JP3233232B2 - Moving picture decoding method and apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for decoding data representing a moving picture .

[0002]

2. Description of the Related Art Generally, in a continuous moving image, the preceding and succeeding images and the image of interest are very similar. Therefore, if the image to be coded is a forward prediction coded image, the difference from the image ahead in time is calculated and transmitted. Take the difference with the forward image, take the difference with the temporally backward image, take the difference with the interpolated image created from the temporally forward image and the backward image, and find the smallest of these differences By transmitting the difference, the redundancy in the time axis direction is reduced to reduce the amount of transmitted information.

[0003] In addition, motion compensation is performed to generate a forward prediction coded image and a bidirectional coded image. The motion compensation refers to searching for a portion having the smallest difference in the vicinity of a target block of a previous image in a block unit composed of a plurality of pixels, and calculating a difference therefrom to reduce transmission data.

The difference between the image data obtained as described above is not transmitted as it is, but a discrete cosine transform (DCT) is performed in block units. The DCT expresses not a pixel level of an image but how many frequency components of a cosine function are included. 2D
By the CT, an 8 (line) × 8 (pixel) pixel block is also converted into an 8 × 8 cosine function component coefficient block. In the case of a smooth signal, performing DCT concentrates large values around a certain coefficient. Next, when this coefficient is quantized (divided by a value such as 4 or 32),
The coefficient block of 8 × 8 becomes almost 0, and only a large coefficient remains. Therefore, when transmitting an 8 × 8 coefficient block, the coefficient block is transmitted by a variable length code (VLC) such as a Huffman code or the like, in which a set of non-zero coefficients and 0 runs indicating how many 0s follow before the coefficients.

Each image is composed of at least one or a plurality of slices (this will be described later). Then, they are classified into the following four types according to the encoding scheme. (1) Intra-encoded image At the time of encoding, only information closed by only one image is used. In other words, when decoding, an image can be reconstructed using only the information of the intra-coded image itself. In practice, DCT is performed and encoded as it is, without taking the difference. Random access and high-speed reproduction are possible if the intra-coded image is inserted everywhere. (2) Forward predictive coded image The forward predictive coded image is located at a temporally earlier position as a predictive image (image serving as a reference for calculating a difference) and has already been decoded. Use images. In practice, encoding the difference between the motion-compensated predicted image and encoding the difference without taking the difference (intra-encoding), whichever is more efficient, is called a macroblock (for this, (To be described later). (3) Bidirectional Predicted Coded Image The bidirectional predictive coded image is located earlier in time as a predicted image, and has already been decoded as an intra-coded image or forward-predicted coded image, and is temporally located later. Three types are used: an encoded intra-coded image or a forward predictive-coded image, and an interpolated image made from both. Among these three types of motion-compensated difference coded images and intra-coded images, the most efficient one is selected for each macroblock. (4) DC intra-coded image This is an intra-coded image composed of only DC coefficients of DCT. It does not exist in the same sequence as the other three images.

[0006] A slice is composed of one or a plurality of macroblocks connected in the scanning order of an image. At the beginning of the slice, the difference between the motion vector and the DC component in the image is reset, and the first macroblock has data indicating the position in the image so that it can be restored even if an error occurs. ing. Therefore, the length of the slice and the starting position are arbitrarily changed depending on the error state of the transmission path.

The macroblock is composed of four luminance blocks Y0, Y1, Y2, and Y3 adjacent in the horizontal and vertical directions, and each of the chrominance blocks Cr and Cb at the same position on the image. It consists of. The order of transmission is Y0, Y1, Y2, Y3, C
r and Cb. What to use for the predicted image, whether or not to send the difference, and the like are determined in this unit. A block is composed of 8 × 8 pixels adjacent to each other in luminance or color difference. DCT is performed in this unit.

[0008] A GOP (group of pictures) is composed of one or more intra-coded pictures and zero or more non-intra-coded pictures.

