JPH1188191A - Coding method, decoding method, coder and decoder - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a coding method, a decoding method, a coder and a decoder where all quantization coefficients for a prescribed period are coded with an excellent coding efficiency with a code length shorter than a set level coding length. SOLUTION: A quantization circuit 6 that quantizes a DCT coefficient of an input signal to provide an output of a quantization coefficient sets a code length in the case of coding by a fixed length and in the case that the quantization coefficient exceeds a range corresponding to the code length, the circuit limits the coefficient within a value of the range. A coding circuit 11 outputs a code in the case of applying variable length coding to coded information when there is any corresponding code in a coding table and provides an output of a code resulting from coding the coded information based on at least the quantization coefficient with the set code length and provides an output of the code in which the set code length information DL is set to a predetermined area in its header.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an encoding method, a decoding method, an encoding device and a decoding device, and more particularly to an encoding method and an encoding device for generating encoded data obtained by compressing and encoding a digital information signal. And a decoding method and a decoding device for decoding encoded data.

[0002]

2. Description of the Related Art Conventionally, an MPEG (Moving Picture Experts Group), an international standard, has been used as an encoding method for generating encoded data obtained by compressing and encoding a digital information signal.
Moving image compression methods based on p) 1 and MPEG2 are known. In the conventional encoding method based on MPEG1 or MPEG2, an image signal or a motion compensation prediction signal is subjected to discrete cosine transform (DCT: Discrete Cosine Transform). Then, in the conventional encoding method, the DCT coefficient value is quantized, and the DCT quantized coefficient value is encoded in the following procedure.

First, an input signal is divided into blocks of 8 × 8 pixels, a DCT operation is performed, and the obtained DCT coefficients are quantized for each of a DC component and an AC component, and among the quantized DCT coefficients, The DCT coefficient of the AC component is 8
Scanning is performed by zigzag scanning or the like in a DCT block of x8. The DCT coefficient (DC coefficient) of the DC component encodes a difference value using the DC coefficient of the immediately preceding block as a predicted value.

[0004] Next, in the above scanning,
The number of consecutive zero quantized coefficients (run: RUN) and the non-zero quantized coefficient values (level: LEVEL) are summarized as 2
Perform dimension variable length coding. Then, a variable length code (VLC: Variable Length Co) used in the two-dimensional variable length coding is used.
de) When there is no code corresponding to the table, a run and a level are encoded with a fixed length following a predetermined escape (ESCAPE) code.

Here, the bit width is a problem in the fixed-length level coding. In MPEG1 and MPEG2, the possible range of DCT coefficient values before quantization is [-204
8: 2047], the range of possible quantization coefficient values is also [−2048: 2047].

On the other hand, when encoding at a low bit rate, the quantization step becomes large, and as a result, the range of the quantization coefficient value becomes narrow. At this time, if the code length at the time of coding with a fixed length is long, the coding efficiency is reduced. When encoding at a high bit rate, the quantization step becomes smaller, and as a result, the range of the quantization coefficient value becomes wider. At this time, if the code length at the time of encoding the level at a fixed length is short, it is necessary to perform clip processing, and large distortion occurs.

In MPEG1, in order to solve this problem,
In the fixed-length level coding, when the level value is [-127: 127] as shown in Table 1 below, the code length is set to 8 bits, and the level value is in the range of [-128: -255] and [128: 255]. ], The code length is 16 bits, and two levels of code lengths are provided according to the level values.

[0008]

[Table 1]

[0009]

However, in the above-mentioned conventional method, as can be seen from Table 1, since the range can be expanded only by 128 in the positive and negative directions in 8-bit units, the level becomes large. The coding efficiency is extremely reduced. Further, when decoding a high-level code, it is necessary to perform a judgment process in units of 8 bits, which complicates the decoding process.

[0010] The present invention has been made in view of the above points,
An object of the present invention is to provide a coding method capable of improving coding efficiency.

Another object of the present invention is to provide an encoding method, a decoding method, an encoding device, and a decoding device capable of preventing occurrence of encoding distortion caused by limiting a level value. is there.

Still another object of the present invention is to provide an encoding method capable of encoding all quantized coefficient values in a predetermined period with a code length smaller than a set level code length with high coding efficiency and decoding the code. , A decoding method, an encoding device, and a decoding device.

