal;
Detecting a minimum value and a dynamic range for each block of the blocked residual signal;
Detecting an existing range of the level of the blocked residual signal based on the minimum value and the dynamic range, and generating a selection signal corresponding to the detected existing range;
The quantization value assignment is denser than the quantization value assignment in the other quantization ranges in a predetermined quantization range, and includes the existence range among a plurality of quantization characteristics different in the predetermined quantization range , And a quantization characteristic having the smallest difference between the predetermined quantization range and the existence range is selected by the selection signal, and the block residual signal is quantized by the selected quantization characteristic. When,
An information signal encoding method comprising: transmitting an output generated by the quantization and the selection signal. In an information signal encoding apparatus, wherein at least first and second hierarchical data are formed from an input digital information signal, and the first and second hierarchical data are encoded and transmitted.
Means for forming the second hierarchical data having a lower resolution than the first hierarchical data;
Means for predicting the first hierarchical data from the second hierarchical data;
Means for forming a residual signal between the predicted data and the first hierarchical data;
Means for blocking the residual signal;
Detecting means for detecting a minimum value and a dynamic range for each block of the blocked residual signal;
Means for detecting a presence range of a level of the blocked residual signal based on the minimum value and the dynamic range, and generating a selection signal corresponding to the detected presence range;
The quantization value assignment is denser than the quantization value assignment in the other quantization ranges in a predetermined quantization range, and includes the existence range among a plurality of quantization characteristics different in the predetermined quantization range , In addition, a quantization characteristic having the smallest difference between the predetermined quantization range and the existence range is selected by the selection signal, and a quantizer that quantizes the block residual signal with the selected quantization characteristic. And
An information signal encoding apparatus comprising: means for transmitting the output of the quantization means and the selection signal. In an information signal encoding method in which at least first and second layer data are formed from an input digital information signal, and the first and second layer data are encoded and transmitted,
Forming the second hierarchical data having a lower resolution than the first hierarchical data;
Predicting the first hierarchical data from the second hierarchical data;
Forming a residual signal between the predicted data and the first hierarchical data;
Blocking the residual signal;
Detecting a minimum value and a dynamic range for each block of the blocked residual signal;
Detecting an existing range of the level of the blocked residual signal based on the minimum value and the dynamic range, and generating a selection signal corresponding to the detected existing range;
The quantization value assignment is denser than the quantization value assignment in the other quantization ranges in a predetermined quantization range, and includes the existence range among a plurality of quantization characteristics different in the predetermined quantization range , And a quantization characteristic having the smallest difference between the predetermined quantization range and the existence range is selected by the selection signal, and the block residual signal is quantized by the selected quantization characteristic. When,
An information signal encoding method comprising: transmitting an output generated by the quantization and the selection signal. Quantization value assignment selected in response to a selection signal corresponding to the presence range of the residual signal level detected based on the minimum value of the block and the dynamic range, and other quantization ranges within a given quantization range Among the plurality of quantization characteristics that are denser than the quantized value allocation in FIG. 4 and have different predetermined quantization ranges, the difference between the predetermined quantization range and the existing range is the largest. In an information signal decoding method for an encoding method in which a blocked residual signal is quantized by a small quantization characteristic and a quantized output and the selection signal are transmitted,
Receiving the selection signal and selecting an inverse quantization characteristic;
An inverse quantization step of converting the quantized output into a representative value according to the selected inverse quantization characteristic;
A method of decoding an information signal, comprising: performing block decomposition on the residual signal converted into the representative value and converting the residual signal into an original order.

Images