[0009]

[0005] In the conventional MPEG,
Since the high-speed search is performed by relying on the intra-coded images included in one GOP one by one, it is difficult to perform a smooth high-speed search. In order to solve this problem, it is possible to use an image corresponding to a forward prediction coded image as an intra coded image, but in this case, it is possible to perform coding by taking a difference from a previous image. Nevertheless, since the intra-coding process is performed, there is a problem in terms of coding efficiency, and at the time of normal reproduction, the intra-coded image is reproduced as an inter-coded image. A series of images are sequentially reproduced, and a specific image cannot be skipped), which causes a problem that image quality is deteriorated.

A first object of the present invention is to perform a smooth high-speed search of a moving image without deteriorating the image quality of normal reproduction, improve encoding efficiency, and read encoded data. It is to provide a disk that can save time.

A second object of the present invention is to provide a moving picture decoding method and apparatus capable of performing a smooth high-speed search without deteriorating the picture quality of normal reproduction.

[0012]

[0013]

[0014]

According to a first aspect of the present invention, there is provided a moving picture decoding apparatus comprising: a moving picture decoding section which selects picture data selected at a predetermined interval from continuous picture data of a moving picture (for example, frame number 1 in FIG. 5); 5 and the image data for high-speed search which is image data created from the image data (for example, frame numbers 1, 2, 3, 4 in FIG. 5).
And 5), the image data for normal reproduction which is the data created from the high-speed search image data.
The position data of the image data for normal reproduction (for example, FIG.
(A) address 1) is added, and the position data of the high-speed search image data (for example,
A moving picture decoding apparatus for decoding a data string to which an address 2) of FIG. 3B is added, comprising: a mode selection means for selecting one of a high-speed search mode and a normal reproduction mode; When the means selects the high-speed search mode, first processing means for separating the position data of the normal reproduction image data from the high-speed search image data and performing a predetermined decoding process on the high-speed search image data ( For example, the data separation / inverse VLC circuit 8 in FIG. 4) and when the mode selection means selects the normal reproduction mode, the position data of the high-speed search image data is separated from the normal reproduction image data to obtain the normal reproduction image. A second processing unit (for example, a data separation / inverse VLC circuit 4 in FIG. 4) for performing a predetermined decoding process on data is provided.

According to a second aspect of the present invention , in the moving picture decoding apparatus according to the third aspect, the position data of the normal reproduction image data separated by the first processing means is stored. Position data storage means (for example, the position data memory 6 in FIG. 4) for storing the position data of the high-speed search image data separated by the two processing means is further provided, and the first processing means stores the position data in the position data storage means. The high-speed search data at the position indicated by the position data of the obtained high-speed search image data is decoded, and the second processing means is indicated by the position data of the normal reproduction image data stored in the position data storage means. The decoding processing is performed on the image data for normal reproduction at the position to be reproduced.

According to a third aspect of the present invention, in the moving picture decoding method, image data selected at a predetermined interval (for example, frame numbers 1 and 5 shown in FIG. 5) out of continuous moving image data.
), And data created from continuous image data of moving images (for example, data of frame numbers 1, 2, 3, 4, and 5 in FIG. 5). The position data (for example, address 1 in FIG. 3A) of the normal reproduction image data is added to the high-speed search image data, and the normal reproduction image data is A moving image decoding method for decoding a data string to which position data (for example, address 2 in FIG. 3B) of high-speed search image data is added, which is one of a high-speed search mode and a normal reproduction mode When the high-speed search mode is selected, the position data of the normal reproduction image data is separated from the high-speed search image data, Decoding processing (e.g., reverse VL
C), when the normal reproduction mode is selected, the position data of the high-speed search image data is separated from the normal reproduction image data, and the second decoding processing (for example, inverse VLC) is performed on the normal reproduction image data. ) Is performed.

A moving picture decoding method according to a fourth aspect is the moving picture decoding method according to the fifth aspect, wherein the position data of the separated normal reproduction image data is stored, and the stored normal reproduction is stored. A second decoding process is performed on the normal reproduction image data at the position indicated by the position data of the high-speed search image data, and the position data of the separated high-speed search image data is stored, and the stored high-speed search image is stored. The first decoding process is performed on the high-speed search data at the position indicated by the data position data.

[0018]

[0019]

[0020]

[Action] In the video decoding apparatus of the first aspect, when the high-speed search mode is selected, among the consecutive image data of the moving image, created from image data selected at predetermined intervals The position data of the image data for normal reproduction is separated from the image data for high-speed search, a predetermined decoding process is performed on the image data for high-speed search, and when the normal reproduction mode is selected, the continuous image of the moving image is displayed. The position data of the high-speed search image data is separated from the normal reproduction image data created from the data, and a predetermined decoding process is performed on the normal reproduction image data. Thus, after the position data is separated from the image data,
Since the image data is decoded, the quality of the reproduced image is not deteriorated. In addition, since the image data for high-speed search can be prevented from being included in the image data for normal reproduction,
The image quality of normal reproduction does not deteriorate. Also, since only the high-speed search image data can be reproduced collectively, a smooth high-speed search for moving images can be performed. Furthermore, since the image data for normal reproduction can be encoded by taking the difference from the previous image, the encoding efficiency can be improved.

[0021] In the video decoding apparatus of the second aspect, the high speed when the search mode is selected, high-speed search of the location indicated by the location data of the image data for high-speed search which is stored in the position data storage unit When the data is decoded and the normal reproduction mode is selected, the normal reproduction image data at the position indicated by the position data of the normal reproduction image data stored in the position data storage means is decoded. Therefore, the continuity of the image can be ensured between the image reproduced by the high-speed search and the image reproduced by the normal reproduction.

In the moving picture decoding method according to the third aspect , one of the high-speed search mode and the normal reproduction mode is selected. The position data of the normal reproduction image data is separated from the high-speed search image data generated from the image data selected at predetermined intervals, and the first decoding process is performed on the high-speed search image data. When the normal reproduction mode is selected, the position data of the high-speed search image data is separated from the normal reproduction image data created from the continuous image data of the moving image, and the second data is separated from the normal reproduction image data. Decryption processing is performed. As described above, since the image data is decoded after the position data is separated from the image data, the image quality of the reproduced image is not deteriorated. Further, since the high-speed search image data can be prevented from being included in the normal reproduction image data, the image quality of the normal reproduction does not deteriorate. Also, since only the high-speed search image data can be reproduced collectively, a smooth high-speed search for moving images can be performed. Furthermore, since the image data for normal reproduction can be encoded by taking the difference from the previous image, the encoding efficiency can be improved.

In the moving picture decoding method according to the fourth aspect , the second decoding process is performed on the normal reproduction image data at the position indicated by the stored position data of the normal reproduction image data. The first decoding process is performed on the high-speed search data at the position indicated by the position data of the stored high-speed search image data. Therefore, the continuity of the image can be ensured between the image reproduced by the high-speed search and the image reproduced by the normal reproduction.

[0024]

FIG. 1 shows the structure of an embodiment of the disk of the present invention. High-speed search image data is collectively recorded in a recording area A1 on the outer periphery of the disc, and image data for normal reproduction is recorded in an inner recording area A2.

The high-speed search image data is created from image data selected at predetermined intervals from the continuous image data of the moving image, and the normal reproduction image data is created from the image data continuous with the moving image. You. For example, the image data for high-speed search is an intra code generated based on only the frame data of frame number 1 or only based on the frame data of frame number 5 in the series of frame data shown in FIG. Image data. The image data for normal reproduction is created, for example, by taking the difference from the predicted image using all of the series of frame data of frame numbers 1, 2, 3, 4, and 5 shown in FIG. It is composed of a combination of inter-coded image data and intra-coded image data created based on only the frame data of frame number 1 or only based on the frame data of frame number 5. The intra-coded image data is created by, for example, DCT (Discrete Cosine Transform), quantization, and VLC (Variable Length Coding). The inter-coded image data includes, for example, motion compensation, DCT, quantization, and VLC.
Created by

As shown in FIG. 3, the position data (eg, address 1 in FIG. 3A) of the normal reproduction image data corresponding to the high-speed search image data is added to the high-speed search image data. The position data (for example, address 2 in FIG. 3B) of the high-speed search image data corresponding to the data is added to the normal reproduction image data, and the high-speed search image data and the normal reproduction image data are added. Can be associated.

As the above address 1 and address 2,
For example, the sector number of the disk is used. For example, in the normal reproduction image data, a frame number corresponding to a multiple of 30 is added with the sector number of the high-speed search image data corresponding to the position, and the high-speed search image data is closest to the frame. The sector number of the frame of the image data for normal reproduction is added. By associating the high-speed search image data with the normal reproduction image data in this manner, when the user selects the high-speed search,
When the user accesses the high-speed search image data and switches from the high-speed search to the normal reproduction, the user can return to the normal reproduction image data and perform reproduction by relying on the sector number.

FIG. 2 shows one mode of normal reproduction and high-speed search of the disc of FIG. In this example, first, normal reproduction is started at the position P1 of the recording area A2, and at the position P2 of the recording area A2, the normal reproduction is switched to the high-speed search, and the recording area A1 corresponding to the position P2 of the recording area A2 is switched. A high-speed search is performed from the position, and the recording area A
At the position P3 of No. 1, the mode is switched from the high-speed search to the normal reproduction, and normal reproduction is performed again from the position of the recording area A2 corresponding to the position P3 of the recording area A1, and at the position P4 of the recording area A2, the normal reproduction is changed to the high-speed search. To the recording area A1 corresponding to the position P4 of the recording area A2.
, The high-speed search is performed again at the position P5 of the recording area A1, the high-speed search is switched to the normal reproduction, and the normal reproduction is performed again from the position of the recording area A2 corresponding to the position P5 of the recording area A1. In this way, continuity of the image can be ensured between the image reproduced by the high-speed search and the image reproduced by the normal reproduction.

FIG. 4 shows an embodiment of a moving picture decoding apparatus according to the present invention. In this embodiment, the image data for high-speed search is
This is intra-coded image data created by DCT, quantization, and VLC, and the image data for normal reproduction includes intra-coded image data, motion compensation, DCT,
It is configured to consist of quantization and inter-coded image data created by VLC.

When a normal reproduction command is received by a user's key operation or remote control operation, the selector 2 supplies the data read from the disk to the data separation / reverse VLC circuit 4 as normal reproduction data. Also,
When the selector 2 receives a high-speed search command by a user's key operation or remote control operation, the selector 2 supplies the data read from the disk to the data separation and reverse VLC circuit 8 as high-speed search data.

The data separation / inverse VLC circuit 4 separates the image data for normal reproduction, the position data of the image data for high-speed search, the motion vector, and the information for distinguishing between intra and inter from the input data, The image data is inverted VLC and supplied to the selector 12, and the position data of the high-speed search image data is supplied to the position data memory 6,
The motion vector is supplied to a control input of the frame memory 18.

The data separation / inverse VLC circuit 8 separates the high-speed search image data and the position data of the normal reproduction image data from the input data, and performs inverse VLC on the high-speed search image data to perform a selector 12. To supply the position data of the image data for normal reproduction to the position data memory 6. Further, the data separation / inverse VLC circuit 8 stops the function of the motion compensation circuit 16 when the inverse VLC-processed normal reproduction image data is supplied to the selector 12.

The position data memory 6 stores the position data of the high-speed search image data and the position data of the normal reproduction image data. The CPU 10 stores the position data memory 6
A command is given to the driver control unit of the reproducing head so that the reproducing head is positioned at the position on the disk indicated by the position data stored in the reproducing head.

When receiving the normal reproduction command, the selector 12 supplies the inverse VLC-processed normal reproduction image data supplied from the circuit 4 to the inverse quantization and inverse DCT circuit 14, and receives the high-speed search instruction. In this case, the inverse VLC-supplied high-speed search image data supplied from the circuit 8 is supplied to the inverse quantization and inverse DCT circuit 14. The inverse quantization and inverse DCT circuit 14 performs inverse quantization and inverse DCT on the image data supplied from the SELECTAR 12.

The motion compensation circuit 16 reads out a predicted image (previous image) indicated by a motion vector from the frame memory 18 and performs motion compensation on the inversely quantized and inverse DCT image data supplied from the circuit 14. To reproduce the image and output it to a display unit (not shown),
Write to the frame memory 18.

Next, the operation of the embodiment of FIG. 4 configured as described above will be described. First, when a normal reproduction command is received by a key operation or a remote control operation of the user, the CPU 10 sets the disk position indicated by the position data of the image data for normal reproduction stored in the position data memory 6 via the driver control unit. The playback head is positioned and the image data for normal playback is read from the disk.

The read image data for normal reproduction is supplied to the data separation / inverse VLC circuit 4 via the selector 2. The data separation / inverse VLC circuit 4 separates the position data of the high-speed search image data, the motion vector, and the information for distinguishing between intra and inter, which are added to the image data for normal reproduction, and reverses the image data for normal reproduction. VLC
The position data of the high-speed search image data is stored in the position data memory 6, and the motion vector is supplied to the control input of the frame memory 18. Then, the inverse VLC-processed normal reproduction image data passed through the selector 12 is inversely quantized and inversely DCT-processed by the circuit 14, is motion-compensated by the circuit 16, is restored to the original image, and is displayed on the display unit. And at the same time, stored in the frame memory 18.

Next, when a high-speed search command is received by a user's key operation or remote control operation, the CPU 1 stores the disk position indicated by the position data of the high-speed search image data stored in the position data memory 6.
0 positions the playback head via the driver control,
Reads high-speed search image data from the disk.

The read image data for high-speed search is
The data is supplied to the data separation / inverse VLC circuit 8 via the selector 2. The data separation / inverse VLC circuit 8 separates the position data of the normal reproduction image data added to the high-speed search image data, and converts the high-speed sort image data into the inverse VL.
The data is supplied to the selector 12 to store the position data of the image data for normal reproduction in the position data memory 6, and the function of the motion compensation circuit 16 is stopped. The inverse VLC-processed high-speed search image data passed through the selector 12 is inversely quantized and inversely DCT-processed by the circuit 14, displayed on the display unit without motion compensation by the circuit 16, and Stored in the memory 18.

According to the embodiment of FIG. 4, the normal reproduction and the high-speed search can be alternately performed without deteriorating the continuity of the image.

[0041]

[0042]

[0043]

Effects of the Invention] According to the video decoding apparatus according to claim 1,
In the high-speed search mode, the position data of the normal reproduction image data is separated from the high-speed search image data created from the image data selected at a predetermined interval from the continuous image data of the moving image, and the high-speed search is performed. A predetermined decoding process is performed on the image data. In the normal reproduction mode, the position data of the high-speed search image data is separated from the normal reproduction image data created from the continuous image data of the moving image, and the normal reproduction is performed. Since the predetermined decoding process is performed on the image data for use, the image data is decoded after the position data is separated from the image data, so that the image quality of the reproduced image is not deteriorated. Further, since the image data for high-speed search can be prevented from being included in the image data for normal reproduction, the image quality of normal reproduction does not deteriorate. Also, since only the high-speed search image data can be reproduced collectively, a smooth high-speed search for moving images can be performed. In addition, since the image data for normal reproduction can be encoded by taking the difference from the previous image, the encoding efficiency can be improved.

According to the second aspect of the present invention, in the high-speed search mode, the high-speed search data at the position indicated by the position data of the high-speed search image data stored in the position data storage means is decoded. In the normal reproduction mode, the normal reproduction image data at the position indicated by the position data of the normal reproduction image data stored in the position data storage means is decoded. The continuity of the image can be ensured with the image reproduced by the normal reproduction.

According to the third aspect of the present invention, in the high-speed search mode, the high-speed search image data generated from the image data selected at predetermined intervals from the continuous image data of the moving image. The first decoding process is performed on the high-speed search image data by separating the position data of the normal reproduction image data from the normal reproduction image data. In the normal reproduction mode, the normal reproduction image data is created from the continuous image data of the moving image. Since the position data of the high-speed search image data is separated from the image data and the second decoding process is performed on the normal reproduction image data, the image data is decoded after the position data is separated from the image data. Therefore, the image quality of the reproduced image is not impaired. Further, since the high-speed search image data can be prevented from being included in the normal reproduction image data, the image quality of the normal reproduction does not deteriorate. Also, since only the high-speed search image data can be reproduced collectively, a smooth high-speed search for moving images can be performed. Furthermore, since the image data for normal reproduction can be encoded by taking the difference from the previous image, the encoding efficiency can be improved.

