JP2007060480A - Reverse playback method, playback apparatus, reverse playback program and recording medium with reverse playback program recorded thereon - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform the reverse playback of video/audio streams without increasing a capacity of an image memory buffer, also without extending a decoding means or a video/audio separating means and the like, and without interrupting playback video. <P>SOLUTION: In order to start playback from a final frame in a GOP in reverse playback, a configuration of a frame constituting the GOP is analyzed beforehand, and playback and display are performed. Furthermore, pictures required for reference in picture decoding are not stored nor preserved but the same picture is decoded repeatedly as required, so that any memory buffer for storage is not required and that memory capacity can be reduced. Since it is necessary to perform DeMux of two GOPs if a reference picture exists over the two GOPs like a B1 picture or B0 picture, however, timing is missed. Therefore, GOP DeMux is stopped immediately when the required picture is completely decoded, and DeMux of the next GOP and GOP analysis are started. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The invention according to this patent application (hereinafter simply referred to as “the present invention”) is a video / audio stream in which video information and audio information are multiplexed and encoded (also referred to as encoding), particularly any recording medium (eg, DVD: Reversal playback method, reverse playback apparatus, reverse playback program, and recording medium recording reverse playback program recorded on a digital versatile disk, HDD: hard disk drive, etc. It is.

  A conventional audio / video stream in which video information and audio information are multiplexed and encoded, particularly a reverse reproduction method of an audio / video stream recorded on some recording medium, is described in, for example, Patent Document 1 below. Data reproduction apparatuses, video signal frame advance reproduction processing apparatuses described in Patent Document 2 below, and image decoding apparatuses described in Patent Document 3 below.

  In a conventional video / audio stream backward reproduction method or the like as in the data reproduction apparatus described in Patent Document 1 below, an image having a predetermined number of frames is decoded in advance, stored in a memory buffer, and stored. Are displayed in reverse order. As a result of continuing this display, if the remaining number of images stored / saved in the memory buffer decreases, an image of a predetermined number of frames is decoded again and stored / saved in the memory buffer. Although this method allows a relatively smooth backward reproduction, there is a problem that a large memory buffer is required.

  Also, in the conventional video / audio stream backward playback method, etc., as in the video signal frame advance playback processing apparatus described in Patent Document 2 below, for example, decoding is performed in batches in one GOP unit. These images were stored and saved in a memory buffer, and the saved images were sequentially displayed in the reverse direction. Even with this method, a relatively smooth backward reproduction is possible, but there is still a problem that a large memory buffer is required.

  Also, in the conventional backward reproduction method of video and audio streams, such as in the image decoding device described in Patent Document 3 below, only I and P pictures are stored and saved in a memory buffer and saved. While decoding other pictures (B pictures) using images, the images were displayed in the reverse direction sequentially. In this method, there is a problem that the buffer size becomes large, and in order to reproduce a smooth reverse image without interruption, performance capable of high-speed decoding is required.

  Alternatively, in other conventional video / audio stream backward reproduction methods, only I pictures are reproduced during backward reproduction. This method can be realized with a very simple configuration and does not lead to an increase in cost, but it must be extremely poor in terms of image quality, and reverse playback with a somewhat good image is not possible. There was a problem that could not be expected. This method can be adapted only for the purpose of “what should be displayed can be roughly determined” in reverse playback.

Alternatively, in other conventional video / audio stream reverse playback methods, etc., if an additional decoder or memory buffer is not prepared and a reverse playback is performed while maintaining a certain good image quality, playback is performed. There was a problem that the image had to be cut off. Since this problem is deeply related to the essence of the present invention, it will be described in detail in the detailed description of the present invention in comparison with the embodiment of the present invention.
JP-A-8-32935 JP-A-8-70429 JP-A-10-271444

  As described above, in the backward reproduction method of the conventional audio / video stream, if backward reproduction is performed with good image quality, an additional decoder or memory buffer is required, or these additional components must be added. For example, the image quality may be extremely poor, or the image may be interrupted if reverse playback is performed with a good image quality without adding an additional configuration.