[0013]

In order to achieve the above object, the encoding method according to the first aspect of the present invention provides a coding method based on a quantized coefficient value obtained by quantizing an orthogonal transform coefficient of an input signal. Generating encoded information, setting an encoding table in advance for a part of the encoded information, and performing variable length encoding of the encoded information, the encoding table is included in the encoded information. If there is a corresponding code, the code is output; otherwise, a code indicating that at least the code corresponding to the coding table is not used for the information to be coded, and a code obtained by coding the quantized coefficient value is output. In the encoding method, a code length for encoding a quantized coefficient value with a fixed length is set, and when the quantized coefficient value exceeds a range corresponding to the code length, the value is limited to a value within the range. When encoding the encoded information, the encoding data If there is no code corresponding to the code, a code indicating that at least the code corresponding to the coding table is not used for the information to be coded and a code coded with the code length set with the quantization coefficient value are output. In addition, the code length information set in a predetermined area in the header is arranged and output.

According to the present invention, the quantization coefficient value is encoded with the set code length according to the encoding bit rate or the like, and when the quantization coefficient value exceeds the range corresponding to the code length, the quantization coefficient value falls within the range. Is limited to the value of
The code length is not increased more than necessary. Further, in the present invention, since the quantization coefficient value is encoded with the set code length, the encoding can be performed without reducing the code length less than necessary.

According to a second aspect of the present invention, in the encoding method, a code length for encoding a quantized coefficient value is determined for each orthogonal transform coefficient in accordance with a quantization characteristic, and encoded information is obtained. If there is no code corresponding to the encoding table when encoding,
A code indicating that the information to be encoded is not used and a code encoded with a code length in which the quantization coefficient value is set are output. According to the present invention, the quantization coefficient value can be encoded with the set code length without arranging and transmitting the set code length information in a predetermined area in the header.

The encoding method according to the third aspect of the present invention is required when encoding a fixed-length quantization coefficient value for a period corresponding to a predetermined period of an input signal. Is determined as the maximum code length, and when encoding the encoded information, if there is no code corresponding to the encoding table, at least the code corresponding to the encoding table is used for the encoded information. A code indicating that no quantization is performed and a code obtained by encoding the quantization coefficient value with the maximum code length are output. According to the present invention, the quantization coefficient value may be encoded with a smaller code length together with a code indicating that the code corresponding to the encoding table is not used, rather than encoding the quantization coefficient value with the set code length.

Further, according to the decoding method of the present invention, encoded information is generated based on a quantized coefficient value obtained by quantizing an orthogonal transform coefficient of an input signal. Is set in advance for a part of, when the encoded information is variable-length encoded, if there is a code corresponding to the encoding table in the encoded information, the code is output, Otherwise, a code indicating that at least a code corresponding to the coding table is not used for the information to be coded, and variable-length coded data output by a coding method of outputting a code obtained by coding a quantized coefficient value. In a decoding method for decoding an input signal, a code length for decoding a quantized coefficient value is determined according to a quantization characteristic for each orthogonal transform coefficient, and a code length is determined for decoding encoded information. No code corresponding to the conversion table When decoding the code indicating that is obtained so as to decode code length in decoding the least quantization coefficient values. According to the present invention, decoding can be performed according to the quantization characteristics for each orthogonal transform coefficient.

According to the present invention, there is provided a quantization means for quantizing orthogonal transform coefficients of an input signal to output a quantization coefficient value, and generating encoded information based on the quantization coefficient value output from the quantization means. Then, an encoding table is set in advance for a part of the encoded information, and when the encoded information is subjected to variable length encoding, a code corresponding to the encoding table is included in the encoded information. If the encoding circuit outputs the code, and if not, a code indicating that at least the code corresponding to the encoding table is not used for the information to be encoded, and an encoding circuit that outputs a code obtained by encoding the quantization coefficient value, The encoding apparatus according to claim 5, further comprising: a code length determining circuit for determining a code length when encoding the quantized coefficient value according to the quantization characteristic for each orthogonal transform coefficient. The encoding circuit matches the encoding table. If there is no code to be encoded, a code indicating that the information to be encoded is not used, and a code coded with the code length set with the quantization coefficient value from the code length determination circuit are output, The encoding apparatus according to claim 6, further comprising: a code length determining circuit for determining a code length when encoding the quantized coefficient value in accordance with the quantization characteristic for each orthogonal transform coefficient. If there is no code corresponding to the table, a code indicating that the information to be coded is not used and a code coded by the code length set with the quantization coefficient value from the code length determination circuit are output. It is.

The encoding apparatus according to claim 7 of the present invention further comprises a memory for accumulating the quantized coefficient value output from the quantizing means for a period corresponding to a predetermined period of the input signal;
A maximum code length detection circuit for obtaining a code length necessary for encoding the quantized coefficient value output from the quantization means with a fixed length as a maximum code length. When encoding the encoded information generated based on the encoding coefficient value, if there is no code corresponding to the encoding table, it is necessary not to use at least the code corresponding to the encoding table for the encoded information. In order to output the code shown and the code obtained by encoding the quantized coefficient value with the maximum code length detected by the maximum code length detection circuit, the encoding is performed rather than encoding the quantized coefficient value with the set code length. In some cases, the quantization coefficient value can be encoded with a small code length together with a code indicating that the code corresponding to the table is not used.

The decoding apparatus according to the present invention further comprises a receiving code length determining circuit for obtaining a code length for decoding a quantized coefficient value in accordance with a quantization characteristic for each orthogonal transform coefficient. When a code indicating that there is no code corresponding to at least the coding table is input, a decoding circuit that decodes the code obtained by coding the quantized coefficient value with the code length from the reception code length determination circuit. It is characterized by having.

[0021]

Next, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram illustrating an encoding process according to a first embodiment of the encoding method and apparatus according to the present invention. In FIG. 1, a digital image signal (video signal) to be transmitted is input to an input terminal 1, and level code length information DL is input to an input terminal 2 in units of one frame or a plurality of frames. This level code length information D
For example, L is transmitted by being arranged in a transmission area additionally set in a picture header of the MPEG1 standard, for example.

Next, the operation of this embodiment will be described. The video signal input from the input terminal 1 is supplied to a subtractor 3 and a prediction circuit 4, respectively, and is subtracted from an output motion compensation signal of a prediction circuit 4 to be described later by the subtracter 3, thereby obtaining a motion compensation prediction error signal and After that, it is supplied to the DCT circuit 5, where it is divided into blocks of 8 × 8 pixels and D
A CT operation is performed to obtain a DCT coefficient as an example of an orthogonal transform coefficient. This DCT coefficient has 12 bits and a value range of [−2048: 2047].

The DCT coefficient output from the DCT circuit 5 (referred to as Ci) is supplied to the quantization circuit 6, where it is input from the quantization control circuit 13 to the quantization circuit 6.
In accordance with the quantization control information QUANT for controlling the generated code amount according to the amount of the buffer memory 12 and the quantization matrix Wi, the quantization value QCi is calculated by the following equation.

QCi = (Wi × Ci) ÷ (2 × 16 × QUANT) (1) Therefore, for example, a certain DCT coefficient value Ci is “1080”
Then, the quantization matrix value Wi for this DCT coefficient value
Is "16" and the quantization control information QUANT is "2", the above-described DCT coefficient value Ci is converted to the quantized value QCi "270" by the operation of the equation (1) according to the MPEG1 standard.

Further, the quantization circuit 6 of this embodiment
A range of the quantization value determined according to the level code length information DL input from the input terminal 2 is set, and if a quantization value exceeding this setting range is obtained by the equation (1), the range is set within this range. To the value of. Therefore, for example, when the level code length information DL represents a level code length of 9 bits, the range of the quantization value is limited to [-256: 255]. That is, in this case, the quantization circuit 6 limits the quantization value to the upper limit value "255" even when the quantization value QCi calculated by the above equation (1) is "270".

The DCT quantized coefficient (quantized value QCi) for each block of 8 × 8 pixels obtained by the quantizing circuit 6 is supplied to an encoding circuit 11 described later, while the inverse After being supplied to the conversion circuit 7 and subjected to inverse quantization, it is supplied to the inverse DCT circuit 8 and subjected to inverse DCT operation. The output signal of the inverse DCT circuit 8 is added to the motion compensation signal from the prediction circuit 4 in the adder 9 to be a locally decoded signal, restored to the same image data as the decoder, and stored in the frame memory 10. Is done.

The prediction circuit 4 supplies the image data (video signal) stored in the frame memory 10 as reference image data for the next motion compensated prediction, while receiving a predetermined unframed frame from the input terminal 1. An input video signal is supplied, and a prediction signal is generated by a known method based on these signals, and a motion vector indicating a position where a difference between the input video signal and the prediction signal in a certain pixel region is the smallest is calculated. The detected and compensated motion compensation signal is output to the subtractor 3 and the adder 9 in accordance with the motion vector.

The encoding circuit 11 encodes the DCT quantization coefficient (quantized value QCi) input from the quantization circuit 6 and the prediction control information such as the motion vector input from the prediction circuit 4 into encoded information. On the other hand, when performing variable-length coding, a two-dimensional variable-length coding table corresponding to the non-zero coefficient (level) and the number of consecutive zero coefficients (runs) before it is set in advance, and the encoded information is stored in MPEG1 If the code sequentially obtained by scanning with the prescribed zigzag scan is in the two-dimensional variable length coding table, the code is output. Otherwise, the code to be encoded corresponds to at least the two-dimensional variable length coding table. The run value is output with a fixed length of 6 bits following the escape code indicating that the code to be used is not used, and the level value is output with the level code length corresponding to the level code length information DL.

The encoding circuit 11 outputs to the buffer memory 12 encoded data obtained by adding a picture header including the level code length information DL to the code obtained by performing the two-dimensional variable length encoding as described above. The buffer memory 12 receives the encoded data output from the encoding circuit 11 at an irregular rate, accumulates the data, and outputs it to the output terminal 14 at a constant rate.
The quantization control circuit 13 generates a quantization control signal QU for controlling the generated code amount in accordance with the amount of the buffer memory 12 being full.
The ANT and the quantization matrix Wi are output.

As described above, in this embodiment, the level code length information DL indicating the level length of the coded image data (video signal) is arranged in the header, and the coded image data (video signal) ) Is output to the output terminal 14 as encoded data. Next, a method of decoding the encoded data output by the encoding device of FIG. 1 will be described. FIG. 2 is a block diagram illustrating a decoding process according to the first embodiment of the decoding method and apparatus according to the present invention. In the figure, encoded data input to an input terminal 21 is temporarily stored in a buffer memory 22 and then supplied to a decoding circuit 23 to be decoded. Of the decoded information, the level code length information DL in the header is supplied to and held in the level code length information memory 24.

In the decoding of the two-dimensional variable length code with respect to the DCT quantized coefficient, when the decoding circuit 23 receives the escape code, it receives the run value with a fixed length of 6 bits and the level value with the level code length. Decoding is performed with a level code length corresponding to the level code length information DL read from the information memory 24.

The circuit sections subsequent to the decoding circuit 23 perform the same decoding processing as the conventionally known decoding processing of the MPEG1 standard. That is, it is inversely quantized based on the quantization control information and the quantization matrix received by the inverse quantization circuit 25, and further subjected to inverse DCT operation by the inverse DCT circuit 26 and decoded. Further, the prediction control information decoded by the decoding circuit 23 is supplied to the prediction circuit 28, where it is made a motion compensation prediction signal based on the code from the frame memory 29.

The adder 27 adds the motion compensation prediction signal from the prediction circuit 28 to the decoded code output from the inverse DCT circuit 26 to generate an original video signal.
On the other hand, the data is output to 0 and stored in the frame memory 29 for a predetermined frame, and then supplied to the prediction circuit 28. As described above, in this embodiment, the level code length can be decoded using the level code length information DL that is arranged and transmitted in the header.

Next, a description will be given of a second embodiment of the encoding method, the decoding method, the encoding device and the decoding device according to the present invention. FIG. 3 is a block diagram illustrating an encoding process according to a second embodiment of the encoding method and apparatus according to the present invention. In the figure, the same components as those of FIG. 1 are denoted by the same reference numerals, and description thereof will be omitted.

In FIG. 3, the DCT circuit 5 divides the input prediction error signal into 8 × 8 pixel blocks and performs a DCT operation to obtain a 12-bit value range [−2048: 204].
7] and generates a DCT coefficient, which is
To supply. The quantization circuit 31 converts the quantization control signal QUANT and the quantization matrix Wi, which are input from the quantization control circuit 13 to the quantization circuit 31, for controlling the generated code amount according to the filling amount of the buffer memory 12. Accordingly, the quantization value Q is obtained from the input DCT coefficient value Ci by the above equation (1).
Ci is calculated, and the calculated quantized value QCi is output as it is.

The encoding circuit 32 controls the DCT quantization coefficient (quantized value QCi) for each block of 8 × 8 pixels input from the quantization circuit 31 and predictive control such as the motion vector input from the prediction circuit 4. The information and the zigzag scan specified by MPEG1 are scanned to obtain a non-zero coefficient (level)
And two-dimensional variable-length coding based on the number of consecutive zero coefficients (runs) before it. If there is a code corresponding to the two-dimensional variable length coding table used in the two-dimensional variable length coding among the codes input to the coding circuit 32, the code is output as it is and the two-dimensional variable length code is output. If the code corresponding to the conversion table does not exist, a run value is output in a fixed length of 6 bits following the escape code, and the level value is determined by the level code length determination circuit 33 as the level code. Output as long.

The output value of the DCT circuit 5 is 12 bits [-2
048: 2047], and the range of the quantization value is uniquely obtained from the quantization operation expression and the quantization matrix value and the quantization control information input from the quantization control circuit 13. For example, when the above-mentioned quantization value QCi is “270”, the range of the quantization value is calculated to be [−512: 511] by the level code length determination circuit 33 so as to include this.

The encoding circuit 32 quantizes the level code length based on the quantization value range [-512: 511] indicated by the level code length information input from the level code length determination circuit 33. Is set to 10 bits, which is a code length necessary for encoding with a fixed length. The encoded data output from the encoding circuit 32 is output to the output terminal 34 through the buffer memory 12.

If the accuracy of the input value of the DCT transform is, for example, 9 bits [-256: 255], the DCT
It is sufficient that the output value of the conversion can be represented by a smaller number of bits, such as 11 bits, except for the DC coefficient. For this reason, the output value of the DCT transform for all the coefficients is not set to 12 bits [−2048: 2047], but is regarded as a smaller number of bits such as 11 bits for each coefficient, so that the range of the quantization value can be further narrowed. . Thereby, the level code length can be further reduced.

However, in this case, similarly to the first embodiment, the process of limiting the coefficients exceeding the above-described range of the quantized values for each quantized coefficient to the range of the quantized values obtained above. I do.

Next, a method of decoding the encoded data output by the encoding apparatus shown in FIG. 3 will be described.
FIG. 4 is a block diagram for explaining a decoding process according to the second embodiment of the decoding method and apparatus according to the present invention. In the figure, the same components as those in FIG.
The description is omitted. 4, the coded data input to the input terminal 41 is coded data obtained by the coding apparatus of FIG. 3, which does not have the level code length information DL in the header.

The encoded data is stored in the buffer memory 22
Is input to the decoding circuit 42 via Decoding circuit 42
Performs decoding of a two-dimensional variable-length code on DCT quantized coefficients. The level code length determination circuit 43 includes the decoding circuit 4
The matrix used in the quantization processing of each block and the quantization control information are obtained from the header information in the encoded data decoded in step 2, and the level code length information is calculated.

In the decoding of the two-dimensional variable length code for the DCT quantized coefficient, the decoding circuit 42 decodes the run value with a fixed length of 6 bits when receiving the escape code.
The level value is decoded with a level code length corresponding to the level code length information input from the level code length determination circuit 43.

In this embodiment, the same format-encoded data as in the prior art is transmitted and decoded without providing a dedicated area for transmitting the level code length information DL in the header as in the first embodiment. can do.

Next, a description will be given of a third embodiment of the encoding method, the decoding method, the encoding device and the decoding device according to the present invention. FIG. 5 is a block diagram illustrating an encoding process according to a third embodiment of the encoding method and apparatus according to the present invention. In the figure, the same components as those of FIG. 1 are denoted by the same reference numerals, and description thereof will be omitted.

In the embodiment shown in FIG. 5, the quantized value QCi output from the quantizing circuit 6 is supplied to a quantized coefficient memory 51 and a maximum code length detecting circuit 52, respectively, and their outputs are converted to a quantized control circuit. 13 together with the quantization control information and quantization matrix value from the prediction circuit 4 and the prediction control information from the prediction circuit 4.

Next, the operation of this embodiment will be described. The quantization circuit 6 is input from the quantization control circuit 13 with the 12-bit DCT coefficient extracted from the DCT circuit 5 and having a value range of [−2048: 2047].
According to the quantization control signal QUANT and the quantization matrix Wi for controlling the generated code amount in accordance with the filling amount of the buffer memory 12, the quantization value QCi is obtained by the above equation (1).
Is calculated, and the calculated quantized value QCi is input to the input terminal 2
And outputs a quantized value QCi whose value is limited by the level code length information DL input from the outside via the.

For example, if a certain DCT coefficient value Ci is "108
Assuming that the quantization matrix value Wi for the DCT coefficient value is “16” and the quantization control information QUANT is “5” for the DCT coefficient value, the DCT coefficient value Ci is calculated by the operation of the expression (1) according to the MPEG1 standard. Is the quantization value QCi “10
8 ".

Further, the quantization circuit 6 of this embodiment
As in the first embodiment, the level code length information DL1
If a quantized value exceeding the range of the quantized value determined according to (= DL) is obtained by the equation (1), the output is limited to a value within this set range. When the information DL1 represents a level code length of 9 bits, the range of the quantization value is limited to [-256: 255]. That is, in this case, the quantization circuit 6 determines that the quantization value QCi calculated by the above equation (1) is “108”.
“108” is output as the quantized value QCi.

As described above, the DC of each 8 × 8 pixel block for one frame obtained by the quantization circuit 6 is obtained.
The T quantization coefficient (quantized value QCi) is supplied to the inverse quantization circuit 7, and is also sequentially supplied to the quantization coefficient memory 51 and stored, and is also supplied to the maximum code length detection circuit 52.

The maximum code length detection circuit 52 detects the maximum code length DL2 of each quantized coefficient in one frame.
This is realized by obtaining the maximum value of the absolute value of each quantization coefficient in one frame and outputting a code length for encoding the maximum value. For example, the maximum value of the quantization coefficient (quantization value QCi) in one frame is “10
If it is "8", the maximum code length detection circuit 52 outputs, as the maximum code length DL2, a code length of 8 bits that can encode the range of [-128: 127] including this "108".

The DCT quantization coefficient (quantized value QCi) for each block of 8.times.8 pixels for one frame stored in the quantization coefficient memory 51 is supplied to the encoding circuit 53, and M
Scanning is performed by a zigzag scan specified by PEG1, and two-dimensional variable length coding is performed using a non-zero coefficient (level) and the number of consecutive zero coefficients (runs) before the coefficient.

That is, if the code corresponding to the run value or the level value is present in the two-dimensional variable length coding table used for the two-dimensional variable length coding, the coding circuit 53 converts the code into a code. If the code corresponding to the two-dimensional variable length coding table does not exist, an escape code is output followed by a run value with a fixed length of 6 bits, and the level value is detected by the maximum code length detection circuit 52. Is output with a level code length corresponding to the maximum code length DL2. The encoding circuit 53 outputs, to the buffer memory 12, encoded data obtained by adding a picture header including the level code length information DL1 to the two-dimensionally variable-length encoded code as described above.

According to this embodiment, for example, the level code length DL 1
When encoding can be performed with a smaller code length, the encoding efficiency can be further improved as compared with the first embodiment.

[0055]

Next, an embodiment of the present invention will be described. Now, as an example, it is assumed that DCT quantization coefficients obtained by quantization by the quantization circuits 6 and 31 are shown in Table 2.
However, the quantization control information QUANT is "4".

[0056]

[Table 2] In the conventional MPEG1 encoding of such DCT quantized coefficients, the encoding circuit sets the DCT quantized coefficients in Table 2 to “135”, “20”, “18”, “−64”, “−”.
3 "," -31 "," -30 "," 4 "," 10 ",
"-23", "12", "3",. . . , And a run value (RUN) and a level value (LEVEL) as shown in Table 3 are obtained, and a variable length code (VLC) is obtained.
And the number of bits.

[0057]

[Table 3] On the other hand, for the DCT quantization coefficients in Table 2, FIG.
In the encoding performed by the encoding circuit 11, when the level code length information DL is 11 bits, a variable length code (VLC) and the number of bits as shown in Table 4 can be obtained. Note that zigzag scanning is the same as in the related art.

[0058]

[Table 4] In addition, with respect to the DCT quantization coefficients in Table 2, the quantization control information QUANT is used in the encoding by the encoding circuit 32 in FIG.
Is "1", the level code length information is "11", when the quantization control information QUANT is "2-3", the level code length information is "10", and when the quantization control information QUANT is "4-7". The level code length information is “9”, and the quantization control information QUA is
When NT is “8 to 15”, the level code length information is “8”,
Assuming that the level code length information is “7” when the quantization control information QUANT is “16 to 31”, the level code length information is “9” here because the quantization control information QUANT is “4”. , Variable-length codes (VL
C) and the number of bits are obtained. Note that zigzag scanning is the same as in the related art.

[0059]

[Table 5] As can be seen from Tables 2 to 5, it can be seen that the present invention can improve the coding efficiency, and in particular, the embodiment of FIG. 3 shown in Table 5 can improve the coding efficiency most.

The present invention is not limited to the above embodiment. For example, the compression encoding and decoding methods are M
It is not limited to PEG1 and MPEG2. In addition, the present invention can be applied to other information signals such as audio signals other than video signals as digital information to be encoded.

[0061]

As described above, according to the present invention,
By encoding the quantized coefficient value according to the encoding bit rate and the like with the set code length, and restricting the quantized coefficient value to a value within the range when the quantized coefficient value exceeds a range corresponding to the code length. Since the code length is not increased unnecessarily, the coding efficiency can be improved.Also, by encoding the quantization coefficient value with the set code length, the code can be encoded without reducing the code length unnecessarily. Therefore, it is possible to prevent the occurrence of coding distortion caused by limiting the level value, and to simplify the coding and decoding processes as compared with the MPEG1 system.

Further, according to the present invention, it is possible to encode and decode the quantized coefficient value with the set code length without arranging and transmitting the set code length information in a predetermined area in the header. In addition to the above effects, the coding efficiency can be further improved.

Further, according to the present invention, for a quantization coefficient value for a period corresponding to a predetermined period of an input signal, the code length necessary for encoding the quantization coefficient value with a fixed length is the maximum code length. When encoding the information to be encoded,
If there is no code corresponding to the encoding table, a code indicating that at least the code corresponding to the encoding table is not used for the information to be encoded, and a code obtained by encoding the quantization coefficient value with the maximum code length. Since the output is performed, the coding efficiency can be more flexibly and effectively improved than when the quantization coefficient value is encoded with the set code length.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating an encoding process according to a first embodiment of an encoding method according to the present invention.

FIG. 2 is a block diagram for explaining a decoding process according to the first embodiment of the decoding method according to the present invention.

FIG. 3 is a block diagram illustrating an encoding process according to a second embodiment of the encoding method according to the present invention.

FIG. 4 is a block diagram for explaining a decoding process according to a second embodiment of the decoding method according to the present invention.

FIG. 5 is a block diagram illustrating an encoding process according to a third embodiment of the encoding method according to the present invention.

[Explanation of symbols]

 Reference Signs List 1 digital image signal (video signal) input terminal 2 level code length information input terminal 4 prediction circuit 5 discrete cosine transform (DCT) circuit 6, 31 quantization circuit 11, 32, 53 coding circuit 13 quantization control circuit 14, 34 Encoded data output terminal 21, 41 Encoded data input terminal 23, 42 Decoding circuit 24 Level code length information memory 30 Video signal output terminal 33, 43 Level code length determination circuit 51 Quantization coefficient memory 52 Maximum code length detection circuit

Claims (8)

[Claims] An encoded information is generated based on a quantized coefficient value obtained by quantizing an orthogonal transform coefficient of an input signal, and an encoding table is set in advance for a part of the encoded information. In the variable-length encoding of the encoded information, if there is a code corresponding to the encoding table in the encoded information, the code is output, otherwise, at least for the encoded information In a coding method for outputting a code indicating that a code corresponding to the coding table is not used and a code obtained by coding the quantization coefficient value, a code used when coding the quantization coefficient value with a fixed length. When the quantization coefficient value exceeds a range corresponding to the code length, the quantization coefficient value is limited to a value within the range, and when encoding the coded information, the encoding table includes If there is no corresponding code, the code A code indicating that at least a code corresponding to the coding table is not used for information and a code obtained by coding the quantization coefficient value with a set code length are output, and a predetermined area in the header is output. Wherein the set code length information is arranged and output. 2. An encoded information is generated based on a quantized coefficient value obtained by quantizing an orthogonal transform coefficient of an input signal, and an encoding table is set in advance for a part of the encoded information. In the variable-length encoding of the encoded information, if there is a code corresponding to the encoding table in the encoded information, the code is output, otherwise, at least for the encoded information In a coding method for outputting a code indicating that a code corresponding to the coding table is not used, and a code obtained by coding the quantization coefficient value, the quantization is performed in accordance with a quantization characteristic for each of the orthogonal transform coefficients. Determine the code length when encoding the encoding coefficient value, if there is no code corresponding to the encoding table when encoding the encoded information, a code indicating that the encoded information is not used. , The quantization coefficient value Encoding method and outputting an encoded code in boss was code length. 3. Generating encoded information based on a quantized coefficient value obtained by quantizing an orthogonal transform coefficient of an input signal, and setting an encoding table in advance for a part of the encoded information. In the variable-length encoding of the encoded information, if there is a code corresponding to the encoding table in the encoded information, the code is output, otherwise, at least for the encoded information In a coding method for outputting a code indicating that a code corresponding to the coding table is not used and a code obtained by coding the quantization coefficient value, the quantization section for a period corresponding to a predetermined period of an input signal is provided. For a numerical value, a code length necessary for encoding the quantized coefficient value with a fixed length is determined as a maximum code length, and when encoding the encoded information, a code corresponding to the encoding table is encoded. If not, the said mark A code indicating that at least a code corresponding to the coding table is not used for the quantization information, and a code obtained by coding the quantization coefficient value with the maximum code length. . 4. An encoded information is generated based on a quantized coefficient value obtained by quantizing an orthogonal transform coefficient of an input signal, and an encoding table is set in advance for a part of the encoded information. In the variable-length encoding of the encoded information, if there is a code corresponding to the encoding table in the encoded information, the code is output, otherwise, at least for the encoded information A code indicating that a code corresponding to the coding table is not used, and a variable length coded data output by a coding method of outputting a code obtained by coding the quantization coefficient value, and decoding the input signal. In the decoding method, according to a quantization characteristic for each of the orthogonal transform coefficients, a code length when decoding a quantized coefficient value is obtained, and when the encoded information is decoded, the code length corresponds to the encoding table. There is no sign to do When decoding the code indicating the decoding method characterized by decoding code length in decoding the at least the quantized coefficient values. 5. Quantizing means for quantizing an orthogonal transform coefficient of an input signal to output a quantized coefficient value, and generates encoded information based on the quantized coefficient value output from said quantizing means. An encoding table is set in advance for a part of the encoded information, and when the encoded information is subjected to variable length encoding, if there is a code corresponding to the encoding table in the encoded information, An encoding circuit that outputs the code and outputs a code indicating that at least a code corresponding to the encoding table is not used for the information to be encoded if there is no code, and a code obtained by encoding the quantized coefficient value; Wherein the quantization means sets a code length for encoding the quantized coefficient value with a fixed length, and when the quantized coefficient value exceeds a range corresponding to the code length. Limited to values within that range, The encoding circuit, if there is no code corresponding to the encoding table, a code indicating that at least the code corresponding to the encoding table is not used for the encoded information, the quantization coefficient value An encoding device, which outputs a code encoded with a set code length, and outputs the set code length information in a predetermined area in a header. 6. Quantizing means for quantizing orthogonal transform coefficients of an input signal to output quantized coefficient values, and coded information is generated based on the quantized coefficient values output from the quantizing means. An encoding table is set in advance for a part of the encoded information, and when the encoded information is subjected to variable length encoding, if there is a code corresponding to the encoding table in the encoded information, An encoding circuit that outputs the code and outputs a code indicating that at least a code corresponding to the encoding table is not used for the information to be encoded if there is no code, and a code obtained by encoding the quantized coefficient value; A coding length determining circuit for obtaining a code length when coding the quantized coefficient value in accordance with a quantization characteristic for each of the orthogonal transform coefficients, wherein the coding circuit comprises: Compatible with encoding table If no code is present, a code indicating that the information to be coded is not used, and a code coded by the code length set from the code length determining circuit to the quantized coefficient value are output. Encoding device. 7. Quantizing means for quantizing orthogonal transform coefficients of an input signal to output quantized coefficient values, and coded information is generated based on the quantized coefficient values output from said quantizing means. An encoding table is set in advance for a part of the encoded information, and when the encoded information is subjected to variable length encoding, if there is a code corresponding to the encoding table in the encoded information, An encoding circuit that outputs the code and outputs a code indicating that at least a code corresponding to the encoding table is not used for the information to be encoded if there is no code, and a code obtained by encoding the quantized coefficient value; A memory that stores the quantized coefficient value output from the quantizing unit for a period corresponding to a predetermined period of the input signal; and the quantizer output from the quantizing unit. Conversion factor The a maximum code length detection circuit for obtaining the maximum code length of code length required to encode provided a fixed length, said encoding circuit,
When encoding the encoded information generated based on the quantized coefficient value input from the memory, if there is no code corresponding to the encoding table, at least the encoded information A code indicating that a code corresponding to the encoding table is not used, and a code obtained by encoding the quantized coefficient value with a maximum code length detected by the maximum code length detection circuit. Device. 8. Encoding information is generated based on quantized coefficient values obtained by quantizing orthogonal transform coefficients of an input signal, and an encoding table is preset for a part of the encoded information. In the variable-length encoding of the encoded information, if there is a code corresponding to the encoding table in the encoded information, the code is output, otherwise, at least for the encoded information A code indicating that a code corresponding to the coding table is not used, and a variable length coded data output by a coding method of outputting a code obtained by coding the quantization coefficient value, and decoding the input signal. A receiving code length determining circuit for determining a code length when decoding a quantized coefficient value according to a quantization characteristic for each of the orthogonal transform coefficients, and a code corresponding to at least the coding table. No And a decoding circuit for decoding a code obtained by coding the quantization coefficient value with a code length from the reception code length determination circuit when a code indicating the following is input: .