JPH11346365A - Encoding/compression method and device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To attain encoding in consideration of the visual characteristics and psychological characteristics of a human being by providing a process for operating the first encoding under variable rate control on a first path, and for editing a target encoding compression condition obtained from the result of the first encoding between the first encoding and the second encoding. SOLUTION: A first encoding circuit 161 of an encoding compression device 160 variable length-encodes an inputted video signal. A data storage device 162 stores divided encoded streams generated by first and second encoding circuits 161 and 165, and each kind of encoding parameters. An encoding parameter encoding circuit 164 reads the encoding parameter of the block in which a user is not satisfied with the quality of data in the divided encoded stream encoded by the first encoding circuit 161 from the data storage device 162, and edits the encoding parameter according to an editing command inputted by an inputting means 163, and outputs the parameter at the time of operating encoding by the second encoding circuit 165.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] The present invention relates to an encoding and compression method,
More particularly, the present invention relates to a high-efficiency encoding method for a moving image signal, and more particularly to an encoding and compression method suitable for a storage medium such as an optical disk.

[0002]

2. Description of the Related Art Generally, storing image data as it is requires an enormous amount of memory. Therefore, a technique for efficiently compressing an image and storing the compressed image in a medium is an extremely important technique. On the other hand, an optical disc normally reproduces image data at a constant rate as seen in a CD or an LD. However, images generally include scenes containing a large amount of information and scenes not containing a large amount of information. As an encoding method that effectively utilizes the characteristics of the image, a variable rate encoding method has been proposed (Japanese Patent Laid-Open No. 7-284097).

First, an image encoding circuit generally used in MPEG will be described with reference to FIG. The encoding circuit 2 in FIG. 13 compresses data using the motion compensation DCT method. The motion-compensated DCT method compresses one frame selected periodically from input image data using only data in the frame, and compresses a difference from a previous frame for the remaining frames. This is one of the transmission methods. For the intra-frame compression and the inter-frame compression, typically, a discrete cosine transform, which is a kind of orthogonal basis transform, is used. Also, when calculating the difference between frames, the compression ratio is greatly improved by detecting the motion vector of the image between the previous frame and the motion, and then taking the difference.

Hereinafter, the operation of the encoding circuit 2 shown in FIG. 13 will be further described. In FIG. 13, a solid line represents a data flow, and a broken line represents a control flow. The image data input from the input terminal 30 is stored in the frame buffer 25 and then rearranged in the encoding order by the frame rearrangement circuit 26. The subtracter 10 is used to calculate a difference from the previous frame. The encoding control circuit 22 controls the on / off of the refresh switches 23 and 24 according to the type of picture to be processed. That is, when the picture to be processed is an I picture (intra picture), the encoding control circuit 22 turns off the refresh switches 23 and 24 (intra-frame compression).
As a result, the subtractor 10 does not operate.

The input image data is subjected to discrete cosine transform by a discrete cosine transform circuit (DCT) 11.
Discrete cosine transform is usually performed in two dimensions. 8x8
If a discrete cosine transform is performed for each block, an 8 × 8 coefficient is obtained as a result of the transform. Although the data subjected to DCT is originally a continuous quantity, since it is calculated using a digital circuit, each of the 64 coefficients is obtained as a digital value having a predetermined bit width. This data is then subjected to an optimal bit distribution for each frequency component by the quantization circuit 12. Normally, the low-frequency component is an important component of an image, so that the bit allocation is increased. The high-frequency component is not so important for configuring the image, and thus the bit allocation is reduced. Variable length coding circuit (VL
C) 13 performs variable length coding on the output of the quantization circuit 12. The variable-length coding is a technique of assigning a shorter code length to data having a statistically higher probability of occurrence. This technique removes a statistical redundant component of data. In this method, Huffman codes are often used.

The output of the quantization circuit 12 is the inverse quantization circuit 1
5 restores the quantization. The inverse quantization circuit 15
Contrary to the quantization, the amplitude of each frequency component is returned to the original amplitude. Each coefficient restored to the original amplitude by the inverse quantization is restored to the original image data by the inverse DCT circuit 16. If the restored image data is in-frame image data, the adder 17 does not operate. Thereafter, the restored image data is delayed by the frame memory 18 by a predetermined number of frames. The delayed image data is input to the motion vector detection circuit 20. The motion vector detection circuit 20 calculates a motion amount from the input image data. The motion compensation circuit 19 moves the position of the image data according to the amount of the motion. The motion-compensated image data is used by the subtracter 10 to calculate a difference from the next image data.

[0007] The image data of some frames following the I picture is used to compress the difference from the image data of the previous frame. When the picture to be processed is a P picture or a B picture, the encoding control circuit 22 turns on the refresh switches 23 and 24 (inter-frame compression). The refresh switch 23 is turned on when calculating the difference between frames, and the subtractor 10
Used to operate. Refresh switch 24
Are repeatedly turned on and off at the same cycle as the refresh switch 23. When the switch is on, the adder 17 is operated to add the inter-frame difference data and the previous frame data to restore the frame. The variable-length encoding circuit 13 performs variable-length encoding on inter-frame compressed data. Here, a picture from an I picture to the next I picture is called 1 GOP (group of pictures),
It consists of video signals of around 5 frames (about 0.5 seconds).

The output of the variable length coding circuit (VLC) 13 is output to an output terminal 31 via a buffer circuit 14.
The quantization width determination circuit 21 designates a quantization width for the quantization circuit 12 by checking the state of the buffer circuit 14.
That is, if the buffer circuit 14 outputs data at a predetermined constant rate, it is necessary to generate more data when the amount of data remaining in the buffer circuit 14 is small. And control to increase the number of generated bits. Conversely, when there is too much data remaining in the buffer circuit 14, it is necessary to make it difficult to generate data. Therefore, the quantization width Q is controlled to be slightly larger than before and the number of generated bits is reduced. is there.

That is, the target data amount to be generated in the next section is calculated from the remaining buffer amount, and the quantization width is calculated from the target data amount. If the target data amount is large, the quantization width Q needs to be small. Conversely, if the target data amount is small, the quantization width Q needs to be large. In other words, they are in inverse proportion.

The quantization width determination circuit 21 previously obtains a target data amount for each section of the video signal, and forcibly changes the target data amount for each short section simultaneously with encoding. , Variable rate control is realized.

Referring to FIG. 14, a description will be given of a variable rate control of a two-pass system which is generally performed in the past. In the first pass, a predetermined video signal, for example, a video signal of 2 hours, is input from a terminal 1, and the encoding circuit 2 performs first encoding. The encoding circuit 2 is the same as that shown in FIG.
The feedback loop from the control circuit to the quantization circuit 12 is omitted. That is, regardless of the amount of generated data, the quantization width Q1 determined by the quantization width determination circuit is fixed at a fixed value. However, it may be set independently for each picture.

The resulting data amount D1 is stored in the generated code amount storage circuit 3 for each short section, for example, for every 2 GOPs.
Remember. As an example of the generated data amount D1 and the target data amount D2 at that time, a simple model is shown in FIG.
(B). The horizontal axis shows time (GOP), and the vertical axis shows the amount of generated data. Section A is a standard image, section B is an image that is easy to encode with little motion, and section C is difficult to encode with much motion. An image is shown. 1 / of all time
Assuming three times, the ratio of the sum of the generated data amounts D1 generated in the respective sections is A: B: C = 2:
1: 3. Here, the numerical value itself on the vertical axis depends on the quantization width Q1, but in the present invention, the relative ratio of each section becomes more important than the numerical value itself on the vertical axis. If the total capacity DALL of the optical disc is 3 gigabytes, the allocation amount to the section A is calculated as 3 giga / 6 × 2 = 1 giga. Similarly, the interval B is calculated as 0.5 giga, and the interval C is calculated as 1.5 giga.

In general, if the amount of data generated in the short section generated in the first pass is D1, the sum of the data amount D1 is ΣD1, the target data amount is D2, and the total capacity is DALL, the target data amount of each short section D2 is D2 = D1 × DALL / ΣD
1 is required.

In this manner, for example, a video signal of 2 hours is divided into short sections every 2 GOPs (one second), and each target data amount D2 is equal to the generated data amount D1 stored in the storage circuit 3. It is easily possible by calculation to allocate the input data D2 so that they are substantially proportional and the sum of the respective target data amounts D2 is equal to the total capacity DALL of the optical disk.

The target data amount determination circuit 4 in FIG. 14 obtains the target data amount in this way based on the information in the generated code amount storage circuit 3. The target data amount for each short section determined by the target data amount determination circuit 4 is stored in the target data amount storage circuit 5.

Next, in the second pass, the same video signal as above is inputted from the terminal 1 in FIG. At this time, FIG.
Then, the target data amount storage circuit 5 sequentially sends the previously calculated target data amount to the encoding circuit 2a. Encoding circuit 2
a is almost the same as that of FIG.
1 is greatly different in that it is configured to receive a command from the target data amount storage circuit 5.

According to such a configuration, a code amount is assigned to each section according to the difficulty of encoding for each short section detected by the first encoding, and encoding compression suitable for a storage medium is performed. Will be done.

[0018]

The prior art described above is
Although this is an excellent encoding and compression system that can efficiently record moving image data on a storage medium such as an optical disk by variable rate control, information on the input video obtained by the first encoding, It is difficult to perform optimal encoding in consideration of human visual characteristics only with target encoding compression conditions obtained by automatic calculation from the result of encoding. In addition, since encoding must be performed at least twice, there is a problem that encoding itself takes time and costs. Furthermore, even if a defect such as an error due to dropout is partially detected in the input video signal after encoding, it is necessary to re-encode the entire encoded video signal. It was inefficient.

[0019]

According to the present invention, the first encoding is performed by one-pass variable rate control, and the first encoding is performed between the first encoding and the second encoding. By providing a step of editing the target encoding compression condition obtained from the result,
Encoding in consideration of human visual characteristics and psychological characteristics is enabled, and the second encoding is performed only on the edited portion of the target encoding compression condition, thereby greatly improving the efficiency of the entire encoding. Further, when the GOP structure and the inverse telecine transform structure of the target encoding condition are changed, the optimal code amount is adaptively predicted from the conditions before and after the change in the target generated code amount in the second encoding. And image quality is improved.

Further, the encoding and compressing method according to claim 1 comprises:
What is claimed is: 1. An encoding / compression method for encoding / compressing a video signal at a variable rate, comprising: encoding a complete encoding in a predetermined coding unit for each predetermined coding unit according to a predetermined first coding parameter; Performing a video signal to generate a first coded stream; and, when generating the first coded stream, for each of the predetermined coding units,
Generating a second encoding parameter of the video signal; specifying a sub-interval of the first encoded stream including the encoding unit; and generating a second sub-interval corresponding to the specified sub-interval. Editing the encoding parameters of the above, and encoding the video signal of the specified sub-interval according to the edited second encoded stream to generate a second encoded stream; Replacing at least a part of the section of the second coded stream with a corresponding section of the first coded stream.

Further, the encoding and compression method according to claim 2 is
2. The encoding and compression method according to claim 1, wherein the first,
The second encoded stream is an MPEG stream, and the predetermined encoding unit is a closed GOP (Group of Group) encoded without referring to preceding and following encoding units.
picture).

Further, the encoding and compression method according to claim 3 is
2. The encoding and compression method according to claim 1, wherein the second encoding parameter is a modulation intensity of a quantization step, a quantization matrix, a pixel level modulation degree, a search range of a motion vector, a motion vector, for the image of the video signal. At least one of a vector detection condition and a target code amount.

Further, according to a fourth aspect of the present invention, there is provided an encoding / compression method comprising:
2. The encoding and compression method according to claim 1, wherein the first,
The second encoded stream is an MPEG stream, and the MPEG stream has a prediction method (prediction).
method), the step of editing the second encoding parameter corresponding to the specified sub-interval comprises editing the MPEG space of a specific picture type of the plurality of different types. And editing only the second encoding parameter.

Further, the encoding and compression method according to claim 5 is
2. The encoding and compression method according to claim 1, wherein the first,
The second encoded stream is an MPEG stream, and the second encoding parameter has a GOP structure.

Further, the encoding and compression method according to claim 6 is characterized in that:
2. The encoding and compression method according to claim 1, wherein the first,
The second encoded stream is an MPEG stream in which a telecine-converted video signal is generated by inverse telecine conversion and includes a copy field, and the second encoding parameter is at least the video signal to be a copy field. Is information for designating the field of (1).

Further, the coding and compression apparatus according to claim 7 is
What is claimed is: 1. An encoding / compression device for encoding / compressing a video signal at a variable rate, wherein said encoding / compression unit completes encoding in a predetermined coding unit for each predetermined coding unit according to a predetermined first coding parameter. A first encoding unit that performs on a video signal to generate a first encoded stream;
When generating the coded stream, the coding parameter generating means for generating a second coding parameter of the video signal for each of the predetermined coding units, and the first code including the coding unit Means for designating a partial section of the encoded stream, editing means for editing the second encoding parameter corresponding to the designated partial section, and a designated section in accordance with the edited second encoded stream. A second encoding unit that encodes the video signal in the partial section to generate a second encoded stream; and encodes at least a part of the section of the second encoded stream with the first code. And a replacement unit for replacing the corresponding section of the converted stream.

Further, the encoding and compression apparatus according to claim 8 is
The encoding and compression device according to claim 7, wherein the first,
The second encoded stream is an MPEG stream, and the predetermined encoding unit is a closed GOP encoded without referring to preceding and following encoding units.

[0028] The encoding and compression apparatus according to claim 9 is
8. The encoding and compression apparatus according to claim 7, wherein the second encoding parameter is a modulation intensity of a quantization step, a quantization matrix, a pixel level modulation degree, a search range of a motion vector, and a motion vector for the image of the video signal. At least one of a vector search condition and a target code amount.

According to a tenth aspect of the present invention, in the encoding and compression apparatus according to the seventh aspect, the first and second encoded streams are MPEG streams, and the MPEG stream is Forecasting method (pred
and editing the second encoding parameter corresponding to the specified sub-interval, wherein the second encoding parameter is composed of a plurality of types of pictures of different types. Is a step of editing only the encoding parameter of (1).

The encoding and compression apparatus according to claim 11 is the encoding and compression apparatus according to claim 7, wherein the first and second encoded streams are MPEG streams, and The parameterization is a GOP structure.

According to a twelfth aspect of the present invention, in the encoding and compression apparatus according to the seventh aspect, the first and second encoded streams are obtained by converting a telecine-converted video signal into an inverse telecine transform. A generated MPEG stream, including a copy field,
Is information for designating at least a field of the video signal to be a copy field.

[0032]

(Embodiment 1) FIG. 16 is a block diagram conceptually showing a basic configuration of the present invention. FIG.
6, reference numeral 160 denotes an encoding / compression device, 161 denotes a first encoding circuit that performs variable length encoding of an input video signal, 162 denotes a divided encoded stream generated by the first encoding circuit 161, And a data storage device 164 for storing coding parameters such as a generated code amount and a quantization width, and a divided coded stream generated by a second coding circuit described later. For the data of the section where the user was not satisfied with the image quality in the segmented coded stream, the coding parameter of the section was read from the data storage device 162 and input from the input means 163. This is an encoding parameter editing circuit that edits with an editing command and outputs parameters when encoding is performed by the second encoding circuit 165.

FIG. 18 is a flowchart for generating a final encoded stream. In FIG. 18, first, a division point is designated to divide an input video signal into predetermined coding units (step S1).
Next, a divided coded stream is generated by the first coding for each of the predetermined coding units (step S2), and at the same time, the modulation intensity of the quantization step, the quantization matrix,
An encoding parameter such as a pixel level modulation degree, a motion vector search range, a motion vector search condition, or a target code amount is generated and stored in the first storage device ST1 (step S3). The operator designates a partial section that did not satisfy the image quality in the first encoding (step S4), and edits the encoding parameters of the partial section (step S5).
For this designated section, a new divided coded stream is generated by the second coding, and the second storage device S
It is stored in T2 (step S6). Further, a divided encoded stream newly generated by the second encoding,
The other encoded streams are selected and combined with the divided encoded streams generated by the first encoding to generate a final encoded stream (step S7) and stored in the third storage device ST3. .

The final encoded stream is shown in FIG.
As shown in the figure, the first encoding circuit 161 generates the divided coded streams of strm_at0 to strm_at3 for the divided sections t0 to t3, and the second coded stream for the section t1 in which the image quality cannot be satisfied. By the encoding circuit 165 of
Suppose that an encoded stream of bt1 is generated. At this time, the final encoded stream output from the data storage device 162 is strm_at0, stre_bt1, strm_at2, strm.
_At3.

FIG. 1 is a block diagram of an encoding and compression apparatus according to Embodiment 1 of the present invention. First, the first encoding will be described. In the encoding circuit 2 in the first encoding, an encoded stream 9 which is an output of the variable length encoding circuit 13 is output from the encoding circuit 2 as a result of the first encoding, and is stored in a data storage device such as a hard disk. It is stored in a device or the like and is input to the generated code amount measuring circuit 27. The generated code amount measuring circuit 27 measures an error between a predetermined average bit rate and a bit rate of an actual coding result in a predetermined period in the past, and determines an error from the predetermined average bit rate as a quantization width. Output to the decision circuit 21. The quantization width determination circuit 21 determines the quantization width from the error from the average bit rate and the predetermined maximum rate and minimum rate so that the error decreases.

For example, when a complex video continues, the coding bit rate generated as a result of actual coding increases, and the error increases in the positive direction, the quantization width is increased. Prevent the error from expanding any further. Conversely, when the actual coding bit rate decreases continuously and the error increases in the negative direction, control is performed so that the quantization width is reduced and the coding bit rate is increased. . At this time, at the same time, the quantization width is determined in consideration of the behavior of the virtual buffer of the decoder so as to satisfy one or both of the predetermined maximum rate and minimum rate.

If the predetermined period in the past is set to a certain length, the quantization width can be considered to be substantially constant locally, and the bit rate can be varied according to the complexity of the image. Thus, a predetermined average bit rate can be satisfied.

In the first encoding, a scene change detected in advance from a specific point of an input video, for example, a start point of each chapter constituting a movie, a difference between frames, or a change in an error value of a motion vector is performed. The generated coded stream is divided at each point or at a specified predetermined periodic point. In other words, encoding is performed so that the GOP up to that point ends in the frame immediately before the input video frame corresponding to the above-described division point, and a new GOP starts from the frame corresponding to the division point. At this time, G to start anew
The OP is a closed GOP that performs encoding without referring to the video signal of the immediately preceding GOP.

The divided encoded stream of variable rate control generated by the first encoding in this manner is stored in a data storage device such as a hard disk device. In addition, encoding parameters such as a generated code amount and a quantization width are similarly stored in a data storage device such as a hard disk device in a form corresponding to the divided encoded stream. Next, the second encoding will be described.

The second encoding is performed on a section where the image quality is not satisfactory in the first encoding, and a new encoded stream is generated again. For other portions, the encoded stream generated in the first encoding can be used as it is.

Next, the editing process of the target encoding condition will be described. The operator designates the section for performing the second encoding from the input means 7 and the quantization width determination circuit 21 in the section.
The amount of correction of the modulation intensity of the quantization step determined by the above can be designated. The modulation intensity correction amount is stored by the correction amount storage circuit 40. Quantization width modulation circuit 4
1 changes the modulation intensity according to the designation.

In general, the modulation of the quantization width is performed according to the complexity (activity) of the input image. The activity of the j-th macroblock is normalized by ACT (j) and the activity value of the previous frame, and N_ACT (j) is calculated by AC_j in the previous picture.
Assuming that the average of T (j) is AVG_ACT and the weighting coefficient in modulation is N, it can be obtained by the following equation. N_ACT (j) = (N * ACT (j) + AVG_AC
T) / (ACT (j) + N * AVG_ACT)

By specifying the value of N at this time, the operator can guide the modulation intensity to an optimum value.
AVG_ACT generally averages ACT (j) in the immediately preceding picture. However, if the average of a plurality of previous pictures is averaged by an operator's specification, the AVG_ACT sharply increases. Even if the complexity changes, it works so as to make the modulation degree of the quantization step gentle, and it is possible to suppress a sudden change in the quantization width. As a result,
It is possible to suppress a sudden change in image quality. The operator can specify not only the modulation intensity but also the modulation itself.

As described above, the target encoding / compression condition is edited. The second encoding is performed on the section including the section where the target encoding / compression condition is edited, using the edited target encoding / compression condition.

As a result, the operator performs the modulation operation of the quantization step in the section where the sufficient image quality cannot be obtained by the automatic assignment, thereby performing the optimum encoding for the human visual characteristic. It becomes possible. Also, by performing the second encoding only on the problematic part, the encoded stream generated by the first encoding can be used as it is for the other parts, and as a total The time required for encoding and compression can be greatly reduced.

(Embodiment 2) FIG. 2 is a block diagram of an encoding and compression apparatus according to Embodiment 2 of the present invention. Compared with FIG. 1, a quantization matrix editing circuit 42 is provided, in which an operator edits a quantization matrix used in the quantization circuit 12a from the input means 7 and edits a quantization width in conjunction with the editing. Are different.

In the second embodiment, as an embodiment,
The first encoding and the second encoding are the same as those of the first embodiment, except for the editing process. Hereinafter, the process of editing the quantization matrix according to the second embodiment will be described.

For example, when the original quantization level is small, and when the quantization step is directly changed, the level greatly changes, and as a result, there is a section where quantization distortion affects human vision. The operator can perform an operation of editing the quantization matrix for the section, reduce the quantization matrix value as a whole, and increase the scale of the quantization step.

Of course, it is also possible to manipulate the quantization matrix value of a particular frequency component. Instead of editing individual quantization matrix values, several quantization matrix value tables are prepared in advance, and the operator selects one of the tables from the input means. This can also be achieved.

Thus, the operator performs the editing operation of the quantization matrix in the section where the sufficient image quality cannot be obtained by the automatic assignment, so that the code particularly suitable for the improvement of the image in the intra-frame compression can be obtained. It is possible to provide a generalized compression device.

(Embodiment 3) FIG. 3 is a block diagram showing an encoding and compression apparatus according to Embodiment 3 of the present invention. Compared with FIG. 1, a pixel level modulation circuit 43 for modulating each pixel level of an input image is added, and a correction amount storage circuit 4
0 is different in that the pixel level modulation circuit 43 is corrected.

In the third embodiment, as an embodiment,
The first encoding and the second encoding are the same as those in the first embodiment except that the editing process is different from that in the first embodiment. The editing process of the pixel level modulation according to the third embodiment will be described below.

In general, human visual characteristics have a low resolution particularly at high and low luminance levels, so that clipping a portion exceeding a certain luminance level has no significant effect. Since the spatial frequency in the band can be suppressed, the amount of bits generated by encoding can be suppressed, and as a result, many bits can be allocated to a place (band) where the bit allocation is desired to be increased. FIG. 4 shows a graph of a modulation circuit having such input / output characteristics.

The designation of the pixel level modulation in the pixel level modulation circuit 43 does not edit the individual level modulation values, but prepares several pixel level modulation value tables in advance, and the operator It can also be realized by selecting one of those tables from the input means.

The editing process according to the third embodiment has been described above. Although the pixel level modulation circuit 43 is provided at a stage prior to the frame rearranging circuit 26a, it may be disposed at a stage subsequent to the frame rearranging circuit 26a, and does not affect the effect of the third embodiment. Thus, the operator can provide an encoding and compression device suitable for intra-frame compression by specifying a pixel level for a section in which sufficient image quality cannot be obtained by automatic assignment. become.

(Embodiment 4) Based on FIGS. 5 and 6,
An encoding and compression apparatus according to Embodiment 4 of the present invention will be described. FIG. 5 is a block diagram of an encoding and compression apparatus according to Embodiment 4 of the present invention. A detection condition changing circuit 44 for changing the detection condition of the motion vector detection circuit 20 as compared with FIG.
In that the operator inputs the detection conditions of the motion vector detection circuit 20 from the input means 7.

FIG. 6 is a block diagram showing an example of the motion vector detecting circuit 20. On the basis of the first motion vector detected by the first motion vector detection circuit 201,
By using a two-stage configuration in which the second motion vector detection circuit 202 further performs motion vector detection, it is possible to perform motion vector detection in a wider range.
The switch 204 makes it possible to forcibly set the motion vector to zero by turning it off. Generally, a motion vector of an image has a high correlation between consecutive frames. Therefore, a motion vector detected in a previous frame is stored in the motion vector storage circuit 203, and
In the next frame, if the reference motion vector of the first motion vector detection circuit 201 is used, the search range is apparently widened, and it is possible to detect a motion vector with higher correlation. When the switch 205 is turned off, use of the reference motion vector can be prohibited.

Also, the first motion vector detecting circuit 201
, A sub-sample of the input and reference images is performed, a first motion vector is detected in a wide range for this sub-sample image, and a second motion vector detection circuit 20
In No. 2, by performing the motion vector detection in detail based on the first motion vector, the motion vector can be detected in a wider range.

In the coding / compression apparatus having the above configuration,
The operator can change the motion vector detection condition in a section where a sufficiently satisfactory image quality cannot be obtained in the first coding, and perform the second coding on the section.

The editing process according to the fourth embodiment will be described. For example, in the case of a rapidly moving image, the first motion vector detection circuit 201 uses the motion vector detected in the previous frame as a reference motion vector, and further uses a sub-sampled image as an input and reference image. If specified, motion vectors can be detected from a wide search range. Conversely, in the case of an image with little motion, the first motion vector detection circuit specifies that the motion vector detected in the previous frame is not used and that no sub-sampling is performed. Although narrow, a highly accurate motion vector search becomes possible. Thus, the operator edits the motion vector detection conditions for a section in which sufficient image quality cannot be obtained by the automatic assignment, thereby reducing spatial and temporal redundancy in motion vector detection. Then, it becomes possible to perform optimal encoding.

(Embodiment 5) FIG. 7 is a block diagram of an encoding and compression apparatus according to Embodiment 5 of the present invention. 1 in that a target data amount editing circuit 45 for editing the target data amount of each short section stored in the target data amount storage circuit 5 is added. The target data amount editing circuit allows the operator to specify a section in which the target data amount is changed and to specify the change amount through the input means 7. The target data amount editing circuit 45 corrects the target data amount of the specified section stored in the target data amount storage circuit 5 by the specified change amount according to the specified amount.

FIG. 8 is a block diagram showing an example of the target data amount editing circuit. The change amount storage circuit 451 includes:
The change amount of the target data amount specified by the operator is stored,
The picture type ratio storage circuit 452 stores the allocation ratio of the target data amount for each of the I, P, and B picture types. The target data amount correction circuit 453 calculates the above change amount,
The target data amount is corrected using the allocation ratio, the GOP structure of the short section held in the target data amount storage circuit 5, and the target data amount.

The change amount storage circuit 451 can specify a change amount for each picture type or a change amount other than a specific picture type. Specify the quantity, P,
By changing the allocation ratio of only I pictures while keeping the allocation ratio between B pictures, the target data amount of only I pictures can be increased or decreased. In addition, 2 of P and B
By specifying the amount of change for each type of picture and appropriately changing the allocation ratio, it is possible to increase or decrease the amount of target data for P and B pictures while maintaining the amount of target data for I pictures. As a result, the operator can determine, for a section in which sufficient image quality cannot be obtained by the automatic assignment,
The target data amount can be corrected, and the spatial and temporal redundancy in setting the target data amount is reduced, so that optimal encoding can be performed.

(Embodiment 6) FIG. 9 is a block diagram showing an encoding and compression apparatus according to Embodiment 6 of the present invention. Compared to FIG. 1, G output from the input means 7 and the encoding circuit 2
A coding parameter storage circuit 62 for storing coding parameters including an OP structure and a coding parameter editing circuit 63 for editing coding parameters are added. A sixth embodiment of the present invention will be described with reference to FIGS.

FIG. 10 shows a GOP structure of M = 3.
FIG. 11 shows a GOP structure of M = 2. FIG.
0 and FIG. 11, (a) shows the original image frame order to be coded and its picture type (I, P or B).
Is shown. (B) shows the encoding processing frame order and its picture type. The M value indicates the cycle of the I and P pictures.

Hereinafter, a sixth embodiment of the present invention will be described. In the first encoding process, the encoding circuit 2 performs the encoding process in the GOP structure of M = 3 shown in FIG. The frame rearranging circuit 26 performs the encoding process by rearranging the order of the original image frames in FIG. 10A in the order of the encoding processing frames in FIG. 10B. The encoding parameters obtained in the first encoding process are stored in the encoding parameter storage circuit 62. The encoding parameter stored in the encoding parameter storage circuit 62 is input by the input means 7
For example, it can be changed by an input process such as designating an M value and a range of the M value. For example, GO of M = 2 shown in FIG.
It can be changed to a P structure. As a result, a frame having a picture type different from the picture type obtained by the first encoding process may occur. For example,
This is the second frame of the original image frame shown in FIGS. 10 and 11 (the B1 frame in FIG. 10 and the I1 frame in FIG. 11).

The result of the changed encoding parameter is output to the target data amount determining circuit 4. The target data amount determination circuit 4 changes and determines at least the target data amount of the frame changed from the picture type at the time of the first encoding process by changing the M value.

There are at least two methods for this determination. The first method is a method that uses the generated code amount obtained in the first encoding process. In the target data amount determination circuit 4, the generated code amount of the frame to be determined is
This is a method of calculating the target data amount based on the data generation amount of the frame closest to the frame with the same picture type in the first encoding process.

In the second method, the first encoding processing is performed on the range including the frame in which the encoding parameter is changed by using the GOP structure data based on the changed M value by the input from the input means 7. In this method, the generated code amount is similarly stored in the generated code amount storage circuit 3, and the target data amount is calculated based on the generated code amount. The target data amount determination circuit 4 determines the target data amount of the second encoding process by the above-described two methods or other methods.

The determined data amount is stored in the target data amount storage circuit 5. The encoding parameter based on the changed M value is stored in the encoding parameter storage circuit 62.

The second encoding process performs encoding based on the encoding parameters from the encoding parameter storage circuit 62 and the data amount from the target data amount storage circuit 5. In the encoding circuit 2a, the frame rearrangement circuit 26 and the encoding control circuit 22 are controlled by data from the encoding parameter storage circuit 62, and the quantization circuit 12 is controlled by data from the target data amount storage circuit 5. .

As described above, the operator sets the GOP for the section where the sufficient image quality cannot be obtained by the automatic assignment.
It is possible to correct the target data amount accompanying the change in the structure (M value), and it is possible to perform encoding that is optimal for human visual characteristics.

(Embodiment 7) When video data is converted between a film of 4 frames / second and a video image of 30 frames / second, telecine conversion for converting video image to film, and film to video A process called inverse telecine conversion for converting to video is performed.

The encoding and compressing apparatus according to the seventh embodiment includes:
In the block diagram of the encoding and compression apparatus shown in FIG. 9, encoding parameters output from the encoding circuit 2 are stored in an encoding parameter storage circuit 62 that stores encoding parameters including information on inverse telecine conversion, The difference is that the encoding parameter is edited by the encoding parameter editing circuit 63 by the input from the input means 7.

Embodiment 7 of the present invention will be described with reference to FIGS. FIG. 12 shows telecine conversion and inverse telecine conversion processing. In FIG. 12, (a)
Is an original film image, and a vertical line indicates one frame of the film. (b), (c) and (d) show video images. In FIG. 12, 1 indicates the top field of the first frame, and 1 ′ indicates the bottom field of the first frame. The same applies to the second and subsequent frames. FIG.
(b) is a video image obtained by telecine-converting the film image of FIG. 12 (a), and is one of the video images input to the encoding device. FIG. 12C shows a result obtained by performing an inverse telecine conversion on the video image of FIG. 12B as a pre-process of the encoding process. In FIG. 12 (c), TFF (Top First Field) and RFF (Repeat First Field) indicate the configuration of a frame that is subjected to inverse telecine conversion and is coded, and TFF = 1 indicates that the top field is time. RFF = 1 indicates that a temporally earlier field is repeated as a copy field. FIG. 12D shows the structure of a frame that is subjected to inverse telecine conversion and is subjected to encoding processing, similarly to FIG. 12C.

The first encoding process is performed by the encoding circuit 2 by an inverse telecine conversion process as shown in FIG. The frame rearranging circuit 26 performs, for example, the inverse telecine conversion shown in FIG. 12C on the video image shown in FIG. That is, it is determined whether or not the input video signal is a copy field in units of fields having the same phase, for example, based on the difference information or the like. Like that. The coding parameters obtained in the first coding process, in the seventh embodiment, the telecine pattern represented by TFF and RFF are:
It is stored in the encoding parameter storage circuit 62.

Next, the editing process according to the seventh embodiment will be described. The telecine pattern stored in the encoding parameter storage circuit 62 has a scene boundary between the fields 3 'and 4, as shown in FIG. In the case where the field is encoded as another frame, the input field 7 is used to input a copy field in FIG.
Is changed to a telecine pattern so as not to become a copy field, that is, the telecine pattern shown in FIG.

As a result, unlike the telecine pattern obtained by the first encoding process, in the case of FIG. 12 (d), the number of frames to be encoded increases, and as a result, the picture in the first encoding process is increased. A case occurs in which a frame of a picture type different from the type occurs. The changed encoding parameter, here the result of the telecine pattern, is output to the target data amount determination circuit 4. The target data amount determination circuit 4 changes and determines at least the target data amount of the frame changed from the picture type at the time of the first encoding process by changing the telecine pattern.

As in the sixth embodiment, there are at least two determination methods. The first method is a method that uses the generated data amount obtained in the first encoding process. The target data amount determination circuit 4 determines the data generation amount of the frame to be determined from the data generation amount of the frame that is the same as the picture type in the first encoding process and is closest to the frame. It is a method of calculating the amount.

In the second method, a GOP based on a changed telecine pattern is applied to a range including a frame whose coding parameter has been changed by an input from the input means 7.
A variable rate encoding process similar to the first encoding process is performed on the structured data, and the generated code amount is similarly stored in the generated code amount storage circuit 3, and the target data amount is set based on the generated code amount. This is the calculation method. Target data amount determination circuit 4
Is the second method described above or other methods.
Is determined for the target data amount of the encoding process.

The determined data amount is stored in the target data amount storage circuit 5. The coding parameter based on the changed telecine pattern is stored in the coding parameter storage circuit 62.

The second encoding process performs encoding based on the encoding parameters from the encoding parameter storage circuit 62 and the data amount from the target data amount storage circuit 5. In the encoding circuit 2a, the frame rearrangement circuit 26 and the encoding control circuit 22 use the data from the encoding parameter storage circuit 62, here the telecine pattern, and the quantization width determination circuit 12 uses the target data amount storage circuit 5 Is controlled by data from

As described above, the operator can correct the target data amount accompanying the change of the telecine conversion pattern for the section where the sufficient image quality cannot be obtained by the automatic assignment. , The spatial and temporal redundancy is reduced, and optimal encoding can be performed.

In the present invention, it is also possible to treat the second encoding result as the first encoding result again, execute the target encoding compression condition editing step of the present invention, and perform the second encoding. It is. Further, the first encoding and the second encoding may be performed by the same encoding circuit. Also, the first
Even if not all of the encoding has been completed, the target encoding / compression condition editing step may be performed and the second encoding may be performed.

Although the editing of the inverse telecine conversion pattern has been described in the seventh embodiment, the editing of the target encoding / compression condition is performed in accordance with the quantization width modulation intensity, the quantization matrix, and the quantization matrix described in the first to sixth embodiments. A plurality of conditions may be edited, including pixel level modulation, motion vector search conditions, and GOP structure.

[0086]

As described above, according to the encoding / compression method and the encoding / compression method according to the present invention, a variable-rate encoded stream divided at a specific division point is generated and the video signal is analyzed. Or a first encoding step of calculating a target encoding compression condition for each of the predetermined encoding units from information obtained as a result of encoding the video signal, and performing partial encoding on the video signal Since a step of editing a parameter related to the target encoding compression condition is provided between the second encoding and the second encoding, it is possible to perform encoding in consideration of human visual characteristics and psychological characteristics. effective.

Further, by performing the second encoding only on the section in which the parameters relating to the target encoding compression condition are edited, it is possible to obtain an effect that the efficiency of the entire encoding can be greatly improved. .

[Brief description of the drawings]

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

FIG. 2 is a block diagram of an encoding and compression apparatus according to a second embodiment of the present invention.

FIG. 3 is a block diagram of an encoding and compression apparatus according to a third embodiment of the present invention;

FIG. 4 is an explanatory diagram of an operation according to the third embodiment of the present invention;

FIG. 5 is a block diagram of an encoding and compression apparatus according to a fourth embodiment of the present invention.

FIG. 6 is a block diagram of a motion vector detection circuit according to a fourth embodiment of the present invention.

FIG. 7 is a block diagram of an encoding and compression apparatus according to a fifth embodiment of the present invention.

FIG. 8 is a block diagram of a target data amount editing circuit according to a fifth embodiment of the present invention.

FIG. 9 is a block diagram of an encoding and compression apparatus according to Embodiments 6 and 7 of the present invention.

FIG. 10 is an operation explanatory view of Embodiment 6 of the present invention.

FIG. 11 is an operation explanatory view of Embodiment 6 of the present invention.

FIG. 12 is a diagram illustrating an operation according to the seventh embodiment of the present invention.

FIG. 13 is a block diagram of a general image encoding circuit.

FIG. 14 is a block diagram of a conventional encoding / compression device.

FIG. 15 is a view for explaining the relationship between the generated data amount and the target data amount.

FIG. 16 is a block diagram conceptually showing a basic configuration of the present invention.

FIG. 17 is a diagram for explaining a basic operation of the encoding / compression device of the present invention.

FIG. 18 is a flowchart of a basic operation performed by the encoding / compression device of the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Input video signal 2 Encoding circuit 2a Encoding circuit 3 Generated code amount storage circuit 4 Target data amount determination circuit 5 Target data amount storage circuit 6 Target data amount editing circuit 7 Input means 9, 9a Encoded stream 11, 11a DCT circuit 12, 12a Quantization circuit 13, 13a Variable length encoding circuit 14, 14a buffer 15 Inverse quantization circuit 21, 21a Quantization width determination circuit 25, 25a Frame buffer 26, 26a Frame rearrangement circuit 27 Generated code amount measurement circuit 40 Correction amount storage circuit 41 Quantization width modulation circuit 42 Quantization matrix editing circuit 43 Pixel level modulation circuit 44 Detection condition changing circuit 45 Target data amount editing circuit 62 Coding parameter storage circuit 63 Coding parameter editing circuit 160 Coding compression device 161 First encoding circuit 162 Data storage device 163 Input means 164 Coding parameter editing circuit 165 Second coding circuit 201 First motion vector detection circuit 202 Second motion vector detection circuit 203 Motion vector storage circuit 204, 205 Switch 451 Change amount storage circuit 452 Picture type ratio Storage circuit 453 Target data amount correction circuit S1 Step for generating coded stream S2 Step for generating coded stream S3 Step for generating coded stream S4 Step for generating coded stream S5 Step for generating coded stream Step S6 Step for Generating an Encoded Stream S7 Step for Generating an Encoded Stream ST1 First Storage Device ST2 Second Storage Device ST3 Third Storage Device

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Kashiwagi ▲ Yoshi ▼ Ichiro 1006 Kadoma, Kazuma, Osaka Prefecture Inside Matsushita Electric Industrial Co., Ltd.

Claims (12)

[Claims] 1. An encoding and compression method for encoding and compressing a video signal at a variable rate, comprising: a code that is completed in a predetermined coding unit for each predetermined coding unit in accordance with a predetermined first coding parameter. Performing a conversion on the video signal to generate a first coded stream; and when generating the first coded stream, a second coding of the video signal for each of the predetermined coding units. Generating an encoding parameter of the following, specifying a sub-interval of the first encoded stream composed of the encoding units, and editing the second encoding parameter corresponding to the specified sub-interval Performing encoding on the video signal of the specified sub-interval according to the edited second encoding parameter.
And a step of replacing at least a part of the section of the second coded stream with a corresponding section of the first coded stream. Compression method. 2. The encoding and compression method according to claim 1, wherein the first and second encoded streams are MPEG streams, and the predetermined encoding unit refers to preceding and succeeding encoding units. Closed GOP encoded without
(Group of picture). 3. The encoding and compression method according to claim 1, wherein the second encoding parameter is a modulation intensity of a quantization step, a quantization matrix, a pixel level modulation degree, and a motion vector for the image of the video signal. Or at least one of a search range, a motion vector detection condition, and a target code amount. 4. The encoding / compression method according to claim 1, wherein the first and second encoded streams are MPEG streams, and the MPEG streams are based on a prediction method (prediction met).
hod), the step of editing the second encoding parameter corresponding to the specified sub-interval includes the step of editing an MPEG stream of a specific picture type of the plurality of different types. A step of editing only the second encoding parameter. 5. The encoding and compression method according to claim 1, wherein the first and second encoded streams are MPEG streams, and the second encoding parameter has a GOP structure. Characteristic encoding / compression method. 6. The encoding and compression method according to claim 1, wherein the first and second encoded streams are MPs obtained by converting a telecine-converted video signal by inverse telecine conversion.
An encoding compression method, comprising: an EG stream, including a copy field, wherein the second encoding parameter is information specifying at least a field of the video signal to be a copy field. 7. An encoding / compression device for encoding / compressing a video signal at a variable rate, wherein the encoding / compression device completes a predetermined coding unit for each predetermined coding unit according to a predetermined first coding parameter. Performing first encoding on the video signal and generating a first encoded stream; and generating the first encoded stream, for each of the predetermined encoding units, Coding parameter generating means for generating a second coding parameter of the video signal; specifying means for specifying a partial section of the first coded stream comprising the coding unit; Editing means for editing the corresponding second encoding parameter; and performing encoding on the video signal of the specified partial section according to the edited second encoding parameter.
A second encoding unit that generates a second encoded stream; and a replacing unit that replaces at least a part of the section of the second encoded stream with a corresponding section of the first encoded stream. An encoding / compression device characterized by the above-mentioned. 8. The encoding and compression apparatus according to claim 7, wherein the first and second encoded streams are MPEG streams, and the predetermined encoding unit refers to preceding and succeeding encoding units. An encoded compression apparatus characterized in that the encoded GOP is a closed GOP encoded without any error. 9. The encoding and compression apparatus according to claim 7, wherein said second encoding parameter is a modulation intensity of a quantization step, a quantization matrix, a pixel level modulation degree, and a motion vector for an image of said video signal. Encoding search apparatus, at least one of a search range, a motion vector search condition, and a target code amount. 10. The encoding and compression apparatus according to claim 7, wherein said first and second encoded streams are MPEG streams, and said MPEG streams are of a plurality of types having different prediction methods. Editing the second encoding parameter corresponding to the specified sub-interval, the second encoding parameter of an MPEG stream of a specific picture type among the plurality of different types. Encoding only the editing step. 11. The encoding and compression apparatus according to claim 7, wherein the first and second encoded streams are MPEG streams, and the second encoding parameter has a GOP structure. Encoding and compression apparatus. 12. The encoding and compression apparatus according to claim 7, wherein the first and second encoded streams are MPs in which a telecine-converted video signal is generated by inverse telecine conversion.
An encoding and compression apparatus, comprising: an EG stream, including a copy field, wherein the second encoding parameter is information specifying at least a field of the video signal to be a copy field.