  The present invention has been made in order to solve the above-mentioned problems, and it is possible to reproduce a reproduced video without increasing the capacity of a decoded image memory buffer and without adding decoding means, video / audio separation means, and the like. An object of the present invention is to enable backward reproduction of a video / audio stream with good image quality without interruption.

  In order to solve the above problems, the present invention adopts the following configuration.

The invention according to claim 1 of the present patent application is
A method for backward reproduction of a video stream comprising an image composition group including an independent encoded image, a forward reference encoded image, and a bidirectional reference encoded image,
During the time between the decoding of one image belonging to a certain image composition group and the decoding of other images belonging to the same image composition group, either the video / audio separation process or the image structure analysis process of the next image composition group Or do both,
This is a reverse playback method.

The invention according to claim 2 of the present patent application is
The one image is an independently encoded image;
The reverse reproduction method according to claim 1.

The invention according to claim 3 of the present patent application is
The first image of the image group is an independently encoded image;
A second image of the image composition group is a bi-directional reference encoded image that refers to the independent encoded image and an image of the previous image composition group;
The second image is an image that is reproduced first when reproducing in the forward direction than the first image,
When the decoding for reproducing the first image is completed, the video / audio separation processing of the image composition group including the first image is interrupted,
Starting video / audio separation processing and image configuration analysis processing of the previous image configuration group after the interruption;
The reverse reproduction method according to claim 1.

The invention according to claim 4 of the present patent application is
The video stream is an MPEG2 encoded video stream,
The image composition group is a GOP,
The independent coded image is an I picture;
The forward reference encoded image is a P picture;
The bi-directional reference encoded image is a B picture.
The reverse reproduction method according to claim 3.

The invention according to claim 5 of the present patent application is
The first image is an I2 picture;
The second image is a B0 picture or a B1 picture;
5. A backward reproduction method according to claim 4.

The invention according to claim 6 of the present patent application is
The first image is an I5 picture,
The second image is any one of B0 to B4 pictures;
5. A backward reproduction method according to claim 4.

The invention described in claim 7 according to the present patent application is
A memory buffer for four images as a decoded image memory buffer;
The reverse reproduction method according to claim 3.

The invention described in claim 8 according to the present patent application is
The reverse reproduction is reverse slow reproduction;
The backward reproduction method according to any one of claims 1 to 7.

The invention according to claim 9 of the present patent application is
A playback device for a video stream comprising an image configuration group including an independent encoded image, a forward reference encoded image, and a bidirectional reference encoded image,
The first image of the image group is an independently encoded image;
A second image of the image composition group is a bi-directional reference encoded image that refers to the independent encoded image and an image of the previous image composition group;
The second image is an image that is reproduced first when reproducing in the forward direction than the first image,
During reverse playback of the video stream,
When the decoding for reproducing the first image is completed, the video / audio separation processing of the image composition group including the first image is interrupted,
Starting video / audio separation processing and image configuration analysis processing of the previous image configuration group after the interruption;
This is a video stream playback device.

The invention according to claim 10 of the present patent application is
A computer-executable backward reproduction program for backward reproduction of a video stream including an image configuration group including an independent encoded image, a forward reference encoded image, and a bidirectional reference encoded image,
The first image of the image group is an independently encoded image;
A second image of the image composition group is a bi-directional reference encoded image that refers to the independent encoded image and an image of the previous image composition group;
The second image is an image that is reproduced first when reproducing in the forward direction than the first image,
When the decoding for reproducing the first image is completed, the video / audio separation processing of the image composition group including the first image is interrupted,
Starting video / audio separation processing and image configuration analysis processing of the previous image configuration group after the interruption;
This is a reverse playback program.

The invention described in claim 11 according to the present patent application is
An information recording medium readable by a computer,
An information recording medium on which the backward reproduction program according to claim 10 is recorded.

  The present invention adopts the above-described means, without increasing the capacity of the decoded image memory buffer, without adding decoding means, video / audio separation means, etc., and without interrupting the reproduced video. The audio stream can be played in the reverse direction.

  Embodiments of the present invention will be described below with reference to the drawings.

(First embodiment)
FIG. 1 shows a first embodiment of a reverse reproduction method, reverse reproduction apparatus, reverse reproduction program and reverse reproduction program according to the present invention, in which video information and audio information are recorded. FIG. 2 shows a block configuration diagram of the media player 100 that performs a reverse playback method and the like of multiplexed and encoded video and audio streams.

  The media player 100 records a video / audio stream in which video information, audio information, and other various types of information are multiplexed and encoded in the video / audio stream recording unit 104, and reproduces it. Specifically, the audio / video stream recording unit 104 is, for example, a magnetic medium such as a DVD (digital versatile disk), an HDD (hard disk drive), or a recording medium using various semiconductors such as an SD card. Alternatively, any information recording apparatus capable of recording / storing information by combining optical or electronic principles and these may be used. Further, it is not always necessary that the media player 100 is built in, and for example, some external device may be used.

  Also, for example, MPEG2 encoding, TS (transport stream) multiplexing encoding, or the like can be used as a multiplexing method or encoding method for video information, audio information, and other various types of information. However, the present invention is not limited to these methods. However, any multiplex encoding method may be used.

  The media player 100 is activated by a user operation using, for example, some user interface (not shown), and an operation according to the operation content is executed under the control of the host control unit 102. Alternatively, the user interface itself may be included in the host control unit 102. Furthermore, it is not necessarily operated by the user, but is activated by control from another device or apparatus via a network or communication line (not shown), and the operation according to the control content is executed by control of the host control unit 102. It doesn't matter. This network or network interface or communication line may be included in the host control unit 102. Alternatively, it is not necessarily a user operation or control from other devices, but an operation in accordance with preprogrammed content by detecting some preset condition, for example, the passage of time or the occurrence of some event. May be executed under the control of the host control unit 102. The contents of control performed by the host control unit 102 include that the video / audio stream is input from the outside and recorded in the video / audio stream recording unit 104, and the video / audio stream recorded in the video / audio stream recording unit 104. May be read out and reproduced, and other various processes may be performed.

  FIG. 1 shows the main constituent blocks of the media player 100 for performing these processes, and the broken arrows connecting these main constituent blocks indicate the flow of content information such as a video / audio stream and the solid lines connecting the main constituent blocks. The arrows indicate the flow of control information. However, these main component blocks are merely representative examples, and the media player 100 may include many other components, and content information between these main components may be included. Also, the flow of information and the flow of control information are merely representative, and the media player 100 may include many other content information and control information flows.

  However, since these control contents and control methods are not directly related to the essence of the present invention, detailed description thereof will be omitted. However, the most relevant to the essence of the present invention is to read out and play back the video / audio stream recorded in the video / audio stream recording unit 104, and in particular, to play back the video / audio stream in the reverse direction at low speed (slow) Therefore, only these will be described in detail.

  These various reproductions are controlled by the reproduction control unit 106, which is activated by the control from the host control unit 102.

  When the playback control unit 106 is activated, the playback method, for example, forward playback, reverse playback, high speed playback, normal speed playback, low speed playback, etc. Accordingly, the designated video / audio stream is read from the video / audio stream recording unit 104 in the designated order and sent to the video / audio separation unit 108 to analyze the video / audio stream and to separate the audio stream and the video stream. Executed. However, these control flows and processing contents are merely schematic and simplified examples, other control flows may be performed, and other various controls and processes may be included. . The order of reading out the video / audio stream, the analysis of the video / audio stream, and the separation of the audio and video streams are directly related to the essence of the present invention, and will be described in detail later.

  The audio stream separated from the video / audio stream is sent to the audio decoder 122, decoded into a reproducible audio signal, temporarily stored in the audio memory buffer 126, and sent to the audio output memory 132 at a timing synchronized with the reproduction timing. And reproduced by a speaker (not shown). The audio decoder 122 requires different decoders depending on various audio data encoding methods. Since this is not directly related to the essence of the present invention, detailed description thereof is omitted.

  The video stream separated from the video / audio stream is sent to the video decoder 124, decoded into a reproducible video signal, temporarily stored in the video memory buffer 128, and sent to the video output memory 134 at a timing synchronized with the playback timing. Are reproduced on a display device such as a CRT or a liquid crystal display device (not shown). This video decoder 124 also requires a different decoder depending on the encoding method of various video data.

  Since the video stream recording method, reading order, decoding method, and decoding order are directly related to the essence of the present invention, they will be described in detail with reference to FIGS.

  The present invention is directed to a video stream including an image configuration group including an independent encoded image, a forward reference encoded image, and a bidirectional reference encoded image. That is, the video stream is configured in a state where one piece of image data called a picture, a frame, or a screen is connected, that is, continuously arranged. One piece of image data is image data constituting a picture that is displayed at a time when a video stream is reproduced. For example, one piece of image data is displayed about every 33.3 milliseconds. However, the display speed is not necessarily limited to this.

  This single image data can be classified into an independent encoded image, a forward reference encoded image, and a bidirectional reference encoded image. An independent encoded image is image data that is encoded and decoded using only the image data, and is sometimes called an I picture. The forward reference coded image is image data that is encoded and decoded using the image data and the forward image data, and is sometimes called a P picture. The bi-directional reference encoded image is image data that is encoded and decoded using the image data, the front image data, and the rear image data, and is sometimes called a B picture.

  A plurality of these I pictures, P pictures, and B pictures are collected to form one image configuration group. This image composition group may be referred to as GOP (Group Of Picture), and may also be referred to as VOBU (Video Object Unit), picture composition group, frame composition group, image group, etc. use.

  For example, as shown in FIGS. 2 to 3, one I picture, three P pictures, and eight B pictures can be collected to form one GOP.

  FIG. 2 shows an example of the picture configuration of this GOP.

  As shown in FIG. 2, the video stream is configured in a state in which a plurality of GOPs are connected, that is, continuously arranged. However, in FIG. 2, for convenience of illustration, only one GOP, the GOP immediately before that GOP, and the GOP immediately after that GOP are shown.

  Each GOP is composed of one I picture, three P pictures, and eight B pictures as described above, and the order of recording and reading is, for example, as shown in FIG. 2, I2, B0, B1, This is the order of the pictures P5, B3, B4, P8, B6, B7, P11, B9, and B10. However, I, P, and B indicate an I picture, a P picture, and a B picture, respectively, and the numbers after the alphabet IPB indicate the order of reproduction described later.

  As can be seen from the above, as shown in FIG. 2, when these video streams are reproduced in the forward direction, B0, B1, I2, B3, B4, P5, B6, B7, P8, B9, B10, P11 are arranged in this order. When these video streams are played back in the reverse direction, they are displayed in the order of P11, B10, B9, P8, B7, B6, P5, B4, B3, I2, B1, B0.

  Also, FIG. 3 shows the relationship of pictures used when encoding or decoding each picture. As described above, since the I picture is image data that is encoded and decoded only by the image data, there is no picture used for encoding and decoding.

  Since the P picture is encoded and decoded using the image data and the preceding image data, the P picture is encoded and decoded using the previous I picture or P picture. In the example of FIG. 3, the P5 picture is encoded and decoded using the previous I2 picture, the P8 picture is encoded and decoded using the previous P5 picture, and the P11 picture is the previous P8 picture. Use to encode and decode.

  Since the B picture is encoded and decoded using the image data, the front image data, and the rear image data, the B picture is encoded and decoded using the I picture or P picture before and after the B picture. In the example of FIG. 3, the B3 picture is encoded and decoded using the previous I2 picture and the subsequent P5 picture, and the B4 picture is encoded and decoded using the previous I2 picture and the subsequent P5 picture. The B6 picture is encoded and decoded using the previous P5 picture and the subsequent P8 picture, and the B7 picture is also encoded and decoded using the previous P5 picture and the subsequent P8 picture. Performs encoding and decoding using the previous P8 picture and the subsequent P11 picture, and also encodes and decodes the B10 picture using the previous P8 picture and the subsequent P11 picture.

  The B0 and B1 pictures are encoded and decoded using the P11 picture of the previous GOP and the I2 picture of the GOP to which the B0 picture belongs. The use of the P11 picture of the GOP immediately preceding the GOP to which the B0 picture and the B1 picture belong is most relevant to the essence of the present invention, so particular attention is required.

  Next, the timing and the passage of time when the video stream configured in this way is reproduced in the reverse direction at a low speed of 1/5 will be described with reference to FIGS. Since playback is in the reverse direction, the next GOP shown in FIG. 3 is played back first, then the one GOP shown in FIG. 3 is played, and then the previous GOP shown in FIG. 3 is played back. To do. This designation will be used in the same way in the following description. Similarly, the playback order in one GOP must be P11, B10, B9, P8, B7, B6, P5, B4, B3, I2, B1, B0. This is as described above. Then, assuming that one GOP is going to be reproduced from now on, this GOP is GOP (± 0). GOP (−1) is the GOP immediately preceding this GOP (± 0) (the GOP immediately preceding in the order of recording and reading, and the GOP that is reproduced immediately after the reverse reproduction).

  FIG. 4 is a diagram (part 1) showing the timing and time passage at the time of 1/5 low-speed backward playback, from the start of playback of this GOP (± 0) to the playback of the B9 picture and the decoding of the P8 picture. Show.

  As shown in FIG. 4, when 1/5 low-speed backward reproduction of GOP (± 0) is started, DeMux is first performed. The DeMux refers to the separation of the video stream and the audio stream executed by the video / audio separation unit 108 shown in FIG. 1, and may be demultiplexing, demultiplexer, or the like. This DeMux requires, for example, about 160 milliseconds in one example, and the GOP (± 0) configuration analysis is performed in parallel with DeMux. This GOP configuration analysis is analysis of GOP configuration information (sometimes referred to as frame configuration information, picture configuration information, picture type configuration information, etc.), and an example of each GOP as shown in FIGS. This means analyzing the configuration and creating a header information table or the like representing the configuration. However, the creation of a header information table or the like is an example, and it is not always necessary to create such a table or information. The time required for DeMux is also an example, and is not necessarily limited to this time. .

  Once the GOP (± 0) DeMux is completed and the header information table and the like are created, the GOP (± 0) DeMux is performed again, and according to the contents described in the created header information table, each The picture is decoded. At this time, a method of storing the result of the first DeMux without repeatedly performing the same GOP (± 0) DeMux is conceivable, but this method requires a lot of memory for storage. Repeating DeMux of the same GOP (± 0) is more efficient because it does not require a memory for storage and only uses the same DeMux circuit repeatedly.

  In FIG. 4 (the same applies to the following drawings), the picture to be decoded is displayed in the column “VDec”.

  Since reverse playback is currently being performed, the first picture to be played in this GOP (± 0) is a P11 picture, and a P8 picture that is the reference picture for decoding the P11 picture In order to decode the P8 picture, the reference picture P5 picture must be decoded, and in order to decode the P5 picture, the reference picture I2 picture must be decoded. Since there is no reference picture in the I2 picture, it can be decoded by itself. Therefore, as shown in FIG. 4, in order to decode the P11 picture, decoding is performed in order of the I2 picture, the P5 picture, and the P8 picture, and then the P11 picture is decoded. Each decoded picture is referred to for decoding the next picture, so it must be temporarily stored. The I2 picture is stored in FM (Frame Memory) 0, and the I2 picture stored in FM0 Is used to decode the P5 picture and store it in FM1, use the P5 picture stored in FM1 to decode the P8 picture, store it again in FM0, and use the P8 picture stored in FM0. The P11 picture is decoded and stored in FM2. Then, this P11 picture is transferred to the video output memory 134 (Vout), and playback of the P11 picture is started when this transfer is completed. In this way, the decoding of one GOP (± 0) DeMux and one picture (P11 picture) is completed over approximately 160 milliseconds.

  In addition, since the reverse playback at 1/5 times speed is currently being executed, the P11 picture needs to be displayed for about 167 milliseconds, and for this reason, the P11 picture has about 167 milliseconds. In the meantime, it is stored in Vout.

  Next, similarly, the B10 picture, which is the previous picture (because of backward reproduction), is decoded and reproduced. Similarly to the above, in order to decode the B10 picture, the reference pictures P8 picture and P11 picture must be decoded, and in order to decode the P11 picture, the reference picture P8 picture must be decoded. In order to decode the P8 picture, the reference picture P5 picture must be decoded. To decode the P5 picture, the reference picture I2 picture must be decoded. Since there is no reference picture, it can be decoded by itself.

  Therefore, as shown in FIG. 4, in order to decode the B10 picture, the decoding is performed in order of the I2 picture, the P5 picture, the P8 picture, and the P11 picture, and then the B10 picture is decoded. Each decoded picture is referred to for decoding the next picture, so it must be temporarily stored. The I2 picture is stored in FM0, and a P5 picture is used using the I2 picture stored in FM0. Is decoded and stored in FM1, P8 picture is decoded using P5 picture stored in FM1 and stored again in FM0, and P11 picture is decoded using P8 picture stored in FM0. The B10 picture is decoded using the P8 picture stored in the FM0 and the P11 picture stored in the FM1 and stored in the FM3.

  As can be seen from the above description, FM2 and FM3 are alternately used as the video memory buffer 128. Two frame memories FM0 and FM1 are used as temporary memory buffers for decoding.

  When the display of the P11 picture is completed, the B10 picture is transferred to the video output memory 134 (Vout), and the reproduction of the B10 picture is started from that point. As can be seen from FIG. 4, when the decoding of the B10 picture is completed, the display time of the P11 picture has not ended yet, so a time gap is generated between the display of the P11 picture and the display of the B10 picture. There is nothing. Similarly, the B10 picture is stored in Vout for about 167 milliseconds from this point.

  In the decoding of each picture, the same I2 picture or P5 picture is not repeatedly decoded as described in the description of the previous DeMux, but the once decoded I2 picture or P5 picture is stored and the next picture is decoded. It is also possible to use this method. However, in such a method, at least four frame buffers are required only to store the I2 picture, the P5 picture, the P8 picture, and the P11 picture. In the method of repeatedly decoding the same I2 picture or P5 picture described in the present invention, it is only necessary to store and save the picture immediately before being referred to for decoding, so that there are only two frame buffers FM0 and FM1. The memory for storage can be reduced without using it.

  Although it seems seemingly useless to decode the same I2 picture or P5 picture at first glance, it is efficient and wasteful because it only uses the same decoding circuit repeatedly.

  Next, similarly, the B9 picture, which is the previous picture, is decoded and reproduced. Similarly to the above, in order to decode the B9 picture, the reference pictures P8 picture and P11 picture must be decoded, and in order to decode the P11 picture, the reference picture P8 picture must be decoded. In order to decode the P8 picture, the reference picture P5 picture must be decoded. To decode the P5 picture, the reference picture I2 picture must be decoded. Since there is no reference picture, it can be decoded by itself.

  Therefore, as shown in FIG. 4, in order to decode the B9 picture, the I2 picture, the P5 picture, the P8 picture, and the P11 picture are sequentially decoded, and then the B9 picture is decoded. Each decoded picture is referred to for decoding the next picture, so it must be temporarily stored. The I2 picture is stored in FM0, and a P5 picture is used using the I2 picture stored in FM0. Is decoded and stored in FM1, P8 picture is decoded using P5 picture stored in FM1 and stored again in FM0, and P11 picture is decoded using P8 picture stored in FM0. The B9 picture is decoded using the P8 picture stored in the FM0 and the P11 picture stored in the FM1 and stored in the FM2. When the display of the B10 picture is finished, the B9 picture is transferred to the video output memory 134 (Vout), and playback of the B9 picture is started from that point. As can be seen from FIG. 4, when the decoding of the B9 picture is completed, the display time of the B10 picture has not ended yet, so a time gap is generated between the display of the B10 picture and the display of the B9 picture. There is nothing. Similarly, the B9 picture is stored in Vout for about 167 milliseconds from this point.

  Next, similarly, the P8 picture, which is the previous picture, is decoded and reproduced. This timing is shown across FIGS. 4 to 5 for convenience of illustration.