According to the moving picture decoding method of the fourth aspect , the second decoding processing is performed on the normal reproduction image data at the position indicated by the stored position data of the normal reproduction image data, and Since the first decoding process is performed on the high-speed search data at the position indicated by the position data of the high-speed search image data, the time between the image reproduced by the high-speed search and the image reproduced by the normal reproduction is obtained. Image continuity can be ensured.

[Brief description of the drawings]

FIG. 1 is a plan view showing a configuration of an embodiment of a disk of the present invention.

FIG. 2 is an explanatory diagram showing one mode of normal reproduction and high-speed search of the disc of FIG. 1;

FIG. 3 is an explanatory diagram showing an example of a format of image data for normal reproduction and image data for high-speed search recorded on the disc of FIG. 1;

FIG. 4 is a block diagram showing a configuration of an embodiment of a moving picture decoding apparatus according to the present invention.

FIG. 5 is an explanatory diagram showing an example of a series of frame data used for creating image data for normal reproduction and image data for high-speed reproduction.

[Explanation of symbols]

 2 selector 4 data separation and inverse VLC circuit 6 position data memory 8 data separation and inverse VLC circuit 10 CPU 12 selector 14 inverse quantization and inverse DCT circuit 16 motion compensation circuit 18 frame memory A1 high-speed search image data recording area A2 normal reproduction Image data recording area P1 Normal playback start position P2 Switching position from normal playback to high-speed search P3 Switching position from high-speed search to normal playback P4 Switching position from normal playback to high-speed search P5 Switching position from high-speed search to normal playback

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ identification symbol FI H04N 7/24 H04N 5/93 B 7/13 Z (58) Investigated field (Int.Cl. ⁷ , DB name) G11B 27 / 00 G11B 20/12 G11B 27/10 H04N 5/92 H04N 5/93 H04N 7/24

Claims (4)

(57) [Claims] 1. High-speed search image data, which is image data created from image data selected at predetermined intervals, among continuous image data of a moving image, and image data created from continuous image data of the moving image. And normal reproduction image data, which is data obtained by the high-speed search.
A moving image decoding apparatus that decodes a data sequence in which the position data of the normal reproduction image data is added, and the position data of the high-speed search image data is added to the normal reproduction image data, Mode selection means for selecting one of a high-speed search mode and a normal reproduction mode; and when the mode selection means selects the high-speed search mode, the position data of the normal reproduction image data is read from the high-speed search image data. First processing means for separating and performing predetermined decoding processing on the high-speed search image data; and when the mode selection means selects a normal reproduction mode, the high-speed search is performed from the normal reproduction image data. Second processing means for separating the position data of the image data for normal use and performing a predetermined decoding process on the image data for normal reproduction. Moving picture decoding apparatus according to claim. 2. Position data for storing the position data of the image data for normal reproduction separated by the first processing means, and for storing the position data of the image data for high-speed search separated by the second processing means. Storage means, wherein the first processing means decodes high-speed search data at the position indicated by the position data of the high-speed search image data stored in the position data storage means, processing means, video decoding according to claim 1, wherein the processing decode the normal reproduction image data of the location indicated by the location data of the normal reproduction image data stored in the position data storage means Device. 3. High-speed search image data, which is image data created from image data selected at predetermined intervals, among continuous image data of a moving image, and image data created from continuous image data of the moving image. And normal reproduction image data, which is data obtained by the high-speed search.
A moving image decoding method for decoding a data string in which the position data of the normal reproduction image data is added and the position data of the high-speed search image data is added to the normal reproduction image data, When one of the high-speed search mode and the normal reproduction mode is selected and the high-speed search mode is selected, the position data of the normal reproduction image data is separated from the high-speed search image data, and the high-speed search image data is separated. Performs the first decryption process on the
A moving image decoding method, comprising separating position data of the high-speed search image data from the normal reproduction image data and performing a second decoding process on the normal reproduction image data. 4. A method for storing position data of the separated normal reproduction image data, wherein the position data of the separated normal reproduction image data is stored in a position indicated by the stored normal reproduction image data. Performs decryption processing of
Storing the position data of the separated high-speed search image data, and performing a first decoding process on the high-speed search data at the position indicated by the stored high-speed search image data; 4. The moving picture decoding method according to claim 3 , wherein: