JPH10164577A - Video encoding device - Google Patents

Abstract

(57) [Summary] An object of the present invention is to provide a moving picture coding apparatus capable of performing variable rate coding of a moving picture in real time. SOLUTION: A moving picture coding apparatus 100 according to the present invention.
When encoding a moving image, a value obtained by multiplying the generated code amount in the screen by the average value of the quantization width in the screen is determined as the screen complexity, and the current value with respect to the average value of the past screen complexity is calculated. A target code amount or a target quantization width of a screen is determined based on the ratio of the screen complexity, and the coding or quantization of the screen is performed using the target code amount or the target quantization width. It is.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving picture coding apparatus, and more particularly to a moving picture coding apparatus capable of performing variable rate coding in real time.

[0002]

2. Description of the Related Art In recent years, the MPEG2 system has been widely used as a highly efficient coding method for moving images. In the MPEG2 system, a screen is divided into macroblocks of 16 × 16 pixels, and the macroblocks are further divided into blocks of 8 × 8 pixels. Then, a discrete cosine transform is performed for each block to obtain a discrete cosine transform coefficient. The discrete cosine transform coefficient is quantized by dividing the obtained discrete cosine transform coefficient by a quantization width and an 8 × 8 quantization matrix corresponding to each frequency component. The quantized discrete cosine transform coefficients are subjected to variable-length coding to form a code sequence.
The quantization width in the quantization can be changed for each macroblock, whereby the generated code amount can be controlled. As a method of controlling the generated code amount in the MPEG2 TEST MODEL3, a fixed rate method for making the generated code amount constant in GOP units has been proposed.

In this conventional fixed-rate encoding apparatus, the code amount allocated to a moving image with a small motion (the amount of information is small) and the code amount allocated to a moving image with a large motion or a complex moving image (a large amount of information). Is the same. Therefore, when the set rate is low, an image with a large amount of information is visually deteriorated in image quality. This degradation in image quality can be solved by increasing the set rate. However, for an image with a small amount of information, the allocated code amount becomes wasteful, for example, when a code string is recorded on a recording medium. However, there is a problem that the recording time is shortened.

In order to solve the above-mentioned problems, Japanese Patent Laid-Open Publication No. Hei 6-141298 discloses a method of controlling a set rate for each unit time so that the total code amount of a moving image becomes a predetermined value. , A variable bit rate encoding device that improves the image quality has been proposed. In this conventional variable bit rate encoding device, the same input image is encoded twice. First, the first encoding (temporary encoding) of an input image is performed using a fixed quantization width.
In the code sequence generated by the provisional encoding, the generated code amount is counted for each unit time and stored as a provisional transfer rate. Then, the target transfer rate is set from the provisional transfer rates counted per unit time so that the total code amount of the input image becomes a predetermined value. Next, the second encoding (actual encoding) of the input image is performed while controlling the generated code amount so as to match the target transfer rate for each unit time.

[0005]

However, in the above-mentioned conventional variable bit rate coding apparatus, two coding operations must be performed in order to obtain a code string. This variable bit rate encoding device requires twice the length of the video and cannot perform actual encoding unless all of the moving image is provisionally encoded once. There was a problem that cannot be used.

SUMMARY OF THE INVENTION The present invention has been made in view of the above situation, and has as its object to provide a moving picture coding apparatus capable of performing variable bit rate coding of a moving picture in real time.

[0007]

According to a first aspect of the present invention, there is provided a moving picture coding apparatus which divides each screen of a moving picture including a plurality of screens into a plurality of blocks. Block conversion means, image conversion means for converting a plurality of blocks of each screen divided by the block conversion means into coefficients, and coefficients of a plurality of blocks of each screen converted by the image conversion means. ,
Quantizing means for quantizing using a quantization width; code string generating means for generating a code string for each screen from coefficients of a plurality of blocks of each screen quantized by the quantization means; A code amount counting means for counting the generated code amount in each screen from the code string of each screen generated by the code string generation means, and an average value for obtaining an average value of a quantization width used in the quantization means. Arithmetic means, and a complexity of calculating a value obtained by multiplying the generated code amount counted by the code amount counting means by the average value of the quantization width calculated by the average value calculating means as the screen complexity of each screen. Calculating means for calculating the average screen complexity of the past screen calculated by the complexity calculating means, and calculating the screen complexity of the current screen calculated by the complexity calculating means with respect to the average screen complexity of the past screen Find the percentage of A target code amount determining means for determining a target code amount of the next screen from the ratio, obtaining a quantization width of the next screen using the target code amount, and outputting the quantization width to the quantization means. And a moving picture coding apparatus characterized by comprising:

According to a second aspect of the present invention, in the moving image encoding apparatus according to the first aspect, the target code amount determining means includes a picture type (I picture, P picture) of a screen. , B picture), the average screen complexity of the second past screen per unit time is calculated from the average screen complexity of the first past screen, and the screen complexity of the first current screen and the second screen The screen complexity of a second current screen per unit time is determined from the screen complexity of two screens other than the current screen of one past, and the above-mentioned average screen complexity of the second past screen is calculated. A target generated code amount for the next screen is determined from the ratio of the screen complexity of the second current screen, and a quantization width of the next screen is obtained using the target generated code amount. Output to the conversion means That is a moving picture coding apparatus.

According to a third aspect of the present invention, there is provided the moving picture coding apparatus according to the first aspect, wherein the moving picture coding apparatus includes a predetermined target average coding rate and the predetermined code. Virtual buffer means for obtaining the allocatable code remaining amount from the generated code amount counted by the amount counting means and holding the allocatable code remaining amount, wherein the target code amount determining means comprises a picture For each type (I picture, P picture, B picture), the average screen complexity of the second past screen per unit time is determined from the average screen complexity of the first past screen,
The screen complexity of the second current screen per unit time is obtained from the screen complexity of the first current screen and the screen complexity of two screens other than the current screen of the first past, and 2
Target generation of the next screen from the ratio of the screen complexity of the second current screen to the average screen complexity of the previous screen and the remaining allocable code per unit time held in the virtual buffer. Determining a code amount, obtaining a quantization width of the next screen using the target generated code amount, and outputting the quantization width to the quantization means. is there.