  Similarly to the above, to decode the P8 picture, the reference picture P5 picture must be decoded, and to decode the P5 picture, the reference picture I2 picture must be decoded, and I2 Since there is no reference picture in the picture, it can be decoded by itself.

  Therefore, as shown in FIGS. 4 to 5, in order to decode the P8 picture, the I2 picture and the P5 picture are sequentially decoded, and then the P8 picture is decoded. Each decoded picture is referred to for decoding the next picture, so it must be temporarily stored. The I2 picture is stored in FM0, and a P5 picture is used using the I2 picture stored in FM0. Is decoded and stored in FM1, and the P8 picture is decoded using the P5 picture stored in FM1 and stored in FM3. The P8 picture is transferred to the video output memory 134 (Vout) when the display of the B9 picture is finished, and the reproduction of the P8 picture is started from that point. As can be seen from FIGS. 4 to 5, when the decoding of the P8 picture is completed, the display time of the B9 picture has not ended yet, so that there is a temporal difference between the display of the B9 picture and the display of the P8 picture. There is no gap. Similarly, the P8 picture is stored in Vout for about 167 milliseconds from this point.

  Next, similarly, the B7 picture, which is the previous picture, is decoded and reproduced. Similarly to the above, in order to decode a B7 picture, the reference pictures P5 picture and P8 picture must be decoded, and in order to decode a P8 picture, the reference picture P5 picture must be decoded. In addition, in order to decode the P5 picture, the reference picture I2 picture must be decoded. Since there is no reference picture in the I2 picture, it can be decoded by itself.

  Therefore, as shown in FIG. 5, in order to decode the B7 picture, the decoding is performed in order of the I2 picture, the P5 picture, and the P8 picture, and then the B7 picture is decoded. Each decoded picture is referred to for decoding the next picture, so it must be temporarily stored. The I2 picture is stored in FM0, and a P5 picture is used using the I2 picture stored in FM0. Is decoded and stored in FM1, and P8 picture is decoded using P5 picture stored in FM1 and stored in FM0 again. P5 picture stored in FM1 and P8 stored in FM0 are stored in FM0. The B7 picture is decoded using the picture and stored in FM2. When the display of the P8 picture is completed, the B7 picture is transferred to the video output memory 134 (Vout), and playback of the B7 picture is started from that point. As can be seen from FIG. 5, when the decoding of the B7 picture is completed, the display time of the P8 picture has not ended yet, so a time gap is generated between the display of the P8 picture and the display of the B7 picture. There is nothing. Similarly, the B7 picture is stored in Vout for about 167 milliseconds from this point.

  Next, similarly, the B6 picture, which is the previous picture, is decoded and reproduced. Similarly to the above, in order to decode the B6 picture, the reference pictures P5 picture and P8 picture must be decoded, and in order to decode the P8 picture, the reference picture P5 picture must be decoded. In addition, in order to decode the P5 picture, the reference picture I2 picture must be decoded. Since there is no reference picture in the I2 picture, it can be decoded by itself.

  Therefore, as shown in FIG. 5, in order to decode the B6 picture, the decoding is performed in order of the I2 picture, the P5 picture, and the P8 picture, and then the B6 picture is decoded. Each decoded picture is referred to for decoding the next picture, so it must be temporarily stored. The I2 picture is stored in FM0, and a P5 picture is used using the I2 picture stored in FM0. Is decoded and stored in FM1, and P8 picture is decoded using P5 picture stored in FM1 and stored in FM0 again. P5 picture stored in FM1 and P8 stored in FM0 are stored in FM0. The B6 picture is decoded using the picture and stored in FM3. When the display of the B7 picture is completed, the B6 picture is transferred to the video output memory 134 (Vout), and playback of the B6 picture is started from that point. As can be seen from FIG. 5, when the decoding of the B6 picture is completed, the display time of the B7 picture has not ended yet, so a time gap is generated between the display of the B7 picture and the display of the B6 picture. There is nothing. Similarly, the B6 picture is stored in Vout for about 167 milliseconds from this point.

  Next, similarly, P5 picture, which is the previous picture, is decoded and reproduced.

  Similