JP2007081744A - Device and method for encoding moving image - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a moving image encoding device capable of obtaining encoding moving image data with the optimum encoding amount and encoding distortion amount under the condition of an assigned target encoding amount, and to provide its method. <P>SOLUTION: The moving image encoding device for encoding an input picture to have the prescribed target encoding amount comprises: a prefilter means for performing filter processing with respect to the input picture with the use of a given filter characteristic, and generating a filter processing picture; an encoding means for performing quantitization processing with respect to the filter processing picture to be outputted from the prefilter means, encoding the picture, and generating encoding data; a local decoding means for performing local decoding processing with respect to the encoding data to be outputted from the encoding means, and generating local decoding data; and a filter coefficient determining means for determining the filter coefficient of the prefilter means, based on the filter processing picture and the local decoding data. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a moving image encoding apparatus and a moving image encoding method, an image transmission apparatus, and an image transmission method that are composed of a prefilter unit and a moving image encoding unit, and particularly to obtain good image quality even at low bit rates. The present invention relates to an image encoding device, an image encoding method, an image transmission device, and an image transmission method. Furthermore, the present invention relates to a control method for prefilter means and moving picture coding means that are optimal for obtaining high image quality.

  Due to dramatic progress in digital signal processing technology in recent years, recording of moving images to a storage medium and transmission of moving images via a transmission path, which have been difficult in the past, are performed. In this case, each picture constituting the moving image is subjected to compression encoding processing, and the data amount is greatly reduced. One typical technique for this compression encoding process is, for example, the MPEG (Moving Picture Experts Group) system.

  When a series of pictures are compression-encoded under a condition of a constant bit rate in accordance with the MPEG system, the amount of codes varies greatly depending on the scene, the spatial frequency characteristics of the pictures, and the quantization scale value. An important technique for minimizing coding distortion in realizing an apparatus having such coding characteristics is code amount control.

  Many algorithms for realizing the code amount control have been proposed so far. Among them, TM5 (Non-Patent Document 1) proposed in the process of standardization of the MPEG-2 encoding method is one well known.

  This TM5 is composed of the following three steps (steps 1 to 3), and the Q scale is controlled so that the bit rate is constant for each GOP.

  [Step 1: Bit allocation]: The target code amount of the picture to be encoded next is obtained from the remaining code amount in the GOP.

  [Step 2: Code amount control]: The Q scale is obtained from the target code amount obtained in Step 1 according to the state of the virtual buffer.

  [Step 3: Adjust Q Scale]: Determine the final Q scale based on the spatial activity of the macroblock.

  Of the three steps, detailed processing of step 1 that has the greatest influence on coding distortion will be described below.

  Now, as shown in FIG. 3, prior to encoding the 10th picture P3 (P picture) in the current GOP, the target code amount of P3 is obtained. The processing of step 1 is expressed by the following equation.

  Here, Rgop represents the amount of code allocated to the current GOP, Ni, Np, and Nb represent the number of remaining pictures in the current GOP for I, P, and B pictures, bits_rate represents the target bit rate, and picture_rate represents the picture rate. Further, for each of the I, P, and B pictures, the picture complexity Xi, Xp, and Xb is obtained from the encoding result by the following equation.

  Ri, Rp, and Rb are the code amounts obtained as a result of encoding I, P, and B pictures, respectively. Qi, Qp, and Qb are Q scale average values for all macroblocks in the I, P, and B pictures, respectively. It is. From the equations (1) and (2), the target code amounts Ti, Tp, and Tb are obtained for each of the I, P, and B pictures using the following equations.

  However, Kp = 1.0 and Kb = 1.4. Based on the target code amount Tp of the picture P3 obtained from the above processing, the Q scale is obtained in step 2 and subsequent steps.

Further, as a code amount control method using a prefilter, a technique described in Patent Document 1 has been proposed. According to this method, the quantization distortion is reduced by controlling the spatial frequency of each picture input to the encoding means by the LPF (hereinafter referred to as prefilter LPF) which is a prefilter.
For the control of the prefilter LPF, a balance function shown in FIG. 4 is defined to match the quantization distortion and the image sharpness deterioration. The two curves in FIG. 4 correspond to the following two functions F1 and F2, respectively. F1 and F2 are
F1 (motion amount, filter coefficient, Q scale, code amount)
F2 (Filter coefficient, Q scale)
It is expressed. The intersection of the functions F1 and F2 is referred to as a balance point, and the Q scale and the filter coefficient of the pre-filter LPF with the best matching of the code amount and image quality are obtained at this point.

Similarly, a method described in Patent Document 2 has been proposed as another conventional technique for controlling the amount of code using a prefilter.
According to this method, first, the encoding difficulty level Y is obtained using the function F for each of the I, P, and B pictures as follows.

Y = F (cumulative code amount, average Q scale)
Next, the filter coefficient parameter Z is obtained from the following equations from the encoding difficulty levels Yi, Yp, and Yb obtained for I, P, and B, respectively.

Depending on the value of the filter coefficient parameter Z obtained by the equation (4), the actual filter coefficient S is selected from the predetermined values S0, S1 or S3 set in advance from the graph shown in FIG. That is, by providing a width to the filter coefficient Z corresponding to each filter coefficient S, a sudden change in the filter coefficient Z is avoided.
TM5 (Test Model Editing Commitee: "Test Model 5", ISO / IEC JTC / SC29 / WG11 / N0400 (Apr. 1993))) Japanese Patent No. 2894137 JP 2002-247576 A

  TM5 has the following problems. In order to obtain the final Q scale in Step 2 and Step 3, only the difference between the target code amount of the picture and the code amount in the picture up to the current macroblock, and the spatial activity of the macroblock are used.

  That is, TM5 attempts to adjust the visual characteristics while actually performing the encoding process after the target code amount of the picture to be encoded has already been determined. Therefore, there is a problem that the degree of quantitative degradation of image quality and human visual characteristics are not sufficiently reflected.

Furthermore, if camera shake or the like occurs when the moving picture encoding apparatus equipped with TM5 is a portable device or the like, it is difficult to sufficiently reduce the amount of information even if motion detection is performed. This is because since the code amount is controlled only by the Q scale, the correlation between pictures becomes low, and the code amount cannot be reduced sufficiently only by motion detection.
FIG. 6 is a graph showing the state of occurrence of camera shake when shooting for 15 seconds in a mobile device, and FIG. 7 shows each picture when MPEG-4 encoding is performed in the situation where the camera shake occurs. This is the transition of the value of the Q scale used in.

  The picture rate is 30 fps and the bit rate is 4 Mbps. In the graph of FIG. 7, the Q scale rapidly increases during the periods of picture Nos. 30 to 60 and 120 to 150. It can be seen that in the period shown in FIG. 6 corresponding to this period, camera shake in the vertical direction is particularly severe. The reason why the Q scale responds greatly to camera shake in the vertical direction is that the search range for motion detection in the vertical direction is narrower than that in the horizontal direction.

  In Patent Document 1, an attempt is made to solve the problem of TM5 by using a pre-filter LPF, but a large-scale circuit is required to calculate a motion amount that is an argument in the function F1. Also, nothing is mentioned about the definition of the functions of F1 and F2 and how to obtain the intersection point, and the control method and effect of the prefilter LPF are unclear.

  Further, Patent Document 2 tries to solve the problem of Patent Document 1 by avoiding an abrupt change when changing the filter coefficient. However, it is difficult to say that the degree of image quality degradation is still taken into account because the filter coefficient is simply predicted only from the information of the accumulated code amount and the average Q scale.

  The present invention has been made in view of the above-described problems of the prior art, and controls the pre-filter and the encoding means while taking into account the degree of image quality degradation, so that the code amount under the condition of the assigned target code amount. And a moving picture coding apparatus and method capable of obtaining coded moving picture data having an optimum coding distortion amount.

  In order to solve the above-described problem, a moving picture coding apparatus according to the present invention is a moving picture coding apparatus that codes an input picture to a predetermined target code amount, and applies a given filter to the input picture. Prefilter means for performing filtering processing according to characteristics to generate a filtered picture, and encoding for generating encoded data by performing quantization processing on the filtered picture output from the prefilter means and encoding A local decoding unit that performs local decoding processing on the encoded data output from the encoding unit and generates local decoded data, and based on the filtering picture and the local decoded data, Filter coefficient determining means for determining the filter coefficient of the pre-filter means.

  The moving picture coding method according to the present invention is a moving picture coding method for coding an input picture to a predetermined target code amount, and performs a filtering process on the input picture with a given filter characteristic. A prefiltering step for generating a filtered picture, a coding step for generating coded data by performing a quantization process on the filtered picture obtained by the prefiltering step, and encoding the coded picture; A local decoding process for performing local decoding processing on the encoded data obtained in the process to generate local decoded data, and a filtering process in the prefiltering process based on the filter processing picture and the local decoded data And a filter coefficient determining step for determining the filter coefficient.

  Further features of the present invention will become apparent from the best mode for carrying out the present invention and the accompanying drawings.

  According to the present invention having the above-described configuration, the pre-filter and the encoding unit are controlled in consideration of the deterioration degree of the image quality. Therefore, the code amount and the encoding distortion amount under the condition of the assigned target code amount. Can obtain the optimal encoded moving image data.

<First Embodiment>
In the first embodiment according to the present invention, LPF is applied as a prefilter, and MPEG-4 encoding is applied as encoding means. FIG. 1 shows a block diagram of a moving picture coding apparatus according to the first embodiment. The moving image encoding apparatus includes an image processing unit 100, an initial value determining unit 104, and a Q scale determining unit 105.

[Configuration of video encoding device]
First, three means constituting the image processing means 100 will be described. The prefilter means 101 is constituted by a one-dimensional convolution LPF of 3, 5, and 7 taps, and the filter coefficient of each tap is set from the filter coefficient means 107 for each picture. The configuration of the prefilter unit 101 is shown in FIG. The filter coefficients correspond to C1, C2, C3, and C4, and all pixels in the picture are processed in the raster scan order. The sum N of the filter coefficients C1 to C4 used for each tap is set to be a power of 2, and the right shift corresponding to the sum N is performed on the variable SEKIWA in FIG. And

  The encoding unit 102 and the local decoding unit 103 correspond to an MPEG-4 encoding unit. The encoding unit outputs an MPEG-4 encoded stream, and the local decoding unit 103 outputs a decoded picture.

  Next, the initial value determining means 104 will be described. Each time an input picture is input to the moving picture encoding apparatus, one of two picture types, I or P picture, is selected as the picture type to be encoded by the encoding means 102. For example, when the number of pictures constituting a GOV (Group of Vop) is always fixed to 15 pictures, the first picture is selected as the I picture and the remaining 14 pictures are selected as the P pictures. Also, the initial value of the Q scale is determined by using TM5 shown in Non-Patent Document 1 described above according to the selected picture type.

  Next, five means constituting the Q scale determining means 105 will be described. The block distortion detecting means 106 uses the technique called spatial differentiation disclosed in Japanese Patent Application 2003-390752 “Image Processing Apparatus and Method, and Computer Program and Computer-Readable Storage Medium” to calculate the block distortion amount for each macroblock. The block distortion amount of the picture is detected by detecting and averaging all MBs in the picture.

  The filter coefficient selection means 107 determines the filter coefficients C1 to C4 of the prefilter means 101 using the block distortion amount input from the block distortion detection means.

  In table holding means 109, a combination table of filter coefficients C1 to C4 and classes is set in advance before inputting the first picture to the moving picture coding apparatus. FIG. 9 shows a table in which there are 27 combinations of filter coefficients C1 to C4, and each of the 27 filter coefficients is divided into three classes. Note that the filter No. is assigned to No. 1 to No. 27 in descending order of the cut-off frequency, and No. 0 is a state in which the filter is off.

  The encoding history holding unit 108 holds the block distortion amount from the block distortion detection unit 106 and the filter characteristic from the filter coefficient selection unit 107 every time the picture encoding is completed. In this embodiment, it is assumed that the block distortion amount and filter characteristics for the last 10 pictures are retained.

  The Q scale determining means 110 adjusts the Q scale value input from the initial value determining means 104 when the I picture is encoded. The Q scale is increased or decreased with reference to the average of filter class numbers used when encoding a plurality of past pictures. For example, when the filter class number is large, a picture with a high degree of encoding difficulty is input. In such a situation, by reducing the code amount distribution to the I picture, the number of pictures can be reduced. The encoding amount is reduced by increasing the amount of code for a large number of P pictures.

[Principle to control code amount by LPF]
The code amount control is realized using LPF as the configuration of the pre-filter means 101. The purpose of this is to reduce the variance of the input picture of the encoding means 102 by applying LPF to the picture input to the encoding means 102. In an encoding method (Lossy) that causes distortion by the classical RD theory, it is known that the code amount Rc follows the following equation (5).

  However, Θc and Ic are coefficients depending on the coding method, Sf is the variance of the input picture, and MSEc is the coding distortion. That is, it can be seen that the code amount Rc is expressed as a variable of the variance Sf when MSEc is constant.

  10 shows the measurement results of the dispersion characteristics of LPFs corresponding to the filters No. 1, No. 6, 15, 23, and 27 shown in FIG. 9, and FIG. 11 shows the frequency characteristics of each. In FIG. 10, the horizontal axis represents the variance of the LPF input pictures, and the vertical axis represents the variance ratio of the input / output pictures. From FIG. 10, it can be seen that the LPF has a characteristic for changing the dispersion in addition to the characteristic for changing the spatial frequency of the input and output pictures. Furthermore, it can be seen that there is a strong correlation between the cut-off frequency of LPF and the dispersion characteristics, and the LPF with a lower cut-off frequency has a smaller dispersion ratio of input / output pictures.

  FIG. 12 shows the result of measuring the code amount and variance when the filter coefficients C1 to C4 of the prefilter means 101 are changed from filter Nos. 1 to 27 when the encoding means 102 has a constant Q scale. Show. From FIG. 12, it can be seen that the amount of code decreases in accordance with the variance of the input picture of the encoding means 102 as shown in the equation (5).

  Therefore, the code amount can be controlled by adaptively changing the filter coefficients C1 to C4 using LPF as the configuration of the prefilter unit 101.

[Filter No. selection method]
FIG. 2 shows a processing flow GetQscale () of the Q scale determining means 110 that is a means for selecting the filter No. and determining the Q scale.

  In step 201, Bt, which is the average of the block distortion amounts Bc of the previous 10 pictures, is calculated. Needless to say, in this embodiment, the number of pictures to be calculated is not limited to ten. By referring to the calculated average Bt, it is possible to predict the encoding difficulty level of the previous 10 pictures. In the case of a scene where Bt is very large, a filter No. with a low cutoff frequency is always selected, and deterioration due to removal of high-frequency components by the pre-filter means 101 rather than visual deterioration caused by block distortion. Will stand out.

  Therefore, by comparing the average block distortion amount Bt with a predetermined constant LIMIT1 in step 202, the prefilter means 101 avoids a situation where a filter number having a low cutoff frequency is always selected. FIG. 13 shows the relationship between the period of the filter class No. 0 during which the operation of the prefilter means 101 is off and the block distortion amount Bc. In this embodiment, LIMIT1 = 500000.

  In step 203, the filter class number is selected. FIG. 14 shows the relationship between the value of the block distortion amount Bc detected at the time of encoding the immediately preceding picture and the filter class number corresponding to Bc. In this embodiment, the lower limit value of filter class No. 1, that is, the lower limit value of filter No. 1 is Bc = 200000, the lower limit value of filter class No. 2 is Bc = 250000, and the lower limit value of filter class No. 3 is Bc = 3000000. Note that filter class numbers that can be selected according to the filter class number used in the immediately preceding picture are limited, and movement between filter classes is allowed only for the adjacent filter class.

  In step 204, one corresponding filter No. is selected from the filter class No. selected according to the block distortion amount Bc of the immediately preceding picture. In this embodiment, the interval between the filter numbers is set in increments of Bc = 5000. That is, there are 10 types of filters No. 1 to 10 in filter class No. 1, and Bc = 200000 to 204999, Bc = 205000 to 209999, ... Bc = 245000 to 249999 in order from filter No. 1. To do.

  In step 205, the average of the filter class numbers of the previous 10 pictures is calculated. In the same manner as in step 201, the scene encoding difficulty level is predicted from the filter number used in the past 10 pictures. Here, the reason for not referring to Bt calculated in step 201 is that the provisional Q scale is adjusted in the next step 206 in accordance with the operation status of the prefilter unit 101.

  In step 206, the provisional Q scale value input from the initial value determining means is adjusted. In the present embodiment, the provisional Q scale is adjusted only for the I picture. This is because when the target code amount calculation means 106 calculates the target code amount of a picture, the I picture allocates a larger amount of code to the P picture, and the current scene is more difficult to encode. This is to avoid a situation where the amount of code allocated to the P picture is insufficient. In this embodiment, if the average of the 10 pictures immediately before the filter class No. calculated in step 205 is filter class No. 3, 3 is added to the temporary Q scale to obtain the Q scale of the encoding means 102, Add 2 to the filter class No. 2 and 1 to the filter class No. 1 to the temporary Q scale.

[Description of processing flow]
FIG. 15 shows the processing flow of this embodiment, and FIG. 16 shows the transition of the Q scale and PSNR [dB] values when the 15-second sequence in which the camera shake shown in FIG. 6 occurs is encoded. The transition of Bc is shown in FIG. The encoding results shown in FIGS. 16 and 17 are encoding results when the bit rate is 4 Mbps. Before the leading picture is input to the moving picture coding processing apparatus of the present embodiment, the processing of step 1400 and step 1401 is performed. In step 1400, the target bit rate is set in the initial value determining means 104. In step 1401, the filter coefficients C1 to C4 corresponding to the filter class No. and the filter No. in FIG.

  Next, in step 1402, the initial value of the Q scale is calculated. As described above, in this embodiment, the initial value of the Q scale is calculated using TM5 of Conventional Technique 1.

  In step 1403, the flow process shown in FIG. 2 described in [Filter No. Selection Method] is performed. In step 1404, the Q scale value obtained in step 1403 is set in the encoding means 102, and in step 1405, the filter number is set in the filter means 100.

  When the processing from step 1400 to step 1405 is completed, the leading picture is input to the prefilter unit 101 in step 1406, and LPF processing corresponding to the filter coefficients C1 to C4 set in the prefilter unit 101 is performed. For example, when the image processing unit 100 is configured by hardware, the encoding unit 102 sequentially performs encoding processing when pixels in the picture output from the prefilter unit 101 are obtained for the macroblock. Alternatively, when the image processing unit 100 is configured by software, the encoding unit 102 may perform the encoding process after the prefilter unit 100 completes the processing for one picture.

  After all the macroblocks in the picture have been processed by the encoding means 102, the local decoding means 103, and the block distortion detection means 109, a picture to be input to the moving picture encoding processing apparatus of the present embodiment is further added. If it is present, the processing from step 1402 is repeated, and if it is the final picture, the processing of the moving image coding processing apparatus of this embodiment is completed.

  When the encoded stream of the same bit rate is generated from the transition of each Q scale picture in FIG. 16, the Q scale value is suppressed to be lower than that when TM5 is used for encoding. I can see that it is made. In particular, this is effective in suppressing a rapid increase in the Q scale during the periods of pictures Nos. 30 to 60 and 120 to 150, which is a period in which camera shake, which is one of the problems of the prior art, is intense.

<Second Embodiment>
When the filter number selection in the filter coefficient selection unit 107 of the first embodiment is adaptively changed according to the value of the block distortion amount Bc detected by the block distortion detection unit 106 when encoding a plurality of past pictures. Will be described.

  Also in the second embodiment, the block configuration of the video encoding apparatus is the same as that in FIG. 1, and the processing flow is the same as that in FIG. The difference from the first embodiment is the processing flow of GetQscale () in step 1403 in FIG.

  FIG. 18 shows a processing flow of GetQscale () in the second embodiment. Similarly to the first embodiment, the table holding means 109 stores filter class numbers and filter coefficients C1 to C4 corresponding to the filter numbers shown in FIG. The initial value of the block distortion amount Bc corresponding to each filter No. is the value used in the first embodiment shown in FIG. 14, and the lower limit value of the filter class No. 1 is the variable BASE1 and the filter class No. The initial value of the lower limit of 2 is the constant LIMIT2. In the second embodiment, LIMIT2 = 250000.

  In step 1701 in FIG. 18, Bt, which is the average of the block distortion amounts Bc of the previous 10 pictures, is calculated. In step 1702, the average value Bt of the block distortion amounts Bc of the previous 10 pictures is compared with the initial value LIMIT2 of the lower limit value of the filter class No.2. If the value of Bt is greater than or equal to LIMIT2, in step 1703, the value of Bt is assigned to BASE1. In the following cases, 200000 is assigned to BASE1 in step 1704. By this comparison processing, the lower limit value of the block distortion amount Bt at which the LPF is activated by the pre-filter means 101 is increased, and a filter number having a low cutoff frequency is always selected particularly in a scene with a high degree of encoding difficulty. To avoid. The subsequent processing in steps 1705 to 1708 is the same as that in the first embodiment.

FIG. 19 is a diagram showing transition of the block distortion amount Bc and the value of BASE1. At point 1, since the value of Bt is less than or equal to LIMIT2, it is determined that the past 10 pictures are difficult to encode, and an initial value of 200,000 is set in BASE1. On the other hand, at point 2, since the Bt value is LIMIT2 or more, the Bt value 280000 is set in BASE1. Similarly, at point 3, since the Bt value is LIMIT2 or more, the Bt value 260000 is set in BASE1.
As described above, an appropriate filter can be selected and set by adaptively setting the Bt value.

<Other embodiments>
In the present invention, a storage medium in which a program code of software for realizing the functions of the embodiments is recorded is provided to a system or apparatus, and a computer (or CPU or MPU) of the system or apparatus is stored in the storage medium. It is also achieved by reading and executing the code. In this case, the program code itself read from the storage medium realizes the functions of the above-described embodiment, and the storage medium storing the program code constitutes the present invention. As a storage medium for supplying such a program code, for example, a floppy (registered trademark) disk, hard disk, optical disk, magneto-optical disk, CD-ROM, CD-R, magnetic tape, nonvolatile memory card, ROM Etc. can be used.

  Further, by executing the program code read by the computer, not only the functions of the above-described embodiments are realized, but also an OS (operating system) running on the computer based on the instruction of the program code Includes a case where the function of the above-described embodiment is realized by performing part or all of the actual processing.

  Furthermore, after the program code read from the storage medium is written to the memory provided in the function expansion board inserted into the computer or the function expansion unit connected to the computer, the function is based on the instruction of the program code. This includes the case where the CPU of the expansion board or function expansion unit performs part or all of the actual processing, and the functions of the above-described embodiments are realized by the processing.

  In addition, by distributing the program code of the software that realizes the functions of the above-described embodiments via a network, the program code is stored in a storage unit such as a hard disk or memory of a system or apparatus or a storage medium such as a CD-RW or CD-R. Needless to say, this can also be achieved by the computer (or CPU or MPU) stored in the system or apparatus reading and executing the program code stored in the storage means or the storage medium.

