CN100477794C - Moving picture coding device, coding method thereof, and moving picture imaging device - Google Patents

Abstract

When frame n+1 will be inter-frame encoded after frame n has been inter-frame encoded (step S141: inter-frame), the moving image encoding device uses the first method to determine the amount of code used to quantize frame n based on the code amount generated by frame n. +1 quantization step size Q[n+1] (step S143). When frame n+1 will be inter-frame encoded after frame n has been intra-frame encoded (step S142: intra-frame), the moving image encoding device uses the second method to determine the amount of code used to quantize frame n+1 based on the code amount generated by frame n. The quantization step size Q[n+1] (step S144). The value obtained by using the quantization step size determined by using the second method is smaller than the value obtained by using the quantization step size determined by using the first method.

Description

Moving picture coding device, coding method thereof, and moving picture imaging device

technical field

The present invention relates to a moving picture coding device, a moving picture coding method, and a moving picture imaging device for coding a moving picture. In particular, the present invention relates to a technique for determining a quantization step size used in quantizing moving image data.

Background technique

Generally, a moving image imaging device is equipped with a moving image encoding device for compressing moving image data that the moving image imaging device has. This video encoding device compresses data by quantization. Quantization is a method of compressing data by replacing data values with discrete representative values. The interval between representative values is defined by a parameter called the quantization step size. As the quantization step size increases, the compression ratio becomes higher (ie, the code size becomes smaller).

According to an embodiment of a conventional model of a moving picture coding apparatus, a fixed value is used as a quantization step size for intra-coded frames (so-called I-pictures) to determine frames for each inter-frame coding (so-called P-pictures or B-pictures) different quantization step sizes (see for example ISO/IEC-14496-2). In order to determine the quantization step size of each frame used for inter-coding, the following formula 1 is used.

Qp[n+1]=(Cn/dstC)*Qp[n] Formula 1

In this formula, Qp[n] is the quantization step size for quantization of the n-th frame, Cn is the generated code amount of the n-th frame, and dstC is the target amount of codes scheduled to be generated for one frame.

According to Formula 1, the quantization step size Qn[n+1] for the n+1th frame becomes larger as the ratio between the coded amount Cn generated in the nth frame and the target coded amount dstC becomes larger. Equation 1 works in such a way as to suppress the amount of encoding generated at frame n+1. By determining the quantization step size for each frame in the above-described manner, the moving image encoding device makes the amount of code produced by the inter-coded frame closer to the target code amount produced.

Because the amount of code generated by intra-coded frames largely affects image quality, the quantization step size for frames to be intra-coded is usually determined so that the code amount generated by these frames is greater than the target code amount. In this case, the above-described moving image encoding device tries to suppress the amount of encoding generated in frame 2 immediately following the inter-coded frame (frame 1). Therefore, the amount of encoding generated by frame 2 is smaller than the target encoding amount. Then, the video encoding device makes the amount of encoding generated in frame 3 larger than the target amount of encoding. As described above, with the conventional moving image encoding apparatus, there is hunting in the amount of encoding generated in several frames following an intra-encoded frame.

FIG. 1A and FIG. 1B are graphs showing changes in the amount of coding and image quality for each frame in a conventional moving picture coding apparatus.

In this embodiment, frames 1, 11 and 21 are intra-coded frames, and the other frames are inter-coded frames. In the figure, fluctuations in the amount of encoding generated as described above can be observed in the four frames immediately following the intra-coded frame (FIG. 1A).

In general, image quality gradually improves as the amount of codes produced increases. Therefore, fluctuations in the generated encoding amount destabilize image quality (FIG. 1B). Therefore, the image becomes difficult to view.

This problem is not limited to the model in which the quantization step size is determined for each frame, but also exists in the model in which the quantization step size is determined in each block line (line of a macroblock located in the image matrix).

Contents of the invention

An object of the present invention is to provide a moving picture coding device, a moving picture coding method, and a moving picture imaging device capable of suppressing fluctuations in the amount of coding generated in several frames immediately following an intra-coded frame.

The moving picture coding device according to the present invention includes a coding unit for selectively performing inter-frame coding or intra-frame coding on each frame data of a plurality of frames of a moving picture, and the intra-frame coding and inter-frame coding both include quantization; The determining unit is based on determining the quantization step size used for quantizing the data of the second frame according to the amount of code generated in the first frame, wherein when the encoding unit performs inter-frame encoding on the data of the second frame, if the data of the first frame has been frame-encoded Inter-coding, determined by (i) the first method, or (ii) the second method if the data of the first frame has been intra-coded, by using the value obtained by the quantization step size determined by the second method is smaller than the value obtained by using the quantization step size determined by the first method.

According to the above structure, when the first frame is intra-frame encoded, the second method is used to determine the quantization step size of the second frame. Therefore, the quantization step size of the second frame is smaller than that of the first method.

In the conventional technique, the quantization step size of the second frame is determined by the same method (ie, the first method), regardless of the type of predictive coding performed in the first frame.

On the other hand, in the present invention, the quantization step size of the second frame becomes smaller than that of the conventional art, and thus the sudden decrease in the amount of encoding generated in the second frame can be reduced. Therefore, it is possible to suppress fluctuations in the amount of encoding generated in several frames following the intra-coded frame.

The above moving image coding apparatus may also cause the second method to be set in such a manner that the quantization step size determined by using the second method falls within the range of 1/5 to 1/3 of the quantization step size determined by using the first method.

Typically, if the same quantization step size is used, the amount of codes produced in an intra-coded frame is three to five times that of an inter-coded frame. According to the above structure, the quantization step size of the second frame is roughly the same regardless of whether the first frame is intra-frame encoded or inter-frame encoded. Specifically, even when the first frame is intra-frame encoded, the quantization step size should be determined as if the first frame is inter-frame encoded. Typically, the portion of the frame coded between frames does not fluctuate very often. Therefore, by applying the above structure, it is possible to suppress fluctuations in the amount of encoding generated in the frames after the second frame.

The above-mentioned moving image coding apparatus may further include a method for calculating (i) the cumulative code amount generated in the first frame and frames preceding the first frame and (ii) the first frame and frames preceding the first frame A difference obtaining unit that obtains a difference between cumulative target amounts of predetermined encoding, wherein the first method and the second method are set in such a manner that the determined quantization step size increases as the difference of generated encoding amounts becomes larger.

According to the above structure, as the difference between the accumulatively generated code amount and the cumulative target code amount becomes larger, the quantization step size of the second frame also becomes larger. Specifically, it is made to function to suppress the amount of encoding generated in the second frame. Therefore, the accumulatively generated code amount changes to roughly correspond to the cumulative target code amount. Therefore, the generated code amount can be made close to the target code amount for the entire moving image.

The above-mentioned moving image coding apparatus may set the first and second methods in such a manner that the quantization step size becomes larger as the generated code amount becomes larger when the cumulatively generated coded amount is smaller than the accumulated target coded amount, regardless of the difference. value.

In moving images, when the correlation between frames is high, the amount of codes generated in each frame decreases with respect to the target code amount. However, according to the above structure, even in this case, the generated encoding amount does not increase, and there may be an excess free space in the memory capacity. At this time, overflow of the memory capacity can be avoided even when the correlation between frames suddenly drops.

The moving image coding apparatus may further include a buffer for storing encoded data of moving images, wherein when the buffer has a free space larger than a predetermined size, the first and second methods follow the quantization step size Set in such a way that the amount of encoding becomes larger and larger, regardless of the amount of difference.

According to the above structure, when the free space is a predetermined size or more, the device operates to suppress the amount of encoding generated in the second frame. Therefore, overflow of the buffer can be avoided.

In the above moving image encoding device, when the encoding unit performs intra-frame encoding on the data of the second frame, the determination unit sets a fixed value as the quantization step size.

According to the above structure, it is not necessary to determine the quantization step size each time a frame is intra-coded. Therefore, the circuit size of the video encoding device can be reduced.

The above moving image encoding device may also be that when the encoding unit performs intra-frame encoding on the second frame data after encoding the data of the first frame, the determination unit determines the quantization step size by using the third method, and the quantization step size determined by the third method The value obtained by the quantization step size is larger than the value determined by the first method.

According to the above structure, the quantization step size of the second frame is determined by the third method. By adopting this method, the quantization step size of the second frame becomes larger compared with the case of using the first method. Therefore, the amount of codes generated in the second frame becomes smaller than when the first method is used, and the amount of codes generated in the second frame can be prevented from becoming excessively large.

The above-mentioned moving image encoding device may further include a buffer for storing encoded data of the moving image, wherein when the encoding unit intra-encodes the data of the second frame after encoding the first frame data, (i) if the buffer has greater than a predetermined size of the free space, the determination unit sets the quantization step to a fixed value, and (ii) if the free space in which the buffer exists is a predetermined size or less, the determination unit determines the quantization step by using a third method, by using The quantization step determined by the third method obtains a value larger than that determined by the first method.

According to the above structure, the amount of codes generated by the intra-coded frames can be adjusted according to the size of the buffer free space. Therefore, overflow of the buffer can be avoided.

The above moving image encoding device may be that when the encoding unit performs intra-coding on the second frame data after encoding the first frame data, the determining unit determines the quantization step size of the second frame based on the quantization step size used for quantization of the first frame, Regardless of the amount of code produced.

According to the above structure, the quantization step size is determined regardless of the amount of codes generated in the first frame. Therefore, compared with the case where the quantization step size is determined based on the amount of generated codes, it is possible to reduce the processing load of the video coding device.

The above moving image encoding device may also be such that when the encoding unit intra-encodes the second frame data after encoding the first frame data, the determination unit determines the quantization step size for the first frame by multiplying the predetermined value by the quantization step size for the first frame. The quantization step size of two frames.

According to the above structure, the quantization step size for the second frame is determined based on the quantization step size for the first frame.

In the above moving image encoding device, the predetermined value may also be in the range of 5/4 to 4/3.

According to the above structure, as the predetermined value becomes larger, the suppressing effect on the amount of codes generated by intra-coded frames is improved, but the image quality is deteriorated. The simulation results show that by setting the predetermined value in the range of 5/4 to 4/3, it is possible to obtain both the suppression effect of the code amount produced by the intra-coded frame and excellent image quality.

The above moving image coding apparatus may further include a recording unit for recording a table representing a correspondence relationship between the quantization step size of the first frame and the quantization step size of the second frame, wherein the determination unit refers to the table based on The quantization step size of determines the quantization step size used for the second frame.

According to the above structure, the quantization step size for the second frame can be determined based on the quantization step size for the first frame.

The above moving image encoding device may be such that the determining unit determines the quantization step for the second frame based on quantization steps used for a predetermined number of preceding frames in addition to the quantization step for the first frame.

By determining the quantization step size based on more than one frame, although the throughput of the moving image encoding device is increased, the effect of suppressing fluctuations can be improved.

In addition, the moving picture coding device according to the present invention includes: a coding unit for selectively performing inter-frame coding or intra-frame coding on each data of a plurality of block lines in the moving picture, and both the inter-frame coding and the intra-frame coding include quantization a correction unit, used for (i) correcting the coding amount generated in one or more inter-frame coding block lines in the predetermined number of coding block lines by using the first method, and (ii) using the second method to correct the predetermined amount Correct the code amount generated by one or more intra-frame coded block lines in the coded block line, and the value obtained by the coded amount generated by one or more intra-frame coded block lines corrected by the second method is smaller than that obtained by the first method The corrected value of the method; and the determination unit, when the coding unit is to inter-code the data of the next block line, it is used to determine the total amount of codes to be used based on the corrected intra-frame coding block line and the inter-frame coding block line. The quantization step size of the next line data quantization.

According to the above structure, the quantization step size for the next line to be encoded can be made smaller compared to the case where the amount of encoding produced by the encoding method is not corrected. Therefore, the amount of codes generated in the block lines when coding is not easily reduced. Therefore, it is possible to suppress fluctuations in the amount of encoding generated in several block lines immediately following the intra-encoded block.

The above-mentioned moving picture coding apparatus may also be: such that the code amount generated in one or more inter-frame coding block lines by using the second method falls within 1/5 to 1/3 of the code amount corrected by using the first method The method within the range is set to the second method.

In general, if the same quantization step size is used, the amount of code produced by intra-coding block lines is three to five times greater than that produced by inter-coded block lines. According to the above structure, even when the first block line is intra-frame encoded, the determination of the quantization step size of the second block line is handled as if the first frame was inter-frame encoded. Typically, the part of the inter-coded block line does not fluctuate very often. Therefore, by applying the above structure, it is possible to suppress fluctuations in the amount of encoding generated by the multi-block lines subsequent to the second block line.

A moving image imaging device according to the present invention includes an imaging device for capturing an image of a target to generate a moving image; an encoding unit for selectively performing inter-frame encoding or intra-frame encoding on each data in a plurality of frames of the moving image, the frame Both inter-coding and intra-frame coding include quantization; a determination unit for determining a quantization step size for quantization of second frame data according to the amount of coding generated in the first frame, wherein when the coding unit performs inter-frame encoding on the data of the second frame When encoding, either (i) make a decision by using the first method if the data of the first frame has been inter-coded, or (ii) make a decision by using the second method if the data of the first frame has been intra-coded, by using the second The quantization step size determined by the method results in a smaller value than the quantization step size determined by using the first method.

In addition, according to the present invention, the moving picture imaging device includes: an imaging device for capturing an object image to generate a moving picture; a coding unit for selectively performing inter-frame coding or frame encoding on each data in a plurality of block lines of the moving picture Intra-coding, wherein both inter-frame coding and intra-frame coding include quantization; a correction unit for (i) correcting the amount of code generated by one or more inter-coded block lines among the predetermined number of coded block lines using a first method , and (ii) using the second method to correct the amount of code generated in one or more intra-frame coding block lines among the predetermined number of coding block lines, by using the second method to correct the code amount generated in one or more intra-frame coding block lines The value obtained by the coding amount of the line is smaller than the coding amount corrected by the first method; and determining the unit length, when the coding unit is to inter-code the data of the next line, it is used to encode the block line based on the corrected intra-frame and the inter-frame The total amount of codes produced by an encoded block line determines the quantization step size used for quantization of the next block line data.

According to the above structure, the same effect as that of the above-mentioned moving picture encoding device can be obtained.

The moving image encoding method according to the present invention includes: an encoding step of selectively performing inter-frame encoding or intra-frame encoding on each data in multiple frames of the moving image, wherein both the inter-frame encoding and the intra-frame encoding include quantization; the determining step, The quantization step size of the data quantization of the second frame is determined according to the amount of coding generated in the first frame, wherein when the data of the second frame is inter-frame encoded, by (i) if the data of the first frame has been inter-frame encoded, adopt The decision is made by the first method, or (ii) if the data of the first frame has been intra-coded, the decision is made by the second method, wherein the value obtained by the quantization step determined by the second method is smaller than the value determined by the first method The value obtained by the quantization step size.

In addition, the moving image encoding method according to the present invention includes: an encoding step of selectively performing inter-frame encoding or intra-frame encoding on each data of the multi-block lines in the moving image, wherein both the inter-frame encoding and the intra-frame encoding include quantization; correcting Steps, (i) using the first method to correct the amount of code generated by one or more inter-coded block lines in the predetermined number of coded block lines, and (ii) using the second method to correct the code amount generated in the predetermined number of coded block lines The code amount of the one or more intra-coded block lines corrected by the second method is smaller than the code amount corrected by the first method; and The determining step is used to determine the quantization step size to be used for quantization of the next line data based on the corrected intra-coded block line and the inter-coded block line generated by the corrected coding amount when the data of the next block line is inter-coded.

According to the above structure, the same effect as that of the above-mentioned moving picture encoding device can be obtained.

Description of drawings

These and other objects, advantages and features of the present invention will become more apparent from the following description and accompanying drawings in conjunction with specific embodiments of the present invention.

In the attached picture:

FIG. 1A and FIG. 1B are schematic diagrams showing changes in the amount of coding and image quality generated in each frame in a conventional moving image coding device;

FIG. 2 shows the structure of the moving picture imaging device according to the first embodiment;

Fig. 3 shows the embodiment of the detailed structure of code correction unit 12;

FIG. 4 is a schematic diagram showing the operation of the moving picture encoding device 31 according to the first embodiment;

FIG. 5 is a detailed schematic diagram of the operation of determining the quantization step size according to the first embodiment;

6A to FIG. 6D are schematic diagrams illustrating changes in time sequence of the quantization step size, the amount of codes generated in each frame, the image quality, and the total amount of codes generated according to the first embodiment;

FIG. 7 is a detailed schematic diagram of the operation of determining the quantization step size according to the second embodiment;

8A to FIG. 8D are schematic diagrams showing the change in time sequence of the quantization step size, the amount of codes generated in each frame, the image quality, and the total amount of codes generated according to the second embodiment;

FIG. 9 shows the structure of a moving picture imaging device according to a third embodiment;

Fig. 10 shows the embodiment of the detailed structure of difference calculation unit 14 and target coding correction unit 15;

FIG. 11 shows the operation of the moving picture encoding device 33 according to the third embodiment;

FIG. 12 shows the detailed operation of correcting the target code amount according to the third embodiment;

13A to 13D are schematic diagrams showing the change in time sequence of the quantization step size, the amount of codes generated in each frame, the image quality, and the total amount of codes generated according to the third embodiment;

FIG. 14 shows the structure of a moving picture imaging device according to a fourth embodiment;

FIG. 15 is a detailed schematic diagram of the operation of determining the quantization step size according to the fourth embodiment;

FIG. 16 shows the structure of a moving picture imaging device according to a fifth embodiment;

FIG. 17 shows the operation of the moving picture encoding device 35 according to the fifth embodiment;

FIG. 18 is a detailed schematic diagram of determining the quantization step size according to the sixth embodiment;

Fig. 19A to Fig. 19C are schematic diagrams showing the change of the quantization step size, the coding amount generated in each frame and the image quality in time order according to the sixth embodiment;

FIG. 20 shows the structure of a moving picture imaging device according to a seventh embodiment;

Figure 21 shows the embodiment that is stored in the table of table storage unit 19;

Fig. 22 is a detailed schematic diagram of determining the quantization step size according to the seventh embodiment;

FIG. 23 shows the structure of a moving picture imaging device according to an eighth embodiment;

Fig. 24 shows the embodiment of the detailed structure of code correction unit 12;

FIG. 25 shows the operation of the moving picture encoding device 38 according to the eighth embodiment;

Fig. 26 is a detailed schematic diagram of determining the quantization step size according to the eighth embodiment;

Fig. 27 is a schematic diagram showing the change of the encoding amount in time order according to the eighth embodiment; and

FIG. 28 shows the structure of a moving picture imaging device according to a ninth embodiment.

Detailed ways

Preferred embodiments of the present invention are described below with reference to the drawings.

first embodiment

FIG. 2 shows the structure of the moving picture imaging device according to the first embodiment.

The moving image imaging device is provided with an imaging lens 1 , an imaging sensor 2 , a moving image encoding device 31 , a microcomputer 4 , a program memory 5 , and a video stream memory 6 . The imaging lens 1 generates an object image on the imaging sensor 2 . The imaging sensor 2 captures an image of an object to generate moving image data. The moving image encoding device 31 generates a video stream by compressing data encoding of moving images generated by the imaging sensor 2 . The microcomputer 4 generally controls the moving picture device according to programs stored in the program memory 5 . The video stream memory 6 stores video streams generated by the moving image encoding device 31 .

The moving image encoding device 31 includes an encoding unit 11 , an encoding correction unit 12 , and a quantization step size determination unit 13 .

The encoding unit 11 is a so-called MPEG encoder that selectively performs inter-encoding or intra-encoding on each frame data of a moving image. Examples of inter coding include inter predictive coding, discrete cosine transform (DCT), quantization, and variable length coding. Examples of intra coding include intra predictive coding, DCT, quantization, and variable length coding. A selection is made between inter-frame coding and intra-frame coding according to an instruction from the microcomputer 4 . In the first embodiment, a case where one frame is intra-coded among ten frames and the other nine frames are inter-coded is taken as an example.

The quantization step size used for frame quantization is provided by the quantization step size determination unit 13 .

The code correction unit 12 measures the code amount generated in each frame, and corrects the code amount according to the following formulas (Formula 2 and Formula 3) for correction.

When frame n is intra-coded:

Cn’＝Cn*P1 Formula 2

When frame n is inter-coded:

Cn’＝Cn*P2 Formula 3

In the above formula, Cn is the amount of code generated after encoding frame n, and P1 and P2 are correction coefficients with a relationship of P1<P2. In the first embodiment, the case where P1=P where (0<P<1) and P2=1 is taken as an example. The correction coefficient P satisfies the following formula 4.

P=Cinter/Cintra Formula 4

In this formula, Cinter is the code amount generated when inter-frame coding is performed on a specified frame using a specified quantization step size, and Cintra is the code amount generated when intra-frame coding is performed on the same frame using the same quantization step size. The correction coefficient P is used to convert the coding amount generated by intra-frame coding into the coding amount generated by inter-frame coding.

Generally speaking, the amount of codes generated by intra-frame coding for a certain frame is about three to five times that of the same frame by inter-frame coding. Therefore, the correction coefficient P in the first embodiment is set to a fixed value falling within the range of 1/5 to 1/3. Note that the encoding correction unit 12 specifies whether the frame is intra-encoded or inter-encoded according to an instruction from the microcomputer 4 .

The quantization step size determination unit 13 determines the quantization step size for the frame ((n+1) according to Formula 5 and Formula 6 below.

When frame (n+1) will be intra-coded:

Qp[n+1]=Qpintra Formula 5

When frame (n+1) will be inter-coded:

Qp[n+1]=(Cn’/dstC)*Qp[n] Formula 6

In the above formulas 5 and 6, Qpintra is a predetermined value, and dstC is a target encoding amount set in advance.

In the first embodiment, Qpintra is used as the quantization step size when frame (n+1) is intra-coded. When frame (n+1) is subjected to inter-frame coding, the quantization step size is determined based on Cn' and the quantization step size Qp[n], where Cn' is the code amount corrected by the code correction unit 12.

The encoding correcting unit 12 corrects the encoding amount of frame n according to formula 2 when frame n performs intra-frame encoding, and corrects the encoding amount of frame n according to formula 3 when frame n performs inter-frame encoding. An embodiment of a structure for realizing the above is described below.

FIG. 3 shows an example of the detailed structure of the code correction unit 12. As shown in FIG.

The code correction unit 12 includes a code measurement unit 121, a frame code memory 122, registers R1 and R2, a selector S1, and a multiplier M1.

The encoding measurement unit 121 measures the encoding amount of the frame n encoded by the encoding unit 11 . The register R1 holds the correction coefficient P, and the register R2 holds the correction coefficient 1. According to an instruction from the microcomputer 4, the selector S1 applies the correction coefficient P to the multiplier M1 when the frame n is intra-coded, and applies the correction coefficient 1 to the multiplier M1 when the frame n is inter-coded.

The multiplier M1 multiplies the generated code amount Cn by the selected correction coefficient. The frame code memory 122 stores the result of multiplication by the multiplier M1, and delivers the result to the quantization step size determining unit 13.

Through the above-mentioned structure, the coding correction unit 12 can output the corrected coding amount Cn'=Cn*P when the frame n is intra-frame encoded, and the coding correction unit 12 can output the corrected coding amount Cn' when the frame n is inter-frame coding =Cn.

FIG. 4 is a schematic diagram showing the operation of the moving image encoding device 31 according to the first embodiment.

The moving image encoding device 31 starts encoding in response to an instruction from the microcomputer 4 . First, the video coding device 31 intra-codes one frame (step S11). Because frame 1 has no previous frame, frame 1 is always intra-coded. As an initial value of the quantization step size, Qpintra to be used when performing intra coding is used. Then, the moving image coding device 31 stores the code amount generated by the frame 1 in the frame coding memory 122 (step S12). Because frame 1 is intra-coded, C1*P is stored in frame code memory 122 .

Next, the moving image coding device 31 determines the quantization step size for frame 2 based on the generated code amount and the quantization step size for frame 1 (step S14). Then, the moving picture encoding device 31 inter-codes frame 2 using the determined quantization step size (step S15).

The moving image encoding device 31 stores the encoding amount generated in the frame 2 in the frame encoding memory 122 (step S16). Since frame 2 is inter-coded, C2 is stored in frame code memory 122 . Thereafter, the moving image encoding device 31 repeats the operations from step S13 to step S17 until the microcomputer 4 instructs to stop imaging.

Fig. 5 is a detailed schematic diagram of the operation of determining the quantization step size according to the first embodiment.

When the microcomputer 4 instructs to inter-code frame (n+1) (step S141: inter) and frame n has been inter-coded (step S142: inter), the moving image coding device 31 determines the frame to be quantized ( The quantization step size Qp[n+1] of n+1) is (Cn/dstC)*Qp[n] (step S143). When frame n has been intra-coded (step S142: intra), the moving image coding device 31 determines the quantization step size Qp[n+1] for quantizing frame (n+1) to be (Cn*P/dstC)* Qp[n] (step S144).

On the other hand, when the microcomputer 4 instructs the intra-coding frame (n+1) (step S141: intra), the moving image coding device 31 determines the quantization step size Qp[n+ 1] is Qpintra (step S145).

[Effect]

6A to FIG. 6D are schematic diagrams illustrating changes in time sequence of the quantization step size, the amount of codes generated in each frame, the image quality, and the total amount of codes generated according to the first embodiment.

Frames 1, 11 and 21 are intra-coded frames, while the others are inter-coded frames.

Fig. 6A shows the quantization step size. In the first embodiment, the quantization step size for an intra-coded frame is set to a fixed value Qpintra (see step S145 in FIG. 5 ). Therefore, as shown in FIG. 6A, the quantization steps for frames 1, 11, and 21 take the same value. Usually, this fixed value is set so that the amount of codes generated in an intra-coded frame is larger than the target code amount for one frame. This is because the amount of codes generated in intra-coded frames greatly affects the image quality of the entire moving image, and the amount of codes generated in intra-coded frames can cause overall degradation of the image quality of moving images.

Fig. 6B shows the amount of encoding. The generated code amount C1 is much larger than the target code amount. The corrected code amount C1 *P shown superimposed on the generated code amount indicates the code amount corrected by the code correcting unit 12 . This corrected code amount C1*P is converted into a code amount resulting from inter-frame encoding, and has approximately the same size as that of an inter-coded frame such as frame 2 . Therefore, the quantization step size of frame 2 becomes smaller compared to the case where the amount of code generated in frame 1 is not corrected. Therefore, compared with the case where the quantization step size for a frame is determined based on the uncorrected code amount generated by frame 1, the code amount generated by frame 2 becomes larger, and the code amount generated by frame 2 is approximately equal to the frame target code amount Same. Since the code amount generated in frame 2 roughly matches the target code amount for one frame, the code amount generated in frames 3 to 10 also varies around the target code amount. In the above-described manner, it is possible to suppress fluctuations in the amount of codes generated in several inter-coded frames immediately following an intra-coded frame.

Figure 6C shows image quality. Since fluctuations in the amount of encoding generated in several inter-coded frames immediately following an intra-coded frame can be suppressed in the above-described manner, image quality becomes stable.

Fig. 6D shows the total code amount. In the first embodiment, the quantization step size for intra-coded frames is a fixed value Qpintra. The fixed value Qpintra is set so that the code amount generated by the intra-coded frame is greater than the target code amount. Therefore, the total encoding amount is larger than the cumulative target encoding amount.

second embodiment

In the first embodiment, a fixed value Qpintra is used as the quantization step size for intra-coded frames. However, also in the second embodiment, the quantization step size of the intra-coded frame is appropriately determined.

[structure]

The moving picture coding device according to the second embodiment has basically the same structure as the moving picture coding device according to the first embodiment. Therefore, description is also given with reference to FIG. 2 .

The encoding correcting unit 12 measures the amount of encoding generated in each frame, and then corrects the amount of encoding generated according to the following correction formulas (Equations 7, 8, and 9).

When frame (n+1) is to be inter-coded, and frame n is already intra-coded:

Cn’＝Cn*P1 Formula 7

When frame (n+1) is to be inter-coded, and frame n is already inter-coded:

Cn’＝Cn*P2 Formula 8

When frame (n+1) will be intra-coded:

Cn’＝Cn*P3 Formula 9

P1, P2 and P3 are correction coefficients with a relationship of P1<P2<P3. In the second embodiment, the case of P1=P, where (0<P<1), P2=1 and P3=1/P is taken as an example for description. This correction coefficient P satisfies the coefficient of Formula 4.

The quantization step size determination unit 13 determines the quantization step size for frame (n+1) according to Formula 10 below.

Qp[n+1]=(Cn’/dstC)*Qp[n] Formula 10

[operate]

Fig. 7 is a detailed schematic diagram of determining the quantization step size according to the second embodiment.

When the microcomputer 4 instructs to inter-code frame (n+1) (step S241: inter) and frame n has already been inter-coded (step S242: inter), the moving image coding device 31 determines to quantize the frame ( The quantization step size QP[n+1] of n+1) is (Cn/dstC)*Qp[n] (step S243). On the other hand, when frame n has been intra-coded (step S242: intra), the moving picture coding device 31 determines the quantization step size Qp[n+1] for quantizing frame (n+1) as (Cn *P/dstC)*Qp[n] (step S244).

Also, when the microcomputer 4 instructs to adopt intra coding for the predictive coding of the frame (n+1) (step S241: intra), the moving picture coding device 31 determines the quantization step size Qp for quantizing the frame (n+1) [n+1] is (Cn/(dstC*P))*Qp[n] (step S145).

[Effect]

8A to 8D are schematic diagrams showing the change of the quantization step size, the amount of codes generated in each frame, the image quality, and the total amount of generated codes in time order according to the second embodiment.

Fig. 8A shows the quantization step size. In the second embodiment, the quantization step size for an intra-coded frame is appropriately determined (see step S245 in FIG. 7 ). Therefore, as shown in FIG. 8A, the quantization steps for frames 1, 11 and 21 do not take the same value. For example, the quantization step size for frame 11 is determined based on the amount of codes generated in previous frame 10 . Because frame 1 has no previous frame, a fixed value is assigned as the initial value of the quantization step for frame 1.

Fig. 8B shows the amount of encoding. The amount of codes generated from frame 1 to frame 10 is the same as that in the first embodiment. The intra-frame encoding generated in frame 11 and the encoding amount of subsequent frames are basically the same as the target encoding amount.

Figure 8C shows image quality. Image quality becomes stable because fluctuations in the generated encoding amount are suppressed. However, since the amount of codes generated by the intra-coded frames (frames 11 and 21) is small, the image quality degrades overall.

Fig. 8D shows the total code amount. In the second embodiment, the quantization step size for the intra-coded frame is properly determined so that the code amount generated by the intra-frame coded frame is basically equal to the target code amount. Therefore, the total encoding amount is close to the cumulative target encoding amount.

third embodiment

In the first embodiment, the code amount generated in an intra-coded frame is set to be larger than the target code amount. Therefore, the total amount of codes is greater than the cumulative target total amount of codes.

In the third embodiment, a technique for suppressing deterioration of image quality as much as possible while the total amount of encoding is substantially equal to the cumulative target encoding amount will be described below.

[structure]

FIG. 9 shows the structure of a moving image imaging device according to a third embodiment.

The structure of the moving picture coding device 33 according to the third embodiment is basically the same as that of the moving picture coding device 31 according to the first embodiment except for the addition of the difference calculation unit 14 and the target code correction unit 15 . Except for the difference calculation unit 14 and the target code correction unit 15, the same elements as in the moving picture coding device 31 are included in the moving picture coding device of the third embodiment. Therefore, no description will be given here for the same elements.

When encoding frame (n+1), the difference calculation unit 14 calculates the target encoding amount produced by subtracting frame n and each frame before frame n from the total encoding amount produced by frame n and each frame before frame n to obtain The accumulated difference value IntC.

IntC＝∑(Cn-dstC) Formula 11

When the cumulative difference is greater than 0, the target coding correction unit 15 corrects the target coding amount to be smaller when the cumulative difference is larger. When the cumulative difference is 0 or less, the target code correcting unit 15 does not correct the target code amount. Specifically, the correction operation is performed according to the following formulas 12 and 13.

When IntC is greater than 0:

dstC'=dstC-IntC/d Formula 12

When IntC is less than or equal to 0:

dstC'=dstC Formula 13

In this formula, d is an adjustment parameter predetermined to be greater than 1.

The quantization step size determination unit 13 determines the quantization step size for frame (n+1) quantization according to the following formulas 14 and 15.

When frame (n+1) will be intra-coded:

Qp[n+1]=Qpintra Formula 14

When frame (n+1) will be inter-coded:

Qp[n+1]=(Cn’/dstC’)*Qp[n] Formula 15

Note that the code correcting unit 12 operates in the same manner as in the first embodiment, and passes the corrected code amount Cn' to the quantization step size determining unit 13.

An embodiment for realizing the difference calculation structure of the above-mentioned difference calculation unit 14 and target code correction unit 15 will be described below.

FIG. 10 shows an example of the detailed structure of the difference calculation unit 14 and the target code correction unit 15. As shown in FIG.

The difference calculation unit 14 includes an encoding measurement unit 141, registers R1 and R2, and adders A1 and A2.

The encoding measurement unit 141 measures the amount of encoding generated by encoding the frame n by the encoding unit 11 . The register R1 holds the target code amount dstC, and the register R2 holds the cumulative difference between frame n and frames before frame n.

In the adder A1, the target code amount dstC is subtracted from the generated code amount Cn. In adder A2, (Cn-dstC) output from adder A1 is added to the cumulative difference held in register R2. The cumulative difference value IntC output from the adder A2 is then stored in the register R2 and output to the target code correction unit 15 .

The target code correction unit 15 includes a determination unit 151, registers R3, R4, and R5, a selector S1, and an adder A3.

The determination unit 151 determines whether the cumulative difference 151 is greater than 0, and then outputs the result to the control terminal of the selector S1.

According to the judgment result of the determination unit 151, when IntC is greater than 0, the selector S1 transmits the cumulative difference IntC to the multiplier M1, and when IntC is less than or equal to 0, transmits 0 held in the register R4 to the multiplier M1.

This multiplier M1 multiplies the cumulative difference IntC by the parameter 1/d for regulation. This adder A3 subtracts the product result from the target code amount dstC held in the register R5, and outputs the subtraction result to the quantization step size determination unit 13.

Through the above structure, when the cumulative difference is greater than 0, the target code correction unit 15 can realize the operation of correcting the target code amount smaller as the cumulative difference increases, and does not correct the target code when the cumulative difference is equal to or smaller than 0 quantity.

[operate]

FIG. 11 shows the operation of the moving image encoding device 33 according to the third embodiment.

The moving image encoding device 33 starts encoding in response to an instruction from the microcomputer 4 . First, the video encoding device 33 intra-encodes frame 1 (step S31). Because frame 1 has no previous frame, frame 1 is always intra-coded. In the third embodiment, Qpintra is used as the initial value of the quantization step size during intra-frame coding. Then, the moving picture coding device 33 stores the code amount generated by the frame 1 in the frame coding memory 122 (step S32). Because frame 1 is intra-coded, C1*P is stored in frame code memory 122 .

Next, the moving image encoding device 33 corrects the target encoding amount (step S34), and determines the quantization step size for frame 2 based on the encoding amount generated for frame 1, the quantization step size for frame 1, and the corrected target encoding amount (step S35). Then, the moving picture encoding device 33 inter-encodes frame 2 using the determined quantization step size (step S36).

The moving picture encoding device 33 stores the amount of encoding generated by the frame 2 in the frame encoding memory 122 (step S37). Since frame 2 is inter-coded, C2 is stored in frame code memory 122 . Thereafter, the moving picture encoding device 33 repeats the operations from step S33 to step S38 until the microcomputer 4 instructs to stop imaging.

FIG. 12 shows the detailed operation of correcting the target code amount according to the third embodiment.

The difference calculation unit 14 calculates the cumulative difference IntC (step S341).

When the cumulative difference value IntC is greater than 0 (step S342: Yes), the target encoding correcting unit 15 corrects the target encoding amount to dstC-IntC/d. When the cumulative difference is less than or equal to 0 (step S342: No), the target encoding correcting unit 15 does not correct the target encoding amount and maintains dstC (step S344).

The operation of determining the quantization step size is basically the same as that of the first embodiment, and thus will not be described in detail here.

[Effect]

13A to 13D are schematic diagrams showing the change in time sequence of the quantization step size, the amount of codes generated in each frame, the image quality, and the total amount of generated codes according to the third embodiment.

Fig. 13A shows the quantization step size. In the third embodiment, the quantization step size for an intra-coded frame is set to a fixed value Qpintra.

Fig. 13B shows the amount of encoding. The code amount C1 generated in frame 1 is corrected to be the corrected code amount C1*P. The target code amount dstC is corrected to the corrected code amount dstC'. The corrected target code amount dstC' becomes smaller than the target code amount dstC by a code amount by which the generated code amount C1 exceeds the target code amount dstC.

The code amount C2 generated by frame 2 is approximately equal to the corrected target code amount dstC'.

The code amount C3 generated in frame 3 is closer to the target code amount than the code amount C2 generated in frame 2 . This is because the cumulative difference IntC at frame 3 is smaller than the cumulative difference IntC at frame 2 , and thus the corrected target code amount at frame 3 is closer to the target code amount dstC than frame 2 .

Also, in the third embodiment, since the cumulative difference is employed, although the encoding amount of one frame will increase sharply, smooth change can be realized. This also applies to frames that do not come immediately after the intra-coded frame. For example, frame 26 produces a sharp increase in the amount of code, but changes in the amount of code produced in subsequent frames 27 and 28 are smooth.

Fig. 13C shows image quality. Image quality becomes stable because it is possible to suppress fluctuations in the amount of encoding generated in several inter-coded frames immediately following an intra-coded frame. Furthermore, since the amount of codes resulting from inter-coded frames is the same as in the case of the first embodiment, overall image quality does not degrade like the second embodiment.

Fig. 13D shows the total code amount. In the third embodiment, when the cumulative difference value IntC is greater than 0, the target coding amount is corrected to become smaller as the cumulative difference value becomes larger. Therefore, since even if the amount of codes generated by an intra-coded frame exceeds the target code amount, the code amount generated in several frames after the intra-coded frame becomes smaller than the target code amount, the total amount of codes is closer to the cumulative target code amount .

Fourth Embodiment

In the fourth embodiment, the intra frame is determined by switching between the first mode in which the quantization step size is set to a fixed value and the second mode in which the quantization step size is appropriately determined according to the remaining space size of the buffer. Quantization step size for encoded frames.

[structure]

FIG. 14 shows the structure of a moving picture imaging device according to a fourth embodiment.

The moving image encoding device 34 according to the fourth embodiment has a structure in which the space measurement unit 16, the buffer 17, and the transmission unit 18 are added to the moving image encoding device 31 according to the first embodiment. The moving picture coding device 34 has basically the same structure as the moving picture coding device 31 according to the first embodiment except for the space measurement unit 16, the buffer 17, and the transmission unit 18, and thus will not be described here.

The space measurement unit 16 measures the size of the remaining space of the buffer 17 and notifies the quantization step size determination unit 13 of the measurement result.

The buffer 17 temporarily stores the video stream generated by the encoding unit 11 . The transfer unit 18 transfers the video stream stored in the buffer 17 into the video stream memory 6 .

The quantization step size determination unit 13 determines the quantization step size of frame (n+1) quantization according to the following formulas 16 and 17.

When frame (n+1) is to be intra-coded and the free space of buffer 17 is larger than a predetermined size (eg, 20% of the buffer):

Qp[n+1]=Qpintra Formula 16

When frame (n+1) is to be intra-coded and the free space of buffer 17 is less than or equal to the predetermined size

hours or when frame(n+1) will be inter-coded:

Qp[n+1]=(Cn’/dstC’)*Qp[n] Formula 17

In the above formulas, Cn' is the same as Cn' in formulas 7, 8 and 9.

The above description performs the operation in the first mode that operates as in the first embodiment when the free space of the buffer 17 is larger than the predetermined size, and in the same manner as the second embodiment when the free space of the buffer 17 is smaller than or equal to the predetermined size. The second mode performs the same operation.

[operate]

Fig. 15 is a detailed schematic diagram of determining the quantization step size according to the fourth embodiment.

When the microcomputer 4 instructs frame (n+1) to be inter-coded (step S441: inter-frame) and frame n has been inter-coded (step S442: inter-frame), the moving image coding device 34 determines The quantization step size Qp[n+1] of (n+1) quantization is (Cn/dstC)*Qp[n] (step S443). On the other hand, when frame n has been intra-coded (step S442: intra), the moving picture coding device 34 determines the quantization step size Qp[n+1] for frame (n+1) quantization as (Cn* P/dstC)*Qp[n] (step S444).

Also, when the microcomputer 4 instructs to intra-encode the frame (n+1) (step S441: intra) and the free space is smaller than or equal to a predetermined size (step S445: NO), the moving image encoding device 34 determines to use The quantization step size Qp[n+1] of frame (n+1) quantization is (Cn/(dstC*P))*Qp[n] (step S446). If the free space is larger than the predetermined size (step S445: YES), the moving picture encoding device 34 determines the quantization step size Qp[n+1] for frame (n+1) quantization to be Qpintra (step S447).

[Effect]

When the quantization step size used for intra-coded frame quantization takes a fixed value, the amount of codes produced by the intra-coded frame cannot be adjusted even though the buffer almost overflows. However, if the quantization step size used for intra-coded frame quantization is properly determined, the amount of code generated by the intra-coded frame can be suppressed.

Therefore, if the buffer is about to overflow, the present invention has the advantage of reducing the possibility of buffer overflow by properly determining the quantization step size for intra-coded frame quantization.

Fifth Embodiment

In the fifth embodiment, according to the size of the remaining space of the buffer, the number of inter-coded frames is determined by switching between the first mode of not correcting the target code amount and the second mode of correcting the target code amount Quantization step size.

[structure]

FIG. 16 shows the structure of a moving image imaging device according to a fifth embodiment.

The moving image encoding device 35 according to the fifth embodiment has a structure in which the space measurement unit 16, the buffer 17, and the transmission unit 18 are added to the moving image encoding device 33 according to the third embodiment. The moving picture coding device 35 has basically the same structure as the moving picture coding device 33 according to the third embodiment except for the space measurement unit 16, the buffer 17, and the transmission unit 18, and thus will not be described here.

The space measurement unit 16, the buffer 17, and the transmission unit 18 are the same as those of the fourth embodiment.

The quantization step size determining unit 13 determines a quantization step size for frame (n+1) quantization according to the following formulas 18, 19 and 20.

When frame (n+1) will be intra-coded:

Qp[n+1]=Qpintra Formula 18

When frame (n+1) is to be inter-coded and the free space of buffer 17 is larger than a predetermined size:

Qp[n+1]=(Cn’/dstC)*Qp[n] Formula 19

When frame (n+1) is to be inter-coded and the free space of buffer 17 is less than or equal to a predetermined size:

Qp[n+1]=(Cn’/dstC’)*Qp[n] Formula 20

In the above formulas, Cn' is the same as Cn' in formulas 7, 8 and 9, and dstC' is the same as dstC' in formulas 12 and 13.

The above description performs operations in the first mode that operates as in the second embodiment when the free space of the buffer 17 is larger than the predetermined size, and in the same manner as the third embodiment when the free space of the buffer 17 is smaller than or equal to the predetermined size. The second mode performs the same operation.

[operate]

FIG. 17 shows the operation of the moving picture encoding device 35 according to the fifth embodiment.

The moving image encoding device 35 starts encoding in response to an instruction from the microcomputer 4 . First, the video encoding device 35 intra-encodes frame 1 (step S51). Because frame 1 has no previous frame, frame 1 is always intra-coded. In the fifth embodiment, Qpintra is used as the initial value of the quantization step during intra-frame coding. Then, the moving picture encoding device 35 stores the amount of encoding generated by frame 1 in the frame encoding memory 122 (step S52). Because frame 1 is intra-coded, C1*P is stored in frame code memory 122 .

When the free space is less than or equal to the predetermined size (step S54: No), the moving image encoding device 35 corrects the target encoding amount (step S55), and when the free space is larger than the predetermined size (step S54: Yes), the moving image encoding device 35 Step 55 for correcting the target encoding amount is skipped.

The moving image encoding device 35 determines the encoding scale for frame 2 based on the obtained target encoding amount (step S56). Then, the moving picture encoding device 35 inter-encodes frame 2 using the determined quantization step size (step S57).

The moving picture encoding device 35 stores the amount of encoding generated by the frame 2 in the frame encoding memory 122 (step S58). Because frame 2 is inter-coded, C2 is stored in the frame code memory 122 . Thereafter, the moving picture encoding device 35 repeats the operations from step S53 to step S58 until the microcomputer 4 instructs to stop imaging.

[Effect]

The moving picture encoding device 35 according to the fifth embodiment corrects the encoding amount generated at frame (n+1) according to the excess amount of frame n and frames before frame n when the buffer is about to overflow. Therefore, the possibility of buffer overflow can be reduced.

Sixth Embodiment

In the sixth embodiment, when frame (n+1) is intra-coded, the quantization step size is determined based on the quantization step size for frame n. This is the only difference from the first embodiment. Others are the same as the first embodiment, and therefore will not be explained here.

[structure]

If the frame (n+1) is intra-coded, the code correcting unit 12 does not correct the generated code amount.

The quantization step size determining unit 13 determines a quantization step size for frame (n+1) quantization according to the following formulas 21, 22 and 23.

When frame (n+1) will be inter-coded:

Qp[n+1]=(Cn’/dstC)*Qp[n] Formula 21

When frame (n+1) is to be intra-coded, and frame n is already inter-coded:

Qp[n+1]=Pq1*Qp[n] Formula 22

When frame (n+1) is to be intra-coded, and frame n is already intra-coded:

Qp[n+1]=Pq2*Qp[n] Formula 23

Here, Pq1 and Pq2 are predetermined coefficients. Specifically, both coefficients are set within the range of 5/4 to 4/3.

As the coefficients Pq1 and Pq2 gradually become larger, the effect of suppressing the amount of coding generated by intra-coding frames increases. However, this simultaneously causes image quality degradation. The simulation results show that by setting the coefficients Pq1 and Pq2 in the range of 5/4 to 4/3, both the suppression effect of the code amount produced by the intra-coded frame and the excellent image quality can be obtained.

[operate]

Fig. 18 is a detailed schematic diagram of determining the quantization step size according to the sixth embodiment.

When the microcomputer 4 instructs frame (n+1) to be inter-coded (step S641: inter) and frame n has already been inter-coded (step S642: inter), the moving image coding device 31 determines The quantization step size Qp[n+1] of (n+1) quantization is (Cn/dstC)*Qp[n] (step S643). When frame n has been intra-coded (step S642: intra-frame), the moving image coding device 31 determines that the quantization step size Qp[n+1] for frame (n+1) quantization is (Cn*P/dstC )*Qp[n] (step S644).

Also, when the microcomputer 4 instructs frame (n+1) to be intra-coded (step S641: intra) and frame n has been inter-coded (step S645: inter), the moving picture coding device 31 determines to use The quantization step size Qp[n+1] of frame (n+1) is Pq1*Qp[n] (step S646). If frame n has been intra-coded (step S645: intra), the moving image coding device 31 determines a quantization step size Pq2*Qp[n] for quantization of frame (n+1) (step S647).

FIG. 19A to FIG. 19C are diagrams showing the time-sequential change of the quantization step size, the code amount generated in each frame, and the image quality according to the sixth embodiment.

Fig. 19A shows the quantization step size. In the sixth embodiment, the quantization step size for an intra-coded frame is appropriately determined (see step S645 of FIG. 18 ). Therefore, as shown in FIG. 19A, the quantization step sizes of frames 1, 11, and 21 take different values. For example, the quantization step size of the frame 11 is determined based on the quantization step size of the previous frame 10 . Since there is no frame before frame 1, a fixed value is assigned as the initial value of the quantization step size for frame 1.

Fig. 19B shows the amount of encoding. The amount of codes generated from frame 1 to frame 10 is the same as that in the first embodiment. The code amount generated by the intra-frame coded frames of frame 11 and subsequent frames is basically equal to the target code amount.

Fig. 19C shows image quality. Since fluctuations in the amount of encoding generated can be suppressed, image quality becomes stable. However, since the amount of codes generated by the intra-coded frames (frames 11 and 21) remains small, the image quality degrades overall.

The change of the total code amount in the sixth embodiment is basically the same as that in the second embodiment, and thus will not be described here.

In the sixth embodiment, when performing intra coding, the coding step size is appropriately determined. Therefore, in the sixth embodiment, as in the second embodiment, good image effects can be obtained and encoding becomes stable.

In addition, in the sixth embodiment, during intra-frame encoding, the quantization step size is determined regardless of the amount of encoding generated in the previous frame. Therefore, compared with the second embodiment, the amount of processing in the operation of the moving picture encoding device can be reduced.

Seventh Embodiment

In the seventh embodiment, when frame (n+1) is intra-coded, the quantization step size is determined based on the quantization step size for frame n. This is the only difference from the first embodiment. Others are the same as the first embodiment, and therefore will not be explained here.

[structure]

FIG. 20 shows the structure of a moving picture imaging device according to a seventh embodiment.

The moving image encoding device 37 includes an encoding unit 11 , an encoding correction unit 12 , a quantization step size determination unit 13 , and a table storage unit 19 .

The table storage unit 19 stores a table representing the relationship between the quantization step size of frame n and the quantization step size of frame (n+1).

When the frame (n+1) is to be intra-coded, the quantization step size determination unit 13 refers to the table stored in the table storage unit 19 to determine the quantization step for the frame (n+1) based on the quantization step size for the frame n long.

FIG. 21 shows an example of a table stored in the table storage unit 19. As shown in FIG.

When frame n is inter-coded, the quantization step size for frame (n+1) is determined to be Qnew1. When frame n is intra-frame encoded, the quantization step size of frame (n+1) is determined to be Qnew2.

[operate]

Fig. 22 is a detailed schematic diagram of determining the quantization step size according to the seventh embodiment.

When the microcomputer 4 instructs frame (n+1) to be inter-coded (step S741: inter) and frame n has already been inter-coded (step S742: inter), the moving image coding device 37 determines to The quantization step size Qp[n+1] of frame (n+1) quantization is (Cn/dstC)*Qp[n] (step S743). If frame n has been intra-coded (step S742: intra-frame), the moving image coding device 37 determines that the quantization step size Qp[n+1] for frame (n+1) quantization is (Cn*P/dstC) *Qp[n] (step S744).

Also, when the microcomputer 4 instructs frame (n+1) to be intra-coded (step S741: intra) and frame n has been inter-coded (step S745: inter), the moving image coding device 37 determines to use The quantization step size Qp[n+1] of frame (n+1) is Qnew1 (step S746). If frame n has been intra-coded (step S745: intra), the moving picture coding device 37 determines a quantization step size Qnew2 for quantization of frame (n+1) (step S747).

[Effect]

According to the seventh embodiment, the same effect as that of the sixth embodiment can be obtained.

Eighth embodiment

In the eighth embodiment, the quantization step size is determined for each block line. The block line is a line of macroblocks arranged in a matrix in an image.

[structure]

FIG. 23 shows the structure of a moving picture imaging device according to an eighth embodiment.

The moving picture coding device 38 according to the eighth embodiment is basically the same as the moving picture coding device 33 according to the third embodiment, and only the structure of the coding correction unit 12 and the quantization step size determining unit 13 is different from the third embodiment. Therefore, the same elements are not described here.

The code correcting unit 12 measures the generated code amount of each block line, and corrects the generated code amount according to the formula 24 for correction. Note that, in the eighth embodiment, the quantization step size of a block line (n+1) is determined with reference to the amount of encoding produced by a predetermined number of block lines. Here, an embodiment in which one frame includes 7 block lines and the quantization step size for the block line (n+1) is determined with reference to the 7 block lines before the block line (n+1) is described.

Cn’＝CAna*P1+CAnb*P2 Formula 24

In the above formula, CAna is the code amount generated in the block line for intra-coding among the code amounts generated in the latest 7 block lines, and CAnb is in the code amount generated in the latest 7 block lines Amount of codes produced in block lines that are inter-coded. In addition, P1 and P2 are correction coefficients with a relationship of P1<P2. In the eighth embodiment, the case where P1=P (0<P<1) and P2=1 is taken as an example. The correction coefficient P satisfies Formula 4.

The quantization step size determining unit 13 determines a quantization step size for block line (n+1) quantization according to the following formulas 25 and 26.

When block lines (n+1) are to be intra-coded:

Qp[n+1]=Qpintra Formula 25

When the block line (n+1) is inter-coded:

Qp[n+1]=(Cn’/dstC’)*Qp[An] Formula 26

In the above formula, Qp[An] is the average or modulus value of the quantization steps for the 7 block lines to be referenced.

An example of a structure for realizing the calculation of Equation 24 is described below.

FIG. 24 shows an example of the detailed structure of the code correcting unit 12. As shown in FIG.

The code correction unit 12 includes a code measurement unit 121, a block line code memory 123, registers R1 and R2, a selector S1, a multiplier M1, and an adder A1.

The encoding measurement unit 121 measures the encoding amount Cn generated at the block line n that has been encoded by the encoding unit 11 . The register R1 holds the correction coefficient P, and the register R2 holds the correction coefficient 1. According to an instruction from the microcomputer 4, the selector S1 applies the correction coefficient P to the multiplier M1 when the block line n is intra-coded, and applies the correction coefficient 1 to the multiplier M1 when the block line n is inter-coded .

The multiplier M1 multiplies the generated code amount Cn by a specified correction coefficient.

The block line code memory 123 is a shift register composed of seven registers. The calculation results of the multiplier M1 are sequentially stored in the block line code memory 123 . The adder A1 adds up the stored results to obtain the corrected code amount Cn'.

In Fig. 24, C5 to C11 respectively generate code amounts. In the figure, it is assumed that C5 to C7 are intra-frame encoded, and C8 to C11 are inter-frame encoded.

In this case, the calculation result in the adder A1 is the following formula 27.

Cn’＝(C5+C6+C7)*P+(C8+C9+C10+C11) Formula 27

In the above formula, C5+C6+C7 is the total amount of codes CAna generated by intra-frame coding block lines, and C8+C9+C10+C11 is the total amount of codes CAnb generated by inter-frame coded block lines.

With the above structure, the code correcting unit 12 can output the corrected code amount Cn' by Equation 24.

[operate]

FIG. 25 shows the operation of the moving picture encoding device 38 according to the eighth embodiment.

The moving picture encoding device 38 starts encoding in response to an instruction from the microcomputer 4 . First, the video encoding device 38 intra-encodes the block line 1 (step S82). Because block line 1 has no previous block lines, block line 1 is always intra-coded. Qpintra is used as the initial value of the quantization step size when intra-coding a block line. Then, the moving picture coding device 38 stores the code amount generated by the block line 1 in the block line coding memory 123 (step S82). Since the block line 1 is intra-coded, C1*P is stored in the block line coding memory 123 .

Thereafter, the moving image encoding device 38 repeats the operations from step S81 to step S84 until the encoding of the block line 7 is completed, and thus, the intra encoding of the frame 1 is completed.

Once the intra-frame encoding of block line 7 is completed, the moving picture encoding device 38 corrects the target encoding amount (step S86), and based on the total amount of encoding generated by block lines 1 to 7, the quantization step size of block lines 1 to 7, and the correction target encoding amount Quantization step size for block line 8 is determined (step S87). The moving image encoding device 38 inter-codes the block line 8 using the determined quantization step size (step S88).

The moving image coding device 38 stores the code amount generated by the block line 8 in the block line coding memory 123 (step S89). Since the block line 8 is inter-coded, C8 is stored in the block line coding memory 123 . Thereafter, the moving image encoding device 38 repeats the operations of steps S85 to S90 until the microcomputer 4 instructs to end imaging.

Fig. 26 is a detailed schematic diagram of determining the quantization step size according to the eighth embodiment.

When the microcomputer 4 instructs the block line (n+1) to perform inter-frame coding (step S851: inter-frame), the moving image coding device 38 determines the quantization step size Qp[n+ 1] is {(CAna*P+CAnb)/dstC'}*Qp[An] (step S852).

On the other hand, when the microcomputer 4 instructs the block line (n+1) to perform intra-coding (step S851: intra), the moving image coding device 38 determines the quantization step for quantizing the block line (n+1). Long Qp[n+1] is Qpintra (step S853).

The operation of correcting the target code amount is the same as that of the third embodiment, so it will not be described here.

[Effect]

Fig. 27 is a schematic diagram showing the change of the encoding amount in time sequence according to the eighth embodiment.

In determining the quantization step size for the block line 12, the amount of encoding produced by the seven block lines (block lines C5 to C11) included in the A5 section is referred to. Of the 7 block lines located in section A5, block lines 5 to 7 located in subsection A5a are intra-coded and are therefore subject to correction. Block lines 8 to 11 in subsection A5b of the 7 block lines in section A5 are inter-coded and therefore need not be corrected.

As described above, by determining the quantization step size of each block line, the amount of coding can be adjusted and stabilized in units of smaller time segments.

Ninth Embodiment

In the ninth embodiment, the moving image encoding device 39 can perform all types of methods of determining the quantization step size from the first to eighth embodiments, and determine the quantization step size using one of the methods randomly selected by the user.

[structure]

FIG. 28 shows the structure of a moving picture imaging device according to a ninth embodiment.

The moving image imaging device is provided with a mode switching button 7 . Except for this, the moving image imaging device according to the ninth embodiment is the same as the moving image imaging devices according to the first to eighth embodiments, and thus will not be described here.

The mode switching button 7 accepts selection of one of the methods for determining the quantization step size. This selection is notified to the video encoding device 39 via the microcomputer 4 . The moving picture encoding means 39 encodes moving pictures using the selected method for determining the quantization step size.

The moving picture imaging device and moving picture encoding device based on the preferred embodiments have been described above. However, the present invention is not limited to the above embodiments, and the following modifications are also included in the present invention.

(1) In the above preferred embodiments, the code correcting unit 12 performs correction by multiplying Cn and a correction coefficient to realize the formula Qp[n+1]=(Cn*P/dstC)*Qp[n]. However, the interpretation of the formulas is not limited to the above-mentioned embodiments. For example, the following explanation may be adopted, instead of correcting the resulting code amount, the target code amount may be corrected to P/dstC when frame n has been intra-coded. In this case, the encoding correcting unit 12 shown in FIG. 1 is unnecessary, and the quantization step size determination unit 13 selects P/dstC as the target encoding amount when frame n has been intra-frame encoded, and selects 1/dstC as the frame n The target encoding amount when n has been inter-encoded.

(2) In the above preferred embodiments, the correction coefficient P is a fixed value. However, the correction coefficient P is not limited to a fixed value, and can be increased as the quantization step size of the frame to be referred to becomes smaller, as in Formula 28 below.

P=1/(Qn-b)+c Formula 28

In this formula, b and c are adjustable parameters.

Also, the quantization step size may be determined with reference to a straight line or a constant approximate to Formula 28, or with reference to a table assigned to each quantization step size.

(3) In the sixth embodiment, correction of the target encoding amount is performed frame by frame. However, this correction may also be performed for each block line.

(4) The encoding unit 11 is explained as an MPEG encoder. However, the encoding unit 11 is not limited to an MPEG encoder if the encoding unit can encode data using both predictive encoding and quantization.

(5) In the above preferred embodiment, the quantization step size of frame (n+1) is determined by calculation. However, the present invention is not limited thereto. For example, the quantization step size for frame (n+1) may be selected from a plurality of quantization step sizes stored corresponding to combinations of predictive encoding, quantization step size, and generated code amount of frame n. In this case, the quantization step size can be determined without calculation for each frame.

(6) In the above-mentioned preferred embodiments, various parameters of frame n (predictive coding of frame n, quantization step size, and generated code amount) are considered. However, the present invention is not limited to this embodiment, and parameters of an arbitrary frame before frame (n+1) may be used. For example, parameters of frame (n-1) may be used.

(7) In the above preferred embodiments, various parameters of one frame are considered. However, the present invention is not limited thereto. More than one frame parameter can be used. Although the use of parameters of more than one frame improves the processing capacity of the motion image encoding device, the effect of suppressing fluctuations is better when the parameters of more than one frame are used.

Although the present invention has been fully described by way of embodiments with reference to the accompanying drawings, it should be noted that various modifications and improvements to the present invention will be apparent to those skilled in the art. Therefore, unless otherwise such modifications and improvements depart from the scope of the present invention, they should be construed as being included therein.

Claims (19)

1, a kind of dynamic image encoding device comprises:

Coding unit is used for each frame of a plurality of frames of moving image is optionally carried out interframe encode or intraframe coding, and this intraframe coding and interframe encode include quantification; And

Determining unit is used for being identified for the quantization step that second frame data quantize according to the encoding amount that results from first frame, wherein

When coding unit carries out interframe encode to the data of second frame, (i) if the data of first frame have been carried out interframe encode, adopt first method to make decision, if perhaps (ii) the data of first frame have been carried out intraframe coding and adopted second method to make decision, the value that obtains by the quantization step that adopts second method to determine is less than the value that obtains by the quantization step that adopts first method to determine.

2, dynamic image encoding device according to claim 1, it is characterized in that described second method is provided with to drop into by the quantization step that adopts second method to determine by the mode in 1/5 to 1/3 the scope that adopts quantization step that first method determines.

3, dynamic image encoding device according to claim 1 is characterized in that, this device further comprises:

Difference obtains the unit, is used for obtaining (i) and results from first frame and prior to the accumulation encoding amount of each frame of first frame and (ii) first frame and prior to the difference between the accumulation predictive encoding aim parameter in each frame of first frame, wherein

First method and second method are big and mode that increase is provided with along with the difference quantitative change of the encoding amount that produces with the quantization step determined.

4, dynamic image encoding device according to claim 3, it is characterized in that, the encoding amount that produces in accumulation is during less than the cumulative target encoding amount, and first and second methods are big and become big mode and be provided with along with the coding quantitative change that produces with quantization step, and do not consider the difference amount.

5, dynamic image encoding device according to claim 3 is characterized in that, this device further comprises:

The buffer that is used for the storing moving image coded data, wherein

When described buffer had free space greater than pre-sizing, this first and second method was big and become big mode and be provided with along with the coding quantitative change that produces with quantization step, and does not consider the difference amount.

6, dynamic image encoding device according to claim 1 is characterized in that, when coding unit will carry out intraframe coding to the data of second frame, it was quantization step that described determining unit is set fixed value.

7, dynamic image encoding device according to claim 1 is characterized in that,

When the digital coding of described coding unit first frame will be carried out intraframe coding to second frame data later on, determining unit is determined quantization step by adopting third party's method, and the value that the quantization step of determining by employing third party method obtains is greater than the value of determining by first method.

8, dynamic image encoding device according to claim 1 is characterized in that, this device further comprises:

The buffer that is used for the storing moving image coded data, wherein

When coding unit will carry out intraframe coding to the data of second frame later on to first frame data coding, (i) if buffer has the free space greater than pre-sizing, it is fixed value that determining unit is set quantization step, if and the free space of (ii) buffer existence is a preliminary dimension or littler, determining unit adopts third party's method to determine quantization step, and the value that the quantization step of determining by employing third party method obtains is greater than the value of determining by first method.

9, dynamic image encoding device according to claim 1, it is characterized in that, at described coding unit the data of first frame are encoded the back will carry out intraframe coding to second frame data time, determining unit is determined the quantization step of second frame based on being used for the quantization step that first frame quantizes, and does not consider the encoding amount that produces.

10, dynamic image encoding device according to claim 9, it is characterized in that, in the time of will carrying out intraframe coding to second frame data after coding unit is encoded to the data of first frame, determining unit is by being multiply by the quantization step that the quantization step that is used for first frame is identified for second frame by predetermined value.

11, dynamic image encoding device according to claim 10 is characterized in that, described predetermined value is in 5/4 to 4/3 scope.

12, dynamic image encoding device according to claim 9 is characterized in that, this device further comprises:

Record cell is used to write down the table that concerns between the quantization step of expression first frame and the second frame quantization step, wherein

Determining unit is by being identified for the quantization step of second frame based on the quantization step that is used for first frame with reference to described table.

13, dynamic image encoding device according to claim 9 is characterized in that, described determining unit is identified for the quantization step of second frame based on each quantization step of the preceding frame that also is useful on scheduled volume except that the quantization step that is used for first frame.

14, a kind of dynamic image encoding device comprises:

Coding unit is used for each data of many sticks of moving image line are optionally carried out interframe encode or intraframe coding, and interframe encode and intraframe coding include quantification;

Correcting unit, be used for the encoding amount that the piece line of one or more interframe encode that (i) adopt first method to proofread and correct the encoding block line of scheduled volume produces, and (ii) adopting second method to proofread and correct the encoding amount that one or more Intra-coded blocks line in the encoding block line of scheduled volume produces, the value that the encoding amount that produces by the one or more Intra-coded blocks line that adopts second method to proofread and correct obtains is less than the value of proofreading and correct by first method; And

Determining unit when coding unit will carry out interframe encode to the data of next piece line, is used for being identified for based on the coding total amount of Intra-coded blocks line of having proofreaied and correct and the generation of inter-coded block line the quantization step of next piece line data-measuring.

15, dynamic image encoding device according to claim 14, it is characterized in that second method is set to fall into by the encoding amount that adopts second method to produce by the mode in 1/5 to 1/3 the scope that adopts encoding amount that first method proofreaies and correct in one or more inter-coded block lines.

16, a kind of moving image imaging device comprises:

Image device is used to gather target image and produces moving image;

Coding unit is used for each data of each moving image multiframe are optionally carried out interframe encode or intraframe coding, and interframe encode and intraframe coding include quantification; And

Determining unit is used for being identified for the quantization step that second frame data quantize according to the encoding amount that first frame produces, wherein

When coding unit will carry out interframe encode to the data of second frame, by (i) if the data of first frame interframe encode, adopt first method to make decision, if perhaps (ii) the intraframe coding of the data of first frame adopts second method to make decision, the value that obtains by the quantization step that adopts second method to determine is less than the value that obtains by the quantization step that adopts first method to determine.

17, a kind of moving image imaging device comprises:

Image device is used to gather target image and produces moving image;

Coding unit is used for each data of many sticks line of moving image are optionally carried out interframe encode or intraframe coding, and interframe encode and intraframe coding include quantification;

Correcting unit, be used for (i) and adopt the encoding amount of first method correction in the piece line generation of one or more interframe encode of the encoding block line of scheduled volume, and (ii) adopting second method to proofread and correct the encoding amount that one or more Intra-coded blocks line in the encoding block line of scheduled volume produces, the value that the encoding amount that produces by the one or more Intra-coded blocks line that adopts second method to proofread and correct obtains is less than the encoding amount of proofreading and correct by first method; And

Determining unit when coding unit will carry out interframe encode to the data of next piece line, is used for being identified for based on the coding total amount of Intra-coded blocks line of having proofreaied and correct and the generation of inter-coded block line the quantization step of next piece line data-measuring.

18, a kind of dynamic image encoding method comprises:

Coding step optionally carries out interframe encode or intraframe coding to each data of multiframe in the moving image, and interframe encode and intraframe coding include quantification;

Determining step is determined the quantization step that second frame data quantize according to the encoding amount that first frame produces, wherein

In the time will carrying out interframe encode to the data of second frame, by (i) if the data of first frame interframe encode adopt first method to make decision, if perhaps (ii) the intraframe coding of the data of first frame adopts second method to make decision, the value that obtains by the quantization step that adopts second method to determine is less than the value that obtains by the quantization step that adopts first method to determine.

19, a kind of dynamic image encoding method comprises:

Coding step optionally carries out interframe encode or intraframe coding to each data of polylith line in the moving image, and interframe encode and intraframe coding include quantification;

Aligning step, (i) adopt first method to proofread and correct the encoding amount that one or more inter-coded block line in the encoding block line of scheduled volume produces, and (ii) adopting second method to proofread and correct the encoding amount that one or more Intra-coded blocks line in the encoding block line of scheduled volume produces, the value that the encoding amount that produces by the one or more Intra-coded blocks line that adopts second method to proofread and correct obtains is less than the encoding amount of proofreading and correct by first method; And

Determining step in the time will carrying out interframe encode to the data of next piece line, is used for being identified for based on the coding total amount of Intra-coded blocks line of having proofreaied and correct and the generation of inter-coded block line the quantization step of next piece line data-measuring.

Images