According to a fourth aspect of the present invention, there is provided a moving picture coding apparatus comprising: a block converting means for dividing each picture of a moving picture comprising a plurality of pictures into a plurality of blocks; Image conversion means for converting the plurality of blocks of each screen into coefficients, respectively, and quantizing the coefficients of the plurality of blocks of each screen converted by the image conversion means using a quantization width. Means, code string generating means for generating a code string of each screen from coefficients of a plurality of blocks of each screen quantized by the quantization means, and each of the screens generated by the code string generating means A code amount counting means for counting the generated code amount in each screen, an average value calculating means for obtaining an average value of the quantization width used by the quantization means, and a code amount counting means. A complexity calculating means for calculating a value obtained by multiplying the counted generated code amount by an average value of the quantization width calculated by the average value calculating means as the screen complexity of each screen; and The calculated average screen complexity of the past screen is calculated, and the ratio of the screen complexity of the current screen calculated by the complexity calculating means to the average screen complexity of the past screen is calculated. A moving image encoding apparatus comprising: a target quantization width determining unit that determines a target quantization width of a screen and outputs the target quantization width to the quantization unit.

According to a fifth aspect of the present invention, in the moving picture encoding apparatus according to the fourth aspect, the target quantization width determining means includes a picture type (I picture, P picture) of a screen. Picture, B picture), the average screen complexity of the second past screen per unit time is calculated from the average screen complexity of the first past screen,
The screen complexity of the second current screen per unit time is obtained from the screen complexity of the first current screen and the screen complexity of two screens other than the current screen of the first past, and 2
Determining the target quantization width of the next screen from the ratio of the screen complexity of the second current screen to the average screen complexity of the past screen, and outputting the target quantization width to the quantization means. A moving image encoding apparatus.

According to a sixth aspect of the present invention, there is provided the moving picture coding apparatus according to the fourth aspect, wherein the moving picture coding apparatus comprises a predetermined target average coding rate and the predetermined code. The apparatus further comprises virtual buffer means for obtaining the allocatable code remaining amount from the generated code amount counted by the amount counting means and holding the allocatable code remaining amount. For each picture type (I picture, P picture, B picture), the average screen complexity of the second past screen per unit time is calculated from the average screen complexity of the first past screen, and the first current The screen complexity of the second current screen per unit time is obtained from the screen complexity of the second screen and the screen complexity of two screens other than the current screen of the first past, and Screen flat Determining a target quantization width of the next screen from the ratio of the screen complexity of the second current screen to the screen complexity and the remaining amount of codes that can be allocated per unit time held in the virtual buffer; A moving picture coding apparatus characterized in that said target quantization width is output to said quantization means.

According to a seventh aspect of the present invention, there is provided the moving picture coding apparatus according to any one of the first to third aspects, wherein the target generation determined by the target code amount determining means is performed. A moving image encoding apparatus characterized in that the code amount is a target generated code amount for a current screen.

According to an eighth aspect of the present invention, there is provided a video encoding apparatus according to any one of the fourth to sixth aspects, wherein the target quantization width determined by the target quantization width determining means is selected. A moving image encoding apparatus characterized in that a quantization width is a target quantization width for a current screen.

According to a ninth aspect of the present invention, there is provided the moving picture coding apparatus according to any one of the first to third aspects, wherein the target generation amount determined by the target code amount determining means is used. A moving image encoding apparatus characterized in that an upper limit and a lower limit are set for a code amount.

According to a tenth aspect of the present invention, in the moving image encoding apparatus according to any one of the fourth to sixth aspects, the target image determined by the target quantization width determining means is provided. A moving picture coding apparatus characterized in that an upper limit value and a lower limit value are set for a quantization width.

[0017] Further, according to an eleventh aspect of the present invention, in the video encoding apparatus according to any one of the fourth to sixth aspects, the picture type (I picture, P picture, B picture Picture)
A moving picture encoding apparatus characterized in that the target quantization width determined by the target quantization width determination means is multiplied by a predetermined coefficient, and the value is used as the target quantization width.

According to a twelfth aspect of the present invention, in the moving picture coding apparatus according to any one of the first to third aspects, the moving picture coding apparatus obtains the average value of the past screen complexity. Is a constant value for each of the picture types of the screen (I picture, P picture, B picture).

According to a thirteenth aspect of the present invention, in the moving picture coding apparatus according to any one of the fourth to sixth aspects, the moving picture coding apparatus obtains the average value of the past screen complexity. Is a constant value for each of the picture types of the screen (I picture, P picture, B picture).

According to a fourteenth aspect of the present invention, there is provided the moving picture coding apparatus according to the first aspect, wherein the target generated code amount T determined by the target generated code amount determining means is equal to the target generated code amount T. The value K obtained by dividing the current screen complexity by the average of the past screen complexity is a function f (K) that satisfies f (K1) ≦ f (K2) when K1 <K2.
Is a moving picture coding apparatus characterized by satisfying T = f (K).

According to a fifteenth aspect of the present invention, there is provided the video encoding apparatus according to the second aspect, wherein the target generated code amount T determined by the target code amount determining means; A value K obtained by dividing the fourth screen complexity by the third screen complexity is f when K1 <K2.
For a function f (K) that satisfies (K1) ≦ f (K2), T
= F (K).

The video encoding device according to claim 16 of the present invention is the video encoding device according to claim 3, wherein the target generated code amount T determined by the target code amount determination means; The value K obtained by dividing the fourth screen complexity by the third screen complexity is g (K1, K1) when the remaining code amount R per unit time and K1 <K2 are satisfied.
R) ≦ g (K2, R), and when R1 <R2, g
Function g (K, R) that satisfies (K, R2) ≦ g (K, R2)
Is a moving picture coding apparatus characterized by satisfying T = g (K, R).

According to a seventeenth aspect of the present invention, in the moving picture encoding apparatus according to the fourth aspect, the target quantization width Q determined by the target quantization width determining means is equal to or smaller than the target quantization width Q. , The value K obtained by dividing the current screen complexity by the average value of the past screen complexity and the function f satisfying f (K1) ≧ f (K2) when K1 <K2. = F (K).

The video encoding apparatus according to claim 18 of the present invention is the video encoding apparatus according to claim 5, wherein: the target quantization width Q determined by the target quantization width determination means; A value K obtained by dividing the fourth screen complexity by the third screen complexity is f when K1 <K2.
For a function f (K) that satisfies (K1) ≧ f (K2), Q
= F (K).

According to a nineteenth aspect of the present invention, in the moving picture coding apparatus according to the sixth aspect, the target quantization width Q and the target quantization width determined by the target quantization width determining means are used. , The value K obtained by dividing the fourth screen complexity by the third screen complexity is g (K) when the assignable code remaining amount R per unit time and K1 <K2 are satisfied.
For a function g that satisfies g (K, R2) ≧ g (K, R2) when (1, R) ≧ g (K2, R) and R1 <R2, S = g (K, R) A moving picture coding apparatus characterized by satisfying the following.

[0026]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a moving picture coding apparatus according to the present invention when the MPEG2 system is used as a coding method will be described with reference to the drawings. Embodiment 1 FIG. First, Embodiment 1 will be described with reference to the drawings. FIG. 1 is a block diagram of a video encoding device according to the first embodiment. In the figure, reference numeral 100 denotes a moving image coding apparatus according to the first embodiment. As described above, the moving image coding apparatus 100 includes a block converter 101 as a block converting means, a differentiator 102, and a switch 10
3, 111, and an orthogonal transformer 104 as an image converting means.
And a quantizer 105 as a quantizing means, and a code amount controller 1 integrating code amount counting means, average value calculating means, complexity calculating means, target code amount determining means, and virtual buffer means.
06, a variable-length encoder 107 as code generation means,
Buffer 108, inverse quantizer 109, inverse orthogonal transformer 110, motion compensation predictor 112, and frame memory 1
13 and an adder 114. In the first embodiment, a moving image input to such a moving image encoding apparatus 100 is converted into frames 201 to 2 as shown in FIG.
27, and frames 201 to 201
227 are arranged in chronological order with the frame 201 at the top. Also, I, added to each of the frames 201 to 227,
The symbols P and B indicate that each frame is an I picture (intra-frame coded image), a P picture (inter-frame forward prediction coded image), and a B picture (bidirectional prediction coded image), respectively. I have.

First, the basic operation of the moving picture coding apparatus 100 when an input frame is an I picture will be described with reference to the drawings. As shown in FIG. 2B, an input frame constituting a moving image input to the moving image encoding apparatus 100 is divided into 16 × 16 macroblocks by a block converter 101. If the input frame is an I-picture, switch 103 is connected to b, so that the macroblock is input to orthogonal transformer 104. The orthogonal transformer 104, as shown in FIG.
One macroblock is divided into blocks of 8 × 8 pixels, and an orthogonal transform is performed for each block to convert each block into an orthogonal transform coefficient. FIG. 2C shows blocks included in one macroblock when the image format is 4: 2: 0.

Next, the orthogonal transform coefficients are quantized by a quantizer 105. This quantization is performed by dividing the orthogonal transform coefficient by a quantization width given by the code amount controller 106 and an 8 × 8 quantization matrix corresponding to each frequency component. The quantized orthogonal transform coefficients are converted into a code string by the variable length encoder 107 and input to the buffer 108. The code string input to the buffer 108 is read out from the buffer 108 at a predetermined rate in a burst or continuous manner, and becomes data to be stored in a storage medium or the like.

On the other hand, the orthogonal transform coefficients quantized by the quantizer 105 are also input to an inverse quantizer 109, where they are subjected to inverse quantization, and further subjected to inverse orthogonal transform by an inverse orthogonal transformer 110. It becomes a locally decoded image. When the input image is an I picture, the switch 111 is not connected, and the locally decoded image is stored in the frame memory 113 as it is.

Next, the basic operation of the moving picture coding apparatus 100 when the input frames are P and B pictures will be described. An input frame constituting a moving image input to the moving image encoding apparatus is, like the case of the I picture,
First, the block is divided into macroblocks composed of 16 × 16 pixels by the block converter 101. If the input frames are P and B pictures, the switch 103 is connected to a, so that the macroblock is input to the motion compensation predictor 112, and the motion prediction is performed using the locally decoded image stored in the frame memory 113. Done. For example, if the input frame is frame 207 (P picture),
The frame 204 (P
The motion prediction is performed using the locally decoded image of the current picture (picture) as a reference frame. Also, if the input frame is frame 20
6 (B picture), the frame 204 (P
(Picture) and the locally decoded image of the frame 207 (P picture) are used as reference frames to perform motion prediction. The motion vector of the input frame obtained by the motion prediction is
The variable length coder 107 performs variable length coding. The difference between the macroblock of the input frame and the macroblock of the reference image obtained by the motion prediction is obtained by a differentiator 102, and the difference value is a difference macroblock.

The difference macroblock is converted into a code string by an orthogonal transformer 104, a quantizer 105, and a variable-length encoder 107 in the same manner as in the case of an I picture.
8 is input. When the input frame is a P-picture, the quantized orthogonal transform coefficients of the differential macroblock are also input to the inverse quantizer 109 and subjected to inverse quantization. Is performed, the addition with the reference image obtained by the motion prediction is performed by the adder 114 to be a locally decoded image, which is stored in the frame memory 113. If the input frame is a B picture, there is no need to generate a locally decoded image.

In the above operation, the code amount controller 106 counts the code amount input to the buffer 108 and obtains the complexity of the frame using the counted code amount and the quantization width of each frame. , A new quantization width is determined from the complexity and sent to the quantizer 105. The structure and operation of the code amount controller 106 will be described below.

First, when the input frame is a frame 201 (I
The detailed operation of the code amount controller 106 in the case of (picture) will be described below with reference to the flowchart of FIG. FIG. 3 is a flowchart showing the operation of the code amount controller. First, before starting to encode an input frame, the code amount controller 106 determines in step 301 that the variables Xi, X
Initial values of p, Xb, AvgXi, AvgXp, and AvgXb are set. Xi, Xp, and Xb indicate the frame complexity of the I, P, and B pictures immediately before the input frame that is currently being encoded. AvgXi is the average frame complexity for I-pictures prior to the I-picture immediately preceding the input frame currently being coded. A
vgXp, AvgXb are the P, B pictures before the P, B picture immediately before the input frame to be currently encoded.
Average frame complexity for B pictures. For example, when the input frame is the frame 222 (P picture), Xp when encoding the input frame is the frame complexity of the frame 219 (P picture), and Avg
Xp is the average frame complexity of a P picture before (including the frame 216) before the frame 216. The code amount controller 106 determines such variables Xi, Xp, Xb, Av
At the same time as setting the initial values of gXi, AvgXp, and AvgXb, the initialization operation by (Equation 1) and (Equation 2) is performed.

[0034]

(Equation 1)

[0035]

(Equation 2)

In (Equation 1), BitRate indicates a target average code rate, and FrameRate indicates the number of frames per second. N indicates the number of frames constituting one GOP. For example, in FIG. 2A, N = 12.
Therefore, G indicates the average allocated code amount per GOP. L in (Equation 2) is an integer of 1 or more, and R indicates the remaining amount of the virtual buffer.

Next, in step 302, it is determined whether or not the input frame is at the head of the GOP (at the start of encoding). For example, if the frame 201 is the head of the GOP, the process proceeds to step 303. In step 303, the virtual buffer remaining amount R is updated according to (Equation 3).

[0038]

(Equation 3)

In the equation (3), R on the right side indicates the virtual buffer remaining amount before updating, and R on the left side indicates the virtual buffer remaining amount after updating.

At step 304, the setting of the frame target generated code amount Tg for each GOP is set by (Equation 4), (Equation 5),
This is performed by (Equation 6).

[0041]

(Equation 4)

[0042]

(Equation 5)

[0043]

(Equation 6)

Np and Nb indicate the number of frames of P and B pictures included in one GOP, for example, as shown in FIG.
Then, it becomes 3,8. Therefore, Xg indicates the frame complexity in GOP units, and AvgXg indicates the average frame complexity in GOP units.

In step 305, the frame target generated code amount T for the frame 201 is set by using (Xi) (Equation 7).

[0046]

(Equation 7)

In step 306, the frame 201 is encoded. At this time, the MPEG TEST is performed such that the code amount of the frame 201 becomes the frame target generated code amount T for the frame 201 set in step 305.
The frame 201 is encoded using the method of MODEL3. This encoding is performed by MPEG TEST MO
A method other than DEL3 may be used.

When the encoding of the frame 201 is completed, post-encoding processing is performed in step 307. Here, the post-coding process refers to updating the remaining virtual buffer R, updating the average frame complexity AvgXi, and the frame complexity Xi. The update of the virtual buffer remaining amount R is performed according to (Equation 8).

[0049]

(Equation 8)

S in (Equation 8) indicates the generated code amount of the frame 201. Further, R on the right side of (Equation 8) indicates the remaining amount of the virtual buffer before the update, and R on the left side indicates the remaining amount of the virtual buffer after the update. Average frame complexity Avg
Xi is updated by (Equation 9), (Equation 10), and (Equation 11).

[0051]

(Equation 9)

[0052]

(Equation 10)

[0053]

[Equation 11]

SumXi in (Equation 9) indicates the cumulative value of the frame complexity Xi, and is 0 at the start of the encoding of the frame 201. Fi in (Equation 10) indicates the number of I pictures for which encoding has been completed. Frame complexity X
The updating of i is performed according to (Equation 12).

[0055]

(Equation 12)

Q in (Equation 12) indicates the average value of the quantization width for all macroblocks in the I picture frame.

The operation of the code amount controller 106 when encoding the frame 201 (I picture) has been described above.

Next, the details of the operation of the code amount controller 106 when encoding the frame 204 (P picture) which is the next frame of the frame 201 will be described below with reference to the flowchart of FIG. First, the code amount controller 106 performs steps 301 and 302 in the same manner as in the case of the frame 201 (I picture) described above. Since the frame 204 is not the first frame of the GOP, step 304 is skipped and step 304 is performed. In step 304, after setting the frame target generated code amount Tg for each GOP, in step 305, the frame target generated code amount T for the frame 204 is set to X
This is performed according to (Equation 13) using p. Frame 20 described above
In the case of 1 (I picture), the frame target generated code amount T is set using Xi by (Equation 7).

[0059]

(Equation 13)

In step 306, the frame 204 is encoded. At this time, the MPEG TEST is performed such that the code amount of the frame 204 becomes the frame target generated code amount T for the frame 204 set in step 305.
The frame 204 is encoded using the method of MODEL3. This encoding is performed by MPEG TEST MO
A method other than DEL3 may be used.

When the encoding of the frame 201 is completed, post-encoding processing is performed in step 307. The post-coding process here refers to updating the remaining virtual buffer R, updating the average frame complexity AvgXp, and the frame complexity Xp.

The update of the virtual buffer remaining amount R is performed according to (Equation 8).

[0063]

(Equation 8)

The update of the average frame complexity AvgXp is performed according to (Equation 14) to (Equation 16). Frame 2 above
In the case of 01 (I picture), the average frame complexity AvgXi was updated by (Equation 9) to (Equation 11).

[0065]

[Equation 14]

[0066]

(Equation 15)

[0067]

(Equation 16)

SumXp in (Equation 14) indicates the accumulated value of the frame complexity Xp, and is 0 at the start of encoding of the frame 204. Fp in (Equation 15) is the number of P pictures for which encoding has been completed. The update of the frame complexity Xp is performed by (Equation 17). Frame 2 above
In the case of 01 (I picture), the frame complexity Xi was updated by (Equation 12).

