JP2650538B2 - Encoding method, encoding device, and decoding device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an encoding method, an encoding apparatus, and a decoding apparatus used for compressing, transmitting, and recording digital signals such as digital images and digital audio.

[0002]

2. Description of the Related Art In recent years, as a method of encoding a digital image signal, a method of dividing a digital image signal into blocks composed of a plurality of pixels, performing discrete cosine transform for each block, and quantizing and encoding discrete cosine transform coefficients. Is the mainstream. When the discrete cosine transform coefficients in each block are quantized, coefficients that do not become 0 are significant coefficients,
Are classified as insignificant coefficients, each is encoded by a variable length code, and when the insignificant coefficient continues to the end of the block, a method of adding a terminal code and not encoding the insignificant coefficient is generally used. ing. In decoding, highly efficient encoding is realized by recognizing the end of encoded data of a block by the end code and reducing the amount of information required for encoding the insignificant coefficient.

[0003] Also, a well-known digital audio coding method is to divide a block into a plurality of sample points and use a variable length code for each block.

[0004]

However, the method of recognizing the end of the coded data of the block by the terminal code requires a large amount of image quality if an error occurs in the decoding of the terminal code due to external noise or the like in the decoding. Have a significant effect. That is, if the termination code disappears or occurs due to an error, significant blocks disappear or unintended blocks occur. As a result, the position of the block is displaced, and the effect of the error continues even after the error occurs, resulting in remarkable image quality deterioration.

[0005] Also, in a method of recognizing the end of a block by encoding all data included in a block without using a termination code and decoding the encoded data, encoding is performed using a variable length code. In this case, when an error occurs, a significant code disappears or an insignificant code causes a code unit position shift.As a result, even in the block after the error occurs, the position shift of the encoded data occurs, and the effect of the error is reduced. Propagate beyond the block. This is because the synchronization in the decoding of the variable length code is shifted. However, in the DCT coding method, the position shift of the encoded data becomes the frequency shift, and the influence of the error on the image quality becomes great.

In view of the foregoing, the present invention provides a code that realizes highly efficient encoding without reducing the amount of information required for error correction without causing the effects of errors due to noise or the like to propagate to blocks subsequent to the block in which the error occurred. It is an object of the present invention to provide an encoding method, an encoding device and a decoding device.

[0007]

SUMMARY OF THE INVENTION An encoding method according to the present invention comprises:
A method for encoding a digital signal sequence, comprising: dividing the digital signal sequence into a block signal sequence including a plurality of signals; encoding the block signal sequence to generate a block code.
A sequence is obtained, and the total code length of the block code sequence is measured as a block code length.
The end of the block code string is configured to be recognized, and the end added to the end of the block code string.
Code or the total code length of the block code string
Thus, the end of the block code string is recognized.

[0008] The encoding device of the present invention, a digital signal string to a device for processing coded, the block divider for dividing the digital signal sequence to the block signal sequence comprising a plurality of signal, the block Encodes a signal sequence and blocks
And encoder for determining the click code sequence that includes the block code length measuring device for measuring the total code length of the block code sequence, the block code length encoder for encoding the output of the block code length measuring device , and the said further has a second predetermined amount number of the block signal sequence encodes the block code length code sequence which is the output of the block code length coder after the first transmitted or recorded a predetermined amount pieces together Block code
The first predetermined amount of the block code length code sequences are collectively transmitted or recorded after the sequence is transmitted or recorded, or the second predetermined amount of the block code sequences are collectively transmitted or recorded. And a format converter for rearranging the code sequence in such a manner as to perform the above operation.

A decoding device according to the present invention is a device for decoding a code sequence obtained by the coding device, comprising: a block code length decoder for decoding the block code length code sequence; , An end recognizer that recognizes the end of the block code string based on the output of the block code length decoder, and the end recognizer further comprises:
The terminal code added to the end of the block code string
First means for recognizing the end of a block code sequence, wherein the block code a second means for recognizing the end of the block code sequence by length, of either of said first means and said second means And means for selecting the output of the means.

[0010]

As a result, even if an error occurs in decoding due to noise or the like, the end of the code string of the block can be recognized more accurately, and decoding synchronization is restored at the beginning of each block. Does not propagate beyond the block, and the position of the block does not shift, so that it is possible to reduce image quality deterioration due to errors. Further, since the data can be transmitted or recorded with an increased error correction capability only for the block code length data sequence, the amount of information required for error correction can be reduced.

[0011]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing an embodiment of an encoding method, an encoding apparatus and a decoding apparatus according to the present invention. Here, a case where the present invention is applied to a method of using a digital image signal as a digital signal to be encoded, dividing the digital image signal into blocks, performing discrete cosine transform on each block, and encoding the amplitude of a discrete cosine transform coefficient Will be described. In addition, a one-dimensional block is assumed for simplicity of description, and the number of pixels included in one block is eight. It is assumed that the number of blocks to be encoded is N.

First, a first embodiment of an encoding device and a decoding device will be described with reference to FIG. FIG. 1 shows a block divider 3, an encoder 4, a block code length measuring device 5,
Encoding device 1 composed of block code length encoder 6
And a block code length decoder 7, a block end recognizer 8,
2 shows a decoding device 2 including a decoder 9.

First, the encoding device 1 will be described. The input signal sequence is divided by the block divider 3 into image signals into blocks each composed of eight pixels. The encoder 4 performs a discrete cosine transform on each block signal sequence, and quantizes the obtained amplitude signal sequence of transform coefficients. The following signal sequence is considered as an example of a quantized amplitude signal sequence.

{.., 0,0, | 2,0,1,0,0,0,0,0 | 3,0,1,0,0,0,
0,0 | 2,1,0,0,0,0,0,0 | 1,1,...} Where “|” indicates a block segment and is an actual signal. Does not exist in the column. If the image is divided into blocks and subjected to discrete cosine transform and quantized, the coefficient at the front of the block is other than 0 as in the above-mentioned amplitude signal sequence, but generally becomes 0 as approaching the end. is there. Therefore, the encoder 1 replaces the sequence of 0s continuing to the end of the block in the amplitude signal sequence of the block with an end symbol (EOB: End Of Block) indicating the end of the block. In the example of the amplitude signal sequence, the following is performed.

{.., EOB, | 2,0,1, EOB, | 3,0,1, EOB, | 2,1, EOB | 1,1, ..} Next, the obtained amplitude signal sequence Is encoded by a code table and output. (Table 1) shows an example of the code table.

[0016]

[Table 1]

When an example of the amplitude signal sequence is encoded using the code table shown in Table 1, the output of the encoder 4 is as follows.

{.. | 1100110 | 1110110 | 11010 | 0101 ..} However, the end code (EOB) is omitted. This is because at the time of decoding, if the code length of the amplitude code string is known, the end of the block can be recognized, so that decoding can be performed without sending the end code to the decoding device 2. is there.

Next, a block amplitude code sequence obtained by encoding the amplitude signal sequence for each block is input to a block code length measuring device 5. The block code length measuring device 5 measures and outputs the total code length of each input block amplitude code sequence. The output of the block code length measuring device 5 in the case of the above example is as follows.

{.., 7, 7, 5,...} Next, the block code length is input to a block code length encoder 3 and coded according to a code table. Is output. The following is the result of encoding the block code length using an example of the code table shown in (Table 2).

{.. | 111 | 111 | 101 | ..}

[0022]

[Table 2]

By separating the block code length code string and the amplitude code string in this manner, it is possible to transmit or record the data without affecting the block code length code string due to an error generated in the amplitude code string due to noise or the like. Can be. further,
Since the block code length code string and the amplitude code string are separated, the error correction capability can be improved only for the block code length code string, so that the amount of information required for error correction can be reduced.

Next, the decoding device 2 will be described. First, a block code length code string output from the encoding device 1 is decoded by a block code length decoder 7 according to a predetermined code table. The decoded block code length and the amplitude code string output from the encoding device 1 are both the block end recognizer 8
To detect the end of the block amplitude code sequence. The method of detecting the end of the block amplitude code string can be realized by counting the input amplitude code string one bit at a time and comparing the counted result with the block code length.

Next, when the end of the block of the amplitude code string is recognized, the amplitude code string is decoded in the decoder 9. The input code sequence of the decoder 9 is as follows in an example of the input signal sequence.

{.. | 1100110 | 1110110 | 11010 | 0101 ..} where "|" is a block delimiter, which is recognized by the block end recognizer 8, and the terminal code in the decoder 9 (EOB) is added. Using the example of the code table of (Table 1), decoding is performed as follows.

{.., EOB, | 2,0,1, EOB, | 3,0,1, EOB, | 2,1, EOB | 1,1, ..} Next, the terminal code (EOB) is 0 continues to the end of the block
, 0,0, | 2,0,1,0,0,0,0,0 | 3,0,1,0,0,0,0,0 | 2,1 , 0,0,
0,0,0,0 | 1,1, ...}. In this way, by recognizing the end of the amplitude code string of a block based on the block code length and decoding the code, even if an error occurs in the amplitude code string, the code can be synchronized at the beginning of each block, and the code can be crossed over. The effects of the error do not propagate.

Next, a second embodiment of the encoding apparatus and the decoding apparatus will be described with reference to FIG. The encoding device 10 shown in FIG. 2 includes the same block divider 3, encoder 4, block code length measuring device 5, block code length encoder 6, and format converter 12 as in the first embodiment. ing. Further, the decoding device 11 includes a deformat converter 13, a block code length decoder 7, a floc end recognizer 8, and a decoder 9 which are the same as those in the first embodiment.

In the second embodiment, the method of encoding the amplitude signal sequence for each block and encoding the block code length are the same as in the first embodiment. The obtained block code length code string and block amplitude code string are both input to the format converter 12 to form an encoded data string. After outputting the block code length code sequence, the format converter 12 rearranges the data sequence so as to continuously output the amplitude code sequence. In the example of the signal sequence used in the first embodiment, the output of the format converter 12 is as follows.

{·· | 111 | 111 | 101 | ·· || · || 1100110 | 1110110
| 11010 | 0101...} However, “||” represents a break between the block code length code string and the amplitude code string, and does not exist in the actual data string. Following the sequence are N blocks of amplitude code sequences.

Next, the decoding device 11 will be described. A code string output from the encoding device 10 is first input to the deformat converter 13. The deformat converter 13 separates the N block code length code string and the N block amplitude code string. The separated block code length code string and amplitude code string are processed in the same manner as in the first embodiment, and the code string is decoded.

In this way, by recognizing the end of the amplitude code string of the block based on the block code length and decoding the code, even if an error occurs in the amplitude code string, the code can be synchronized at the head of each block, The effect of the error does not propagate beyond. In addition, since the block code length code string and the amplitude code string are separated, transmission or recording can be performed without affecting the block code length code string due to an error generated in the amplitude code string due to noise or the like. Furthermore, since the block code length code string and the amplitude code string are separated, the error correction capability can be improved only for the block code length code string, so that the amount of information required for error correction can be reduced. .

Next, a third embodiment of the encoding device and the decoding device will be described with reference to FIG. 2 and FIG.
The configuration of the device of the third embodiment is the same as that of the second embodiment, except that the block end recognizer 8 is configured as shown in FIG. In the third embodiment, the encoder 4 adds the end code of the block to the amplitude code string of the block by the encoder 4 and transmits or records the amplitude code string including the end code. This is a decoding device that decodes a code string encoded by the device. This is performed to increase the reliability of recognition of the end of the block.

The block end recognition unit 8 shown in FIG. 3 includes an end code detector 14, a block code length end detector 15, and a selector 16. First, the decoded block code length is input to the block code length end detector 15, and the end of the block amplitude code sequence is detected based on the block code length. On the other hand, the end code detector 14 receives the amplitude code string and detects the end code at the end of the block amplitude code string to detect the end of the block amplitude code string. Either the output of the block code length end detector 15 or the output of the end code detector 14 is selected by the selector 16, and the selected one is input to the decoder 9.

The selector 16 selects the one without errors or the one with less false establishments. The process for performing error detection and error correction on the output of the encoding device is a general process. By performing such processing, it is possible to correct the error, and if the error can be detected even if the error cannot be corrected, the data in which no error is detected may be selected. Therefore, when an error occurs in the block code length code string and the error is detected by the error control processing but cannot be corrected, and no error occurs in the amplitude code string, or the error is corrected even if it occurs. If possible, the end of the code string of each block is recognized by the end code in the amplitude code string.

On the other hand, when no error occurs in the block code length code sequence or when the block code length code sequence can be corrected, the end of the code sequence of each block is recognized based on the block code length, as in the first embodiment. Also, the reliability of the block code length data and the amplitude code string can be improved by always comparing the block code length data and the end code of the amplitude code string. Note that the error control process may be performed using a Reed-Solomon code or the like, but is not limited to this, and may be any process that can detect or correct an error.

As described above, if one of the two means is selected and the end of the block is recognized, the end of the block can be recognized more accurately, and the influence of errors caused by noise or the like can be reduced. can do.

Although the block code length code string may be transmitted or recorded once, the reliability can be further improved by transmitting or recording it a plurality of times. In addition, transmission or recording may be performed with a different error correction and error detection capability than the block code length code sequence and the amplitude code sequence. The block code length is the total code length of the amplitude code string of one block. However, the present invention is not limited to this, and the block code length may be the total code length of the amplitude code string of a large block including a plurality of blocks. Good. However, when it is necessary to recognize the end of each block at the time of decoding, the end of each block is recognized by the terminal code. However, when an error occurs in the compounding of the terminal code, it is recognized by the block code length. The effect of the error may propagate to the block up to the end of the block. However, the amount of information required for encoding the block code length is reduced, and if a code string of a large block is recognized based on the block code length, the effect of the error does not propagate to subsequent blocks.

The method of encoding by replacing the connection of 0s to the end of the block with a terminal code is used. However, the present invention is not limited to this method, and an encoding method that does not require a terminal code may be used. When performing encoding that does not require a terminal code, a new terminal code may or may not be added. When a terminal code is added to an encoding method that does not require a terminal code, it is the same as the above-described second or third embodiment. In addition, when a code string can be recognized even if there is no terminal code at the time of decoding, the end of the block can be recognized. Either one of the block end and the block end recognized by the block code length may be selected from the result of error detection or the like.

Although an example of transmitting or recording an amplitude code string after transmitting or recording a block code length code string has been described above, the present invention is not limited to this, and the block code length code string is transmitted after transmitting or recording an amplitude code string. Alternatively, it may be recorded. Further, although the number of pixels included in the block is eight, the number is not limited to this and may be any number. Further, in the above-described embodiment, the number of blocks is set to N. The number may be any, and if the signal is an image signal, it may be one frame or a part of one frame.

The blocks may be one-dimensional or two-dimensional blocks or anything. In the case of a multi-dimensional block, it may be taken out in a predetermined order to form one data string. Further, although the image signal has been described as an example as the signal, any signal may be used as long as it is a data sequence. For example, a prediction error image composed of the prediction error signal predicted from another image frame or an audio signal may be used. It does not matter. Furthermore, an example has been described in which discrete cosine transform is performed for each block. However, the present invention is not limited to this, and other transforms may be used or no transform may be used. Further, the code table used for the above-described encoding is an example, and the present invention is not limited to this.

[0042]

As described above, the encoding method of the present invention provides
A method for encoding a digital signal sequence, comprising: dividing the digital signal sequence into small block signal sequences including a plurality of signals; encoding the small block signal sequence to obtain a small block code sequence; Measuring the total code length of a large block code string composed of a plurality of small block code strings as a block code length, and recognizing the end of the large block code string by the obtained block code length; The first block code added to the end of the block code string, or the second block code added to the end of the large block code string, or the total block length of the large block code string, By recognizing the end of the block, even if an error occurs in decoding due to noise or the like, the end of the code sequence of the block is more accurately recognized. Since the synchronization of decoding is restored at the beginning of each block, the effects of errors do not propagate beyond the blocks, and there is no displacement of the blocks, reducing image quality degradation due to errors. Becomes possible. Further, since the data can be transmitted or recorded with an increased error correction capability only for the block code length data sequence, the amount of information required for error correction can be reduced, and highly efficient encoding can be realized.

An encoding apparatus according to the present invention is an apparatus for encoding a digital signal sequence, comprising: a block divider for dividing the digital signal sequence into small block signal sequences comprising a plurality of signals; An encoder for encoding a small block signal sequence and an encoder for a small block code sequence; a block code length measuring device for measuring a total code length of a large block code sequence including one or a plurality of the small block code sequences; A block code length encoder that encodes the output of the length measuring device, and further transmits or transmits a first predetermined amount of block code length code strings output from the block code length encoder. After recording, a small block code sequence obtained by encoding a second predetermined amount of the small block signal sequences is collectively transmitted or recorded, or the second predetermined amount of the small block signal sequences is Tsu are those equipped with a format converter for rearranging the code sequence to click code sequence collectively transmission or transmission or recording recorded the first predetermined amount number of collectively the block code length code sequence after,
By separating the block code length code string from the amplitude code string, it is possible to transmit or record the data without affecting the block code length code string due to an error generated in the amplitude code string due to noise or the like. Further, since the block code length code string and the amplitude code string are separated, the error correction capability can be improved only for the block code length code string, and the amount of information required for error correction can be reduced. Efficient coding can be realized.

A decoder according to the present invention is a device for performing a decoding process on the code sequence obtained by the coding device, and comprises a block code length decoder for decoding the block code length code sequence. And an end recognizer that recognizes the end of the large block code sequence based on the output of the block code length decoder.
First means for recognizing the end of the large block code string by a first end code added to the end of the small block code string, or a second end code added to the end of the large block code string; By configuring the second means for recognizing the end of the large block code string by the block code length, and means for selecting the output of one of the first means and the second means, Even if an error occurs in decoding due to noise or the like, the end of the block coded data sequence can be more accurately recognized,
Since the synchronization of decoding is restored at the beginning of each block, the effects of errors do not propagate beyond the blocks, and the position of the blocks does not shift, thereby reducing image quality degradation due to decoding errors. It becomes possible.

[Brief description of the drawings]

FIG. 1 is a block diagram of an encoding device and a decoding device according to a first embodiment of the present invention.

FIG. 2 is a block diagram of an encoding device and a decoding device according to second and third embodiments of the present invention.

FIG. 3 is a block diagram of a block end recognizer according to a third embodiment of the present invention;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Encoding device 2 Decoding device 3 Block divider 4 Encoder 5 Block code length measuring device 6 Block code length encoder 7 Block code length decoder 8 Block end recognition device 9 Decoder 10 Encoding device 11 Decoding device 12 Format converter 13 Deformat converter 14 Termination code detector 15 Block code length termination detector 16 Selector

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁶ Identification code Agency reference number FI Technical display location H04N 7/30 H04N 7/133 Z

Claims (6)

(57) [Claims] 1. A digital signal string to a method for processing encoded blanking comprising the digital signal sequence from the plurality of signal
It divides into a lock signal sequence , encodes the block signal sequence to obtain a block code sequence , measures the total code length of the block code sequence as a block code length, and terminates the block code sequence by the obtained block code length. An encoding method characterized by recognizing a symbol. 2. A terminator added to the end of a block code string.
Signal or the total code length of the block code string
2. The encoding method according to claim 1, wherein the end of the block code string is recognized by the following. 3. A device for processing encoded digital signal sequence, blanking comprising the digital signal sequence from the plurality of signal
A block divider for dividing the lock signal sequence, the blocks
Encoder for determining the block code sequence encodes a click signal sequence
A block code length measuring device for measuring a total code length of the block code sequence; and a block code length encoder for coding an output of the block code length measuring device. . 4. A block code block block code length code string the total code length by encoding and the second predetermined amount blocks signal sequence encoded after transmitting or recording a first predetermined amount pieces together columns Collectively transmitting or recording a code string , or said second predetermined number of said block codes
A format converter for rearranging a code string such that the first predetermined number of the block code length code strings are collectively transmitted or recorded after the strings are collectively transmitted or recorded. 3. The encoding device according to 3. 5. A digital signal sequence comprising a plurality of signals.
Is divided into blocks the signal sequence, the block signal sequence there is provided an apparatus for processing decode the transmission or recorded code sequence for each encoding block code sequence, the total of the block code sequence
And a block code length decoder for decoding a block code length code sequence obtained by encoding code length, be provided with a recognizing termination recognizer the end of the block code sequence by the output of the block code length decoder A decoding device characterized by the above-mentioned. 6. An end recognition device is provided at the end of a block code string.
First means for recognizing the end of the block code sequence I by the addition to the termination code, second means for recognizing the end of the block code sequence by the block code length, which is the output of the block code length decoder 6. The decoding apparatus according to claim 5, further comprising: means for selecting an output of one of the first means and the second means.