It is a block diagram which shows the structure of the moving image encoder in 1st Embodiment. It is a figure which shows the processing flow of the Q scale determination means 110 in 1st Embodiment. It is a figure for demonstrating TM5 of a prior art. It is a graph explaining the balance function of a prior art. It is a graph which determines the filter coefficient of a prior art. It is a graph which shows the occurrence situation of camera shake when photographing with a portable device. It is a figure which shows Q scale transition at the time of using TM5. 2 is a diagram showing an internal configuration of prefilter means 101. FIG. 10 is a table showing filter class numbers and filter numbers. It is a graph showing the measurement result of the dispersion characteristic of LPF. It is a graph showing the measurement result of the frequency characteristic of LPF. It is a graph showing the measurement result of the code amount and dispersion ratio by LPF. It is a figure showing the period which selects filter class No.0. It is a figure which shows a response | compatibility with filter class No. and a block distortion amount. It is a processing flow of the moving image encoder in 1st Embodiment. It is a figure which shows the encoding result by 1st Embodiment. It is a figure which shows the encoding result by 1st Embodiment. It is a processing flow of filter No. selection in a 2nd embodiment. It is a figure which shows transition of the block distortion amount in 2nd Embodiment, and the lower limit of filter class No.1.

Claims (12)

A moving image encoding apparatus that encodes an input picture to a predetermined target code amount,
Pre-filter means for generating a filtered picture by performing filtering on the input picture according to a given filter characteristic;
Encoding means for performing quantization processing on the output of the prefilter means and generating encoded data;
Local decoding means for performing local decoding processing on the output of the encoding means and generating local decoded data;
Filter coefficient determining means for determining a filter coefficient of the prefilter means based on the filtered picture and the locally decoded data;
A moving picture encoding apparatus comprising:   2. The moving image according to claim 1, wherein the filter coefficient determination unit detects a block distortion amount from the filtered picture and the local decoded data, and determines the filter coefficient using the block distortion amount. Encoding device. Furthermore, it comprises table means for storing the filter coefficient,
3. The moving picture coding apparatus according to claim 2, wherein the filter coefficient determining unit selects the filter coefficient suitable for correcting the block distortion amount from the table unit according to the block distortion amount. .   The filter coefficient determination means determines the filter coefficient using encoding history information including information on block distortion amounts and filter coefficients of a plurality of pictures preceding at least the encoding order. 4. The moving image encoding apparatus according to any one of 3 above. And a quantization scale adjusting unit that adjusts a quantization scale according to information related to a block distortion amount of a plurality of pictures preceded by the encoding order,
5. The moving picture encoding apparatus according to claim 4, wherein the encoding unit performs the quantization process on the filter processing picture based on a quantization scale obtained by the quantization scale adjustment unit. .   6. The moving image according to claim 4 or 5, wherein the filter coefficient determination unit adaptively determines whether or not to perform the filter processing in the pre-filter unit according to the coding history information. Encoding device. A moving image encoding method for encoding an input picture to a predetermined target code amount,
A pre-filtering step of generating a filtered picture by performing filtering on the input picture according to a given filter characteristic;
An encoding step of performing quantization processing on the filtered picture obtained by the prefiltering step to generate encoded data;
Performing a local decoding process on the encoded data obtained in the encoding step to generate local decoded data; and
A filter coefficient determination step for determining a filter coefficient of the filter processing in the prefiltering step based on the filter processing picture and the locally decoded data;
A moving picture encoding method comprising:   8. The moving image according to claim 7, wherein the filter coefficient determining step detects a block distortion amount from the filtered picture and the locally decoded data, and determines the filter coefficient using the block distortion amount. Encoding method.   9. The moving picture encoding method according to claim 8, wherein the filter coefficient determination step selects the filter coefficient suitable for correcting the block distortion amount according to the block distortion amount.   The filter coefficient determination step determines the filter coefficient using encoding history information including information on a block distortion amount and a filter coefficient of a plurality of pictures preceded by at least the encoding order. 10. The moving picture encoding method according to any one of 9 above. Furthermore, a quantization scale adjustment step of adjusting a quantization scale according to information on block distortion amounts of a plurality of pictures preceded by the encoding order,
The video encoding method according to claim 10, wherein the encoding step performs the quantization process on the filter processing picture based on the quantization scale obtained in the quantization scale adjustment step. .   12. The moving image coding according to claim 10 or 11, wherein the filter coefficient determination step adaptively determines whether or not to perform a filter process in the pre-filter step according to the coding history information. Method.

Images