[0069]

[Equation 17]

Q in (Equation 17) is the average quantization width for all macroblocks in the P picture frame.

The operation of the code amount controller 106 when encoding the frame 204 (P picture) has been described above.

Next, details of the operation of the code amount controller 106 at the time of encoding the frame 202 (B picture) subsequent to the frame 204 will be described with reference to the flowchart of FIG. First, the code amount controller 106 performs steps 301 and 302 as in the case of the frame 201 (I picture) and the frame 204 (P picture). Since the frame 202 is not the first frame of the GOP, step 304 is skipped and step 304 is performed. In step 304, 1G
After setting the frame target generated code amount Tg for each OP, in step 305, the frame target generated code amount T for the frame 202 is set. The setting of the frame target generated code amount T for the frame 202 is performed by (Equation 18) using Xb. For the frame 201 (I picture) described above, Xi is used by (Equation 7), and for the frame 204 (P picture), Xp is used (Equation 13).
Thus, the frame target generated code amount T was set.

[0073]

(Equation 18)

At step 306, the frame 202 is encoded. At this time, the MPEG TEST is performed such that the code amount of the frame 202 becomes the frame target generated code amount T for the frame 202 set in step 305.
The frame 202 is encoded using the method of MODEL3. This encoding is performed by MPEG TEST MO
A method other than DEL3 may be used.

When the encoding of the frame 202 is completed, post-encoding processing is performed in step 307. The post-coding process here refers to updating the remaining virtual buffer R, updating the average frame complexity AvgXb, and the frame complexity Xb. The update of the virtual buffer remaining amount R is performed according to (Equation 8).

[0076]

(Equation 8)

The update of the average frame complexity AvgXb is performed according to (Equation 19) to (Equation 21). Frame 2 mentioned above
For 01 (I picture), (Equation 9) to (Equation 11), and for frame 204 (P picture), (Equation 14)
From (Equation 16), the average frame complexity AvgXi,
AvgXp was updated.

[0078]

[Equation 19]

[0079]

(Equation 20)

[0080]

(Equation 21)

SumXp in (Equation 19) indicates the cumulative value of the frame complexity Xb, and is 0 at the start of encoding of the frame 202. Fb in (Equation 20) is the number of B pictures for which encoding has been completed. The update of the frame complexity Xb is performed by (Equation 17). In addition, the above-mentioned frame 2
The frame complexity Xi and Xb are updated according to (Equation 12) for 01 (I picture) and (Equation 17) for frame 204 (P picture).

[0082]

(Equation 22)

Q in (Equation 22) is an average quantization width for all macroblocks in a B picture frame.

The operation of the code amount controller 106 when encoding the frame 202 (B picture) has been described above. FIGS. 4A and 4B are diagrams showing an example of a temporal change in the generated code amount when the code amount controller 106 operates using the above-described variable bit rate coding method. . The solid line in FIG. 4A indicates the frame complexity Xg, and the broken line indicates 1.
5 shows the average frame complexity AvgXg. In FIG. 4B, the solid line indicates the generated code amount in one GOP unit, and the broken line indicates the target generated code amount in one GOP unit. As described above, the temporal change in the frame complexity in 1 GOP unit and the generated code amount in 1 GOP unit draws a similar curve. That is, this graph shows a moving image with an appropriate target generated code amount according to the frame complexity. This indicates that the image is being encoded.

As described above, in the moving picture coding apparatus according to the first embodiment, the ratio of the frame complexity to be currently coded to the past average frame complexity is obtained, and the target generation of the current frame is determined based on the ratio. Since the code amount is obtained, it is possible to encode the current frame with an appropriate target generated code amount according to the frame complexity, and it is also possible to satisfy the target generated code amount in GOP units. Becomes Therefore, the moving picture coding apparatus according to the first embodiment uses a preset 1 GOP
Even when the target generated code amount per unit is low, the effect of improving the image quality can be obtained as compared with the conventional fixed rate coding method.

Embodiment 2 Next, a moving picture coding apparatus according to a second embodiment of the present invention will be described with reference to the drawings. Code amount controller 10 according to Embodiment 2
Reference numeral 6 denotes a unit integrating the code amount counting unit, the average value calculating unit, the complexity calculating unit, the virtual buffer unit, and the target quantization width determining unit. The code amount controller 106 according to the first embodiment does not include the target quantization width determining unit, but includes the target code amount determining unit instead. FIG. 5 is a flowchart of the operation of the code amount controller 106 according to the second embodiment. Comparing FIG. 5 with FIG. 3 according to the first embodiment, the calculation of the frame target quantization width is performed between the step of calculating the target target code amount (step 305) and the step of encoding (step 502). 501 is different.
Thus, the code amount controller 106 according to the second embodiment
The operations in steps 301 to 305 and step 307 are the same as those in the first embodiment, and a description thereof will be omitted. The operations in steps 305 and 502 will be described below.

In step 305, the frame target generated code amount T is set according to (Equation 7).
The frame target quantization width Qi is set according to 3). However, (Equation 23) is used when the frame currently to be encoded (hereinafter referred to as frame M) is an I picture, and when the frame M is a P or B picture, (Equation 24) , (Equation 25), the frame target quantization widths Qp and Qb are set.

[0088]

(Equation 23)

[0089]

(Equation 24)

[0090]

(Equation 25)

The values of the frame target quantization widths Qi, Qp, and Qb obtained by (Equation 23), (Equation 24), and (Equation 25) are directly used as the quantization widths for all the macroblocks in the frame M. , Frame M is quantized.

Thereafter, the frame M in step 502
And the operation of the post-coding process in step 307 are the same as in the first embodiment.

When the frame M is quantized, the values of the frame target quantization widths Qi, Qp, and Qb are used as they are as the quantization widths for all the macroblocks in the frame M. The average value of the quantization width for each macroblock in the frame is the frame target quantization width Q
If the values are close to i, Qp, and Qb, the value of the quantization width for each macroblock may be changed.
For example, as shown in FIG. 8, an activity calculator 801 is provided in the video encoding device 100, and the quantization width for each macro block is changed according to the activity of each macro block calculated by the activity calculator 801. It doesn't matter.

When quantizing the frame M, the values of the frame target quantization widths Qi, Qp, and Qb are used as they are as the quantization width values for all the macroblocks in the frame M. The quantization widths Qi, Qp,
The value of Qb may be weighted according to the picture type of the frame M. For example, if the frame M is an I or P picture, the frame M is quantized using the frame target quantization widths Qi and Qp as it is, and if the frame M is a B picture, the frame target quantization obtained by (Equation 25) is used. The width Qb is set to 1.
The quantization of the frame M may be performed using the quadrupled value as the frame target quantization width Qb.

As described above, in the moving picture coding apparatus according to the second embodiment, the ratio of the complexity of the current frame to be coded to the average of the past frame complexity is obtained, and the target of the current frame is calculated based on the ratio. Since the quantization width is obtained and the current frame is quantized using the target quantization width, it is possible to encode the current frame with an appropriate target generated code amount according to the frame complexity. It becomes possible, and it is also possible to satisfy the target generated code amount of one GOP unit. Therefore, the second embodiment
Is a moving image encoding apparatus that has previously set 1GO.
Even when the target generated code amount in P units is low, the effect of improving the image quality can be obtained as compared with the conventional fixed rate coding method.

Embodiment 3 Next, a video encoding device according to Embodiment 3 of the present invention will be described with reference to the drawings. FIG. 6 is a block diagram of a video encoding device according to the third embodiment. Thus, the third embodiment
Is a video encoding device 603,
604 in combination with the moving picture coding device 6
Numeral 03 denotes a configuration in which the code amount controller 106 and the buffer 108 are removed from the moving picture coding apparatus 100 according to the first embodiment, and a complexity calculator 602 is added at the subsequent stage of the variable length coding unit 107a. The encoding device 604 is the same device as the video encoding device 100. That is, the moving picture coding apparatus according to the third embodiment is a modification of the moving picture coding apparatus according to the first embodiment.

The moving picture coding device 603 is a device for temporarily coding a frame to be currently coded (for example, a frame M) in order to obtain the frame complexity. The encoding method by the device 603 is almost the same as in the first embodiment. The difference is that code amount control is not performed, and the quantization width used in the quantizer 105a is constant. The complexity calculator 602 counts a code string output from the variable-length encoder 107a, and obtains a code amount for each frame. Then, the frame complexity of the frame M is obtained by multiplying the frame code amount by the quantization width. The obtained frame complexity is determined by the code amount controller 106 of the moving image coding apparatus 604.
b.

The moving image coding device 604 is a device that performs actual coding of the frame M to generate a code string. By the frame memory 601, the input of the frame M to the video encoding device 604 is input at least one frame later than the input to the video encoding device 603. The code amount controller 106b includes a moving image coding device 603.
, A target generated code amount for the frame M is obtained based on the frame complexity of the frame M output from, and the target generated code amount is output to the quantizer 105b. Then, the video encoding device 604 encodes the frame M. The encoding method by the moving image encoding device 604 is described in Embodiment 1.
Is almost the same as that of the above, but the value used for the frame complexity is different. That is, in the first embodiment, the frame complexity Xi, Xp, Xb is the complexity of the frame of the same picture type existing immediately before the frame M, but in the third embodiment, the frame complexity of the current frame M The frame complexity is calculated by the complexity calculator 602.
It was obtained in. Therefore, for example, when encoding the frame 222, Xp is the frame complexity of the frame 222, and AvgXp is the average value of the frame complexity of the P pictures before the frame 219 (including the frame 219). Also, in step 307, (Equation 12)
It is not necessary to calculate the frame complexity by using.

Note that the moving picture coding apparatus 604 may use the moving picture coding apparatus 100 according to the second embodiment. In this case, the code amount controller 106b obtains a target generated code amount for the frame M based on the frame complexity output from the video encoding device 603, obtains a target quantization width from the target generated code amount, and performs quantization. Output to the converter 105b. The quantizer 105b quantizes the frame M using the target quantization width output from the code amount controller 106b. After that, the video encoding device 604 encodes the frame M.

As described above, in the moving picture coding apparatus according to the third embodiment, the ratio of the frame complexity to be coded to the average frame complexity in the past is determined, and the target generation of the current frame is determined based on the ratio. Since the code amount is obtained, it is possible to encode the current frame with an appropriate target generated code amount according to the frame complexity, and it is also possible to satisfy the target generated code amount in GOP units. Becomes Therefore, the moving picture coding apparatus according to the third embodiment uses the preset 1 GOP
Even when the target generated code amount per unit is low, the effect of improving the image quality can be obtained as compared with the conventional fixed rate coding method.

Although the encoding method according to the first to third embodiments has been described based on the MPEG2 method, this method controls the generated code amount by the quantization width (H.261).
System, Motion JPEG system, etc.).

Although the unit of a moving image in the first to third embodiments has been described as a frame, it may be a field.

In the equations (Equation 6) in the first to third embodiments, the ratio between the GOP-converted complexity Xg and the GOP-converted average complexity AvgXg is directly assigned to the code amount Tg.
, But the allocated code amount Tg is Xg
If / AvgXg is a simple increasing function,
Any reflection method may be used. For example, (Equation 6)
Equation (26) can be used instead of

[0104]

(Equation 26)

In the first to third embodiments, the average frame complexity AvgXi, AvgXp, AvgXb
Is the average value from the beginning of the sequence, but this may be the average value of frames for several GOPs immediately before the frame currently being encoded. For example, if the frame currently being encoded is an I picture, and the average value of the frame complexity of the G frames immediately before it is used as the average frame complexity, step 307
(Equation 27) as shown in (Equation 27).

[0106]

[Equation 27]

In Equation (27), AvgXi on the right side is a value before update, and AvgXi on the left side is a value after update.

In the first to third embodiments, the encoding of the first frame of each picture type is performed in step 30.
Xi, Xp, Xb, AvgXi, AvgX set in 1
The frame target generated code amount or the frame target quantization width is obtained from the initial values of p and AvgXb, and the coding is performed using the frame target generated code amount or the frame target quantization width. However, the coding is performed using the fixed quantization width. It may be.

In the first to third embodiments, the input frame is encoded using the value of the frame target generated code amount T as it is.
May be limited to a maximum value and a minimum value. In this case, the flowcharts shown in FIGS. 3 and 4 are as shown in FIGS. 7 (a) and 7 (b), respectively. The process of step 701 of setting the maximum and minimum values in FIG. 7 is performed according to (Equation 28).

[0110]

[Equation 28]

Tmax and Tmin in (Equation 28)
Are the maximum and minimum values for the frame target generated code amount T, respectively. This restriction may be applied to the frame target generated code amount Tg in GOP units.

[0112]

As described above, in the moving picture coding apparatus according to the present invention, when coding a moving picture, the value obtained by multiplying the generated code amount in the screen by the average value of the quantization width in the screen is used. As the complexity, and based on the ratio of the current screen complexity to the average of the past screen complexity, the target generated code amount or the target quantization width of the screen is determined. The target generated code amount or the target quantization width can be allocated to each screen, and the target generated code amount in one GOP can be satisfied. Therefore, the moving picture coding apparatus according to the present invention has an effect that the picture quality is dramatically improved as compared with the conventional fixed rate coding method even when the preset target generated code amount is low. can get.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a video encoding device according to Embodiment 1 of the present invention.

FIG. 2 is a diagram for explaining frames, macro blocks, and blocks constituting a moving image input to the moving image encoding device of the present invention.

FIG. 3 is a flowchart showing an operation of the code amount controller according to the first embodiment.

FIG. 4 is a diagram illustrating an example of a temporal transition of a quantization width and a generated code amount according to the first embodiment.

FIG. 5 is a flowchart showing an operation of a code amount controller according to the second embodiment.

FIG. 6 is a block diagram of a video encoding device according to a third embodiment.

FIG. 7 is a flowchart illustrating an example of an operation of a code amount controller according to the second embodiment.

FIG. 8 is a block diagram illustrating an example of a configuration of a code amount controller according to the second embodiment.

[Explanation of symbols]

100, 603, 604 Video encoding device, 101
Block converter, 102 differentiator, 103, 111 switch, 104 orthogonal transformer, 105 quantizer, 10
6 Code amount controller, 107 Variable length encoder, 108
Buffer, 109 inverse quantizer, 110 inverse orthogonal transformer, 112 motion compensation predictor, 113,601 frame memory, 114 adder, 201-227 frames, 602 complexity calculator, 801 activity calculator.

Claims (19)

[Claims] 1. A block conversion unit that divides each screen of a moving image composed of a plurality of screens into a plurality of blocks, and converts a plurality of blocks of each screen divided by the block conversion unit into coefficients. Image conversion means, quantization means for quantizing coefficients of a plurality of blocks of each screen converted by the image conversion means using a quantization width, and each screen quantized by the quantization means Code string generating means for generating a code string for each screen from the coefficients of the plurality of blocks; and counting the amount of code generated in each screen from the code string for each screen generated by the code string generating means. Code amount counting means for calculating, an average value calculating means for calculating an average value of the quantization width used by the quantization means, and a generated code amount counted by the code amount counting means and the average value calculating means. A complexity calculating means for calculating a value obtained by multiplying the average value of the calculated quantization width as the screen complexity of each screen, and an average screen complexity of a past screen calculated by the complexity calculating means. Calculating the ratio of the screen complexity of the current screen calculated by the complexity calculating means to the average screen complexity of the past screen, and determining the target generated code amount of the next screen from the ratio; A moving image encoding apparatus comprising: a target code amount determining unit that obtains a quantization width of a next screen using a generated code amount and outputs the quantization width to the quantization unit. 2. The moving picture coding apparatus according to claim 1, wherein said target code amount determining means includes a picture type (I) of a screen.
Picture, P picture, B picture), the average screen complexity of the second past screen per unit time is determined from the average screen complexity of the first past screen,
The screen complexity of the second current screen per unit time is obtained from the screen complexity of the current screen and the screen complexity of two screens other than the current screen of the first past in the second past. A target generated code amount for the next screen is determined from the ratio of the screen complexity of the second current screen to the average screen complexity of the past screen, and the quantization width of the next screen is determined using the target generated code amount. And outputting the quantization width to the quantization means. 3. The moving picture coding apparatus according to claim 1, wherein said moving picture coding apparatus assigns a code that can be assigned from a predetermined target average coding rate and a generated code amount counted by said code amount counting means. Virtual buffer means for obtaining the remaining amount and holding the allocatable code remaining amount, wherein the target code amount determining means includes a picture type (I
Picture, P picture, B picture), the average screen complexity of the second past screen per unit time is determined from the average screen complexity of the first past screen,
The screen complexity of the second current screen per unit time is obtained from the screen complexity of the current screen and the screen complexity of two screens other than the current screen of the first past in the second past. From the ratio of the ratio of the screen complexity of the second current screen to the average screen complexity of the past screen and the allocatable code remaining per unit time held in the virtual buffer, the target generated code of the next screen A moving picture coding apparatus characterized in that the amount is determined, the quantization width of the next screen is determined using the target generated code amount, and the quantization width is output to the quantization means. 4. A block converting means for dividing each screen of a moving image composed of a plurality of screens into a plurality of blocks, and converting the plurality of blocks of each screen divided by the block converting means into coefficients. Image conversion means, quantization means for quantizing coefficients of a plurality of blocks of each screen converted by the image conversion means using a quantization width, and each screen quantized by the quantization means Code string generating means for generating a code string for each screen from the coefficients of the plurality of blocks; and counting the amount of code generated in each screen from the code string for each screen generated by the code string generating means. Code amount counting means for calculating, an average value calculating means for calculating an average value of the quantization width used by the quantization means, and a generated code amount counted by the code amount counting means and the average value calculating means. A complexity calculating means for calculating a value obtained by multiplying the average value of the calculated quantization width as the screen complexity of each screen, and an average screen complexity of a past screen calculated by the complexity calculating means. Determining the ratio of the screen complexity of the current screen calculated by the complexity calculating means to the average screen complexity of the past screen, and determining the target quantization width of the next screen from the ratio; A moving image encoding apparatus, comprising: a target quantization width determining unit that outputs a quantization width to the quantization unit. 5. The moving picture coding apparatus according to claim 4, wherein said target quantization width determining means determines a first past picture for each picture type (I picture, P picture, B picture) of the picture. The average screen complexity of the second past screen per unit time is calculated from the average screen complexity of the first screen, and the screen complexity of the first current screen and the two screens other than the current screen of the first past are obtained. The screen complexity of the second current screen per unit time is obtained from the screen complexity, and the following is calculated from the ratio of the screen complexity of the second current screen to the average screen complexity of the second past screen. A moving image encoding apparatus for determining a target quantization width of the screen of (1) and outputting the target quantization width to the quantization means. 6. The moving picture coding apparatus according to claim 4, wherein said moving picture coding apparatus assigns a code that can be assigned from a predetermined target average coding rate and a generated code amount counted by said code amount counting means. Virtual buffer means for obtaining the remaining amount and holding the allocatable code remaining amount, wherein the target quantization width determining means is provided for each picture type (I picture, P picture, B picture) of the screen. Calculating the average screen complexity of the second past screen per unit time from the average screen complexity of the first past screen, and calculating the screen complexity of the first current screen and the current past of the first past. The screen complexity of the second current screen per unit time is obtained from the screen complexity of two screens other than the screen, and the second current screen of the second current screen is compared with the average screen complexity of the second past screen. Screen Determining the target quantization width for the next screen from the ratio of the degree of the degree and the remaining amount of assignable codes per unit time held in the virtual buffer, and outputting the target quantization width to the quantization means. A moving picture coding apparatus, characterized in that: 7. The moving picture coding apparatus according to claim 1, wherein the target generated code amount determined by said target code amount determining means is a target generated code amount for a current screen. A video encoding device characterized by the above-mentioned. 8. The video encoding apparatus according to claim 4, wherein the target quantization width determined by said target quantization width determination means is set as a target quantization width for a current screen. A moving picture coding apparatus characterized by the above-mentioned. 9. The moving picture coding apparatus according to claim 1, wherein an upper limit value and a lower limit value are set for the target generated code amount determined by the target code amount determining means. A video encoding device characterized by the above-mentioned. 10. The moving picture coding apparatus according to claim 4, wherein an upper limit value and a lower limit value are set for the target quantization width determined by said target quantization width determination means. A moving picture coding apparatus characterized by the above-mentioned. 11. The moving picture coding apparatus according to claim 4, wherein each of the picture types of the screen (I picture, P picture, B picture) is determined by said target quantization width determining means. A moving picture coding apparatus characterized in that the target quantization width is multiplied by a predetermined coefficient, and the value is used as the target quantization width. 12. The moving picture coding apparatus according to claim 1, wherein the number of past screens for obtaining the average value of the past screen complexity is a picture type of a screen (I picture, I picture, A moving picture encoding apparatus having a constant value for each of a P picture and a B picture. 13. The video encoding apparatus according to claim 4, wherein the number of past screens for obtaining the average value of the past screen complexity is a picture type (I picture, A moving picture encoding apparatus having a constant value for each of a P picture and a B picture. 14. The moving picture encoding apparatus according to claim 1, wherein the current generated picture amount is determined by the target generated code amount determining means and an average value of the past screen complexity is used to calculate the current screen. The value K obtained by dividing the complexity satisfies T = f (K) for a function f (K) that satisfies f (K1) ≦ f (K2) when K1 <K2. Moving image encoding apparatus. 15. The moving picture coding apparatus according to claim 2, wherein the fourth screen complexity is calculated based on the target generated code amount T determined by the target code amount determining means and the third screen complexity. Is divided by f (K) when K1 <K2.
1) For a function f (K) that satisfies ≦ f (K2), T = f
(K) A moving picture coding apparatus characterized by satisfying (K). 16. The moving picture coding apparatus according to claim 3, wherein the fourth screen complexity is calculated based on the target generated code amount T determined by the target code amount determining means and the third screen complexity. Is divided by the remaining code amount R that can be allocated per unit time, and g (K1, R) ≦ when K1 <K2.
g (K2, R), and when R1 <R2, g (K,
R2) ≦ g (K, R2) A function g (K, R) that satisfies T = g (K, R) with respect to a function g (K, R). 17. The moving picture coding apparatus according to claim 4, wherein the target quantization width Q determined by the target quantization width determination means is determined based on the average value of the past screen complexity. For a value K obtained by dividing the complexity and a function f satisfying f (K1) ≧ f (K2) when K1 <K2, Q =
A moving picture coding apparatus that satisfies f (K). 18. The moving picture coding apparatus according to claim 5, wherein the fourth screen complexity is obtained based on the target quantization width Q determined by the target quantization width determining means and the third screen complexity. Is divided by f (K1) when K1 <K2.
For a function f (K) that satisfies ≧ f (K2), Q = f
(K) A moving picture coding apparatus characterized by satisfying (K). 19. The video encoding apparatus according to claim 6, wherein the fourth screen complexity is determined based on the target quantization width Q determined by the target quantization width determination means and the third screen complexity. The value K obtained by dividing the degree is the above-mentioned allocatable code remaining amount R per unit time and g (K1, K1) when K1 <K2.
R) ≧ g (K2, R), and when R1 <R2, g
For a function g such that (K, R2) ≧ g (K, R2),
A moving picture coding apparatus, wherein S = g (K, R) is satisfied.