JPH07273739A - Digital signal encoding and multiplexing device - Google Patents

Abstract

PURPOSE:To stabilize the overall code generation volume to not only shorten the extent of system delay but also reduce the output/input buffer capacity of an encoder/decoder by assigning modes different by generated code volumes to plural compressing/encoding devices. CONSTITUTION:At the time of encoding, the intra-code encoding mode and the inter-code encoding mode are controlled by an encoding mode control part 900 with respect to plural hybrid encoding parts 720a to 720d. The intra-code encoding mode is especially controlled so that it doesn't overlap with respect to time, and the variance width of the code volume in a multiplexing part 800 is reduced.

Description

Detailed Description of the Invention

[0001]

FIELD OF THE INVENTION The present invention compresses / compresses a plurality of image signals.
The present invention relates to an encoding digital signal encoding and multiplexing / multiplexing transmission device.

[0002]

2. Description of the Related Art In recent years, a high-efficiency compression / encoding system for reducing a required signal rate by removing the redundancy of an information source has been put into practical use. As a typical example of this, there is a so-called hybrid coding method in which a spatial redundancy is removed from a moving image signal by orthogonal transformation such as DCT, and motion detection / motion compensation is further applied to reduce temporal redundancy.

FIG. 7A shows an example of the structure of a hybrid coding device. An analog moving image signal is input to the input terminal 100, the preprocessing circuit 110 subjects the signal to digitalization, filtering, and the like, and the output signal is a subtractor 120 and a motion motion detection circuit 2 for detecting image motion.
00 is input. For convenience of explanation, assuming that the switches 220 and 230 are open, the preprocessing circuit 11
The output signal of 0 is directly input to the DCT circuit 130. The DCT circuit 130 handles an image signal for each block composed of 8 horizontal samples / 8 vertical lines, and DCT (Disc) transform which is one of orthogonal transforms for each block.
ret Cosine Transform) processing is performed. By this DCT transform, the 8 × 8 spatial region is transformed into the 8 × 8 frequency region. Here, as a general property of an image, it is known that spatially randomly arranged signals are concentrated in a specific area by DCT conversion. In the quantizer (Q) 140, D
A minute signal is cut out from the output signal of the CT circuit 130, and a signal concentrated portion is extracted. The originally truncated signal can be regarded as a component that is not important in expressing the original image, and is practically no problem. In this way, the number of signals required to represent the block is reduced by (DCT transform + quantization), in other words, the spatial redundancy is removed. The output of the quantizer 140 is input to the encoder 240. In the encoder 240,
Entropy coding is performed to further improve the coding efficiency by utilizing the fact that the output frequency of the quantization 140 has a deviation in the frequency of occurrence and assigning a code with a shorter code length to a signal with a higher frequency. .

As described above, the method in which the switches 220 and 230 are opened and the compression / encoding is performed by utilizing only the spatial property of the image is called intra encoding. Next, inter coding with the switches 220 and 230 closed will be described. The intra-coded image is Q ^-1 (inverse quantizer) 150 and DCT.
^-1 (inverse DCT circuit) 160 returns to the original spatial area, and then is input to the delay device 180 to form one image. The delay unit 180 obtains a signal delay corresponding to the time for forming one screen, and the intra-coded signal is input to the motion compensation circuit 190 and the motion detection circuit 200. In the motion detection circuit 200, the image signal currently being encoded output from the pre-processing circuit 110 and the signal delayed by the one screen period are input, and by comparing the two, a motionless region and a motion are detected. Area is detected, and the amount of movement is detected in a moving area.
In the motion compensation circuit 190, the motion information obtained from the motion detection circuit 200 is used to perform a spatial movement corresponding to the amount of motion in the motion area, and the input signal is output as it is in the still area with no motion amount. To be done. The output of the motion compensation circuit 190 is added to the subtractor 120, and the difference from the signal currently being encoded is taken. Since not only the still image area but also the moving area is spatially moved by motion compensation, the subtractor 120
Output becomes almost zero, and a small output is obtained in a region where motion compensation is insufficient such as a change in the shape of an object on the screen. As described above, in the inter coding mode, not only the temporally stationary area but also the area is moved by motion compensation, and the redundancy can be removed in the area that can be overlapped with the previous screen. Also,
Even in a region where motion compensation is insufficient, the difference value from the previous screen is the DCT circuit 130, the quantizer 140, and the inverse quantizer 150.
And the inverse DCT circuit 160, and then input to the adder 170, and the other input end of the adder 170 receives the previous screen signal which is the output signal of the motion compensation circuit 190. Images can be played back. Also in this inter-coding mode, the encoder 240
Then, the output signal of the quantizer 140 is entropy-encoded, and the motion amount obtained from the motion detection circuit 200 is also output together with the code of the encoded difference signal.

Here, the role of the intra-coded image will be described. When the hybrid coding method is applied to broadcasting or a storage device, it is not decided when the decoding start in the decoder will occur. If this is a broadcast,
It can be considered that the decoding start is requested when the decoding device is powered on or when the channel is changed, and in the storage device, for example, the CDROM reproducing device, the code start is requested when the pickup unit starts to pick up the code at random access. However, in the compressed / encoded signal adopting the hybrid encoding method described above, a normal decoded signal cannot be obtained even if the decoding is started from the inter-coded image. This is because the inter-coded image encodes only the difference value from the previous screen, and the decoder can obtain a normal decoded image only after the intra-coded image is input. . Therefore, in the above-mentioned application field, the intra coding mode is inserted relatively frequently, and the intra coding is normally performed once every several screens to several tens of screens. The intra-encoding cycle is determined in FIG. 7 (A).
The mode control unit 210 of FIG.

Now, in the intra coding mode in which compression / encoding is performed by utilizing the redundancy in the screen and in the inter coding mode in which only the difference value from the previous screen is coded, the code generated per one screen There is a big difference in quantity. FIG. 7 (B)
Is a schematic representation of the amount of generated codes per screen. The cycle of intra-coding is set to 4, and the generated code amount is larger than that of the second-to-fourth and the sixth to eighth in which the first, fifth and ninth screens are intra-coded and inter-coded. ing. Although it depends on the application field, the generated code amount in the normal intra-coding mode is about 2 to several times the code amount generated by the inter-coding. On the other hand, the medium that transmits the code itself has a fixed or nearly fixed transmission rate, and the signal bandwidth determines the transmission rate in broadcasting, and the rate at which the code is picked up in the storage system determines the transmission rate. become.
Therefore, the output buffer 250 shown in FIG. 7A is provided in order to absorb the fluctuation of the generated code amount depending on the encoding mode.

In the output buffer 250, the code amount input varies greatly depending on the encoding mode, but the output code amount is kept constant, and its capacity is determined by the variation width of the generated code amount. When the fluctuation amount is large, that is, when the ratio of the generated code amount between the encoding modes is large, a large capacity output buffer is required. Further, in the decoding device for decoding this hybrid-coded signal, on the contrary, since the code input from the transmission medium at a fixed rate is received, and the code amount required per screen differs depending on the coding mode, Output buffer 2 in encoding device
An input buffer having a function opposite to 50 is required.
The capacity of this input buffer is also determined by the fluctuation range of the generated code amount depending on the encoding mode.

FIG. 8 shows the output buffer capacity in the encoding device, the input buffer capacity in the decoding device, and the code amount accumulated in each buffer at each time, which are required depending on the code fluctuation width. is there. Figure 8 for simplicity
In (A) and (B), a code of 6 is generated in intra coding, and a code of 2 is generated in inter coding. In FIGS. 9 (A) and 9 (B), a code of 9 is generated in intra coding and a code of 1 is generated in inter coding. It was done. In both FIGS. 8 and 9, the intra coding mode occurs at a rate of 1 in 4 and the code transmission rate is 3. Further, the encoding process in the encoding device and the decoding process in the decoding device are ideal, and the processing time is set to zero. Therefore, taking FIG. 8 as an example, the coding is started at time 1, the intra coding mode is selected, and the coding device output buffer code amount is 6. Since the code transmission rate is 3, the transmission of the code is started from time 1 and its inclination is -3. Immediately before time 2, the code amount accumulated in the output buffer is 3 (6-3 = 3), and the code amount is increased to 5 at time 2 due to the inter coding mode. On the other hand, the input buffer code amount of the decoding device starts from time 1 and increases with a gradient of 3, and at time 3 all codes corresponding to the image number (a) are received and immediately decoded. This operation is the same in FIG. Here, in FIG. 8, by setting both the output buffer capacity of the encoding device and the input buffer capacity of the decoding device to 6, stable image transmission can be performed without overflow or underflow of each buffer. There is. On the other hand, in FIG. 9, this is possible by setting the buffer capacity to 9. The difference in capacity between the two is due to the ratio of the generated code amount depending on the encoding mode, that is, the fluctuation range of the code amount, and in FIG. 9 where the fluctuation range is large, a large capacity buffer is required. Further, when comparing FIGS. 8A and 8B, it can be seen that there is a difference in the decoding time of the image number (a). Since the decoding is performed at time 3 in FIG. 8, the total system delay amount including encoding / decoding is 2, whereas in FIG. 9, it is 3. This is also due to the ratio of the generated code amount in the coding mode, and the delay time is longer in FIG. 9 in which the variation width of the code amount is larger.

Generally, the rate at which a code is transmitted is
As described above, it is determined in advance. On the other hand, in the high efficiency coding method, the fluctuation range of the generated code amount depending on the coding mode is relatively large. In the high-efficiency coding method, the coding efficiency is increased by increasing the fluctuation range of the generated code amount, and therefore the buffer capacity that absorbs the fluctuation range is naturally large, and the entire coding and decoding including The amount of system delay tends to increase.

In general, it is desirable that the delay amount of the entire system be small. Especially, in the broadcasting that attaches importance to simultaneity and the teleconferencing system that performs bidirectional encoding / decoding, the delay amount must be restricted. Therefore, due to this restriction, it is necessary to reduce the fluctuation range depending on the coding mode, in other words, perform coding while sacrificing the coding efficiency to some extent.

In addition, the coding efficiency achievable in this coding system, in other words, the code rate read from the output buffer of the coding apparatus and the transmission line capacity for transmitting the code do not have a consistent relationship. It is known that a plurality of encoded images can be multiplex-transmitted in a transmission line capacity used for broadcasting.

FIG. 10 shows an example of the structure of an apparatus for multiplex transmission of a plurality of hybrid-coded images. Different image signals inputted to the input terminals 300, 400 and 500 are processed by the preprocessing circuit 310, by the method shown in FIG.
Hybrid coding unit 32 via 410 and 510
Output buffers 330, 430, 53 that are compressed / encoded by 0, 420, 520 and absorb the fluctuation of the generated code amount.
The multiplexed signal is converted into a fixed rate by 0 and is input to the multiplexing unit 600. In the multiplexing unit 600, each code is multiplexed by a fixed time axis, frequency axis, or other method, and a multiplexed signal is output to the output terminal 610. It will be. The code transmission rate at this time is 3 for each encoding / decoding device.
Therefore, the entire code transmission rate, that is, the multiplexing unit 600
Has an output rate of 12. Therefore, even in the case of multiplex transmission, the output buffer capacity of each encoding device and the input buffer capacity of each decoding device are the same as those shown in FIG. 7A, and the delay amount of the entire system is also increased. Will be the same.

[0013]

In an apparatus for encoding / multiplexing a plurality of images seen in a conventional example, in order to set a high encoding efficiency in a method of fixing an output code rate in each encoding apparatus. Increases the capacity of the output and input buffers. On the other hand, in order to reduce the system delay amount, it was necessary to reduce the coding efficiency and the buffer capacity.

Therefore, an object of the present invention is to eliminate the above-mentioned drawbacks and to reduce the system delay amount in a state where the coding efficiency is set high, and to control the intra coding mode in each coding device. A mode controller is provided,
As a result, the generation time of the intra coding mode of each encoder is shifted, and the total sum of the code amounts generated by each encoder is made constant.

[0015]

SUMMARY OF THE INVENTION The present invention is directed to a plurality of compression / encoding devices having a plurality of modes having different generated code amounts, and a multiplexer for multiplexing output signals of the plurality of compression / encoding devices. And a signal amount control device for controlling the encoding modes of the plurality of compression / encoding devices to reduce the fluctuation width of the code amount in the multiplexing device.

[0016]

According to the above means, the sum of the code amounts generated by the respective coding devices can be made constant by the signal amount control device.

[0017]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 shows an embodiment of the present invention. Input terminal 71
a-71b are supplied with, for example, a plurality of image signals to be multiplexed, and pre-processing circuits 710a-710 are respectively provided.
It is input to d. The preprocessing circuits 710a to 710d are blocks that perform digital and filtering as in the conventional case. Outputs of the preprocessing circuits 710a to 710d are input to hybrid coding units (coding devices) 720a to 720d. This hybrid encoding section 720a-72
0d is also close to the conventional one, but the coding mode control unit 900 is different from the mode control unit 210 incorporated in the conventional example.
The input signal from has been added. Further, there is no encoder output buffer memory in the conventional example, and the multiplexer 800
It has a built-in configuration. The output of each encoding device is input to the buffer memory 810 in the multiplexing unit 800. At this input, the amount of generated codes in the coding mode fluctuates greatly. However, the buffer memory 81
Inside 0, the total sum is averaged and maintained at a constant coding amount as a whole. Buffer memory 810
The output of is subjected to multiplex processing by the multiplex processing unit 820 and is output to the output terminal 830.

FIG. 2 shows the generated code amount of each encoding device under the control of the encoding mode control unit 900 in screen units. The symbol I indicates the generated code amount of one screen that is intra-coded, and the symbol P indicates the generated code amount when inter-coded. Similar to the conventional example, the inter coding mode in each coding device is inserted in cycle 4. Here, the encoding mode control unit 9
00 indicates the coding mode to each coding device, and the intra coding mode is specified to the coding device, for example, 720a → 720b → 720c → 720d → 720a.
... is assigned cyclically.

FIG. 2 also shows the total sum of generated code amounts for each encoder. Coding mode control unit 900
There is one device that performs intra-coding and three devices that perform inter-coding at the time when the coding mode is instructed by, and since this relationship holds at all times, the total amount of generated codes is almost It is constant. Therefore, the buffer memory 81 in the multiplexing unit 800
At 0, the code amount input from each encoding device fluctuates largely as in the conventional example, but the sum of the code amounts is kept almost constant, and the system delay amount is significantly small. Become. This will be further described below.

Here, considering the case where the generated code amount in the intra coding mode is 6 and that in the inter coding mode is 2 for simplification as in the conventional example, consider that at time 1, the coding device 720a The generated code amount is 6, and the coding devices 720b, 720c, and 720d have a value of 2, and the total number of codes at this time is 12. At time 2, the coding device 720b has a value of 6 and the other devices have a value of 2, and the total code amount is 12 as well. In this way, the total sum of the generated code amounts is always 12, which is a constant value. Here, the transmission rate at the output of the multiplexer is set to 12 as in the conventional example.

In FIGS. 3A and 3B, the buffer capacity of the encoder and the buffer capacity of the decoder when multiplexing are 1
The number of codes accumulated in each buffer when 2 is shown. In the encoder, the image data α1, β1, γ1, and ξ1 encoded at time 1 are generated, and the code amount accumulated in the output buffer is 12. Since this is read at the transmission rate 12, the code amount becomes 0 immediately before time 2 and the image data α2 and β encoded again at time 2 are obtained.
2, γ2, ξ2 occur. On the other hand, the code amount in the decoding device starts from 0 at time 1, becomes 12 immediately before time 2, and is immediately decoded. At time 2, it returns to 0 again.

Since such an operation is repeated, the delay amount of the entire system becomes equal to one image period, which is shorter than the delay amount 2 in the conventional example. Further, comparing the system delay amount in the generated code amount (= 9) in the intra coding mode and the generated code amount (= 1) in the inter coding mode shown in FIG. 9 of the conventional example, three images are obtained in the conventional example. Although it is a period, in the embodiment of the present invention, it is one image period as in FIG.

As described above, in the present invention, the system delay is kept constant regardless of the ratio of the generated code amounts in the intra coding mode and the inter coding mode. That is, the delay time can be kept short even if the coding efficiency is increased by increasing the variation of the generated code amount.
Further, when comparing the capacities of the output buffers in the encoding devices, in the example of FIG.
The total sum of capacity is 9 × 4 =
Although it was 36, it is 12 in the present embodiment, and it can be seen that it is greatly reduced. Further, comparing the input buffer capacities in the decoding devices, first, in the case of decoding all four types of transmitted images, in the conventional example, a buffer having a capacity of 9 is required in each device, and the total amount of buffers is required. Is 36, whereas in this embodiment it is 12,
Has been significantly reduced. Further, in the case of decoding only one of the four types of transmitted images, an input buffer having a capacity of 9 is required in the conventional example, and in the present embodiment, it is transmitted in one certain image period. It suffices to have a capacity equal to the maximum value of the code amount per image, and the capacity is 9 and is equal to the conventional example (in the case of the multiplexed code, there are 12 codes in one image period, but the necessary image data Only 9 need be stored, so the capacity can be 9).

In the above description, an example was described in which the generated code amounts in the intra coding mode or the inter coding mode were the same for the four types of images. The present invention can be implemented in a higher-definition encoding than the fourth image or in an application field in which the image size is increased.

FIG. 4 illustrates encoders 720a and 720b.
Is an example in which the generated code amount of is doubled with respect to 720c and 720d. At time 1, inter coding is selected at intra coding 720a, 720c, and 720d at time 720a, at time 2 intra coding is selected at intra coding 720a, 720c, and 720d at time 720b, and at time 3 at 720c. And 720d
In intra coding 720a and 720b, inter coding is selected, and these three cycles are repeated.
Since the code amount of 720a and 720b is twice the code amount of 720c and 720d, the code amount of 720c and 720d combined is the same as 720a or 720b. By selecting intra coding for 720c and 720d simultaneously at time 3, It can be seen that the total sum of generated code amounts is kept constant.

As described above, the present invention can be applied to image compression / encoding systems having different code amounts. Also, FIG.
It is obvious that the present invention can be applied without changing the cycle for selecting intra-encoding of the encoding apparatus as shown in FIG. 6 or even when there is no cycle as shown in FIG. By controlling the encoding modes of a plurality of encoding devices in this way,
Various application examples are conceivable within a range not deviating from the gist of the present invention in which the code amount generated during multiplexing is constant.

[0027]

According to the present invention, modes having different generated code amounts are assigned to a plurality of compression / encoding devices to stabilize the total code generation amount, shorten the system delay amount, and perform the encoding / encoding.
The output / input buffer capacity of the decoding device can be reduced.

[Brief description of drawings]

FIG. 1 is a diagram showing an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing an operation example of the apparatus of the present invention.

FIG. 3 is an explanatory view showing an operation example of the apparatus of the present invention.

FIG. 4 is an explanatory view showing an operation example of the apparatus of the present invention.

FIG. 5 is an explanatory view showing an operation example of the apparatus of the present invention.

FIG. 6 is an explanatory view showing an operation example of the apparatus of the present invention.

FIG. 7 is a diagram showing a conventional encoding device and an operation example thereof.

FIG. 8 is an explanatory diagram showing an operation example of a conventional device.

FIG. 9 is an explanatory diagram showing an operation example of a conventional device.

FIG. 10 is a diagram showing a conventional multiprocessing device.

[Explanation of symbols]

710a to 710d ... Preprocessing circuits, 720a to 720d
... Hybrid encoding unit, 800 ... Multiplexing unit, 810 ...
Buffer memory, 820 ... Multiplex processing unit, 900 ... Coding mode control unit.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical indication H04N 7/081 7/30 7/32 H04N 7/137 Z

Claims (5)

[Claims] 1. A plurality of compression / encoding devices having a plurality of modes having different generated code amounts, a multiplexer for multiplexing output signals of the plurality of compression / encoding devices, and the plurality of compression / encoding devices. A digital signal coding and multiplexing device, comprising: a signal amount control device for controlling each coding mode of the coding device to reduce the fluctuation range of the coding amount in the multiplexing device. 2. The compression / encoding device is a device for compressing / encoding a moving image signal. As a mode in which the generated code amount is different, the first mode is a spatial redundancy of the moving image signal. Is used to perform compression / encoding, and the second mode is a mode in which compression / encoding based on temporal redundancy is added in addition to the first mode. Item 1. A digital signal encoding and multiplexing device according to item 1. 3. A compression / encoding mode using the spatial redundancy described above divides an image signal to create a plurality of blocks,
The compression / encoding mode using the above-mentioned temporal redundancy performs the motion prediction between the images which are temporally separated from each other, and performs the orthogonal transformation for each block and quantizes the output. The digital signal encoding and multiplexing apparatus according to claim 2, wherein motion compensation is performed according to a value. 4. The code amount control device generates a large amount of generated code for a compression / encoding device (a) or a compression encoding device group (A) at a specific time for each compression / encoding device mode. Mode allocation, compression / encoding device (b)
Alternatively, the digital signal coding and multiplexing apparatus according to claim 1, wherein a mode having a small amount of generated code is assigned to the compression coding apparatus group (B). 5. The multiplexing device is provided with a buffer memory for absorbing code amount fluctuations for the purpose of outputting the outputs of a plurality of compression / encoding devices at a fixed rate. The digital signal encoding and multiplexing device according to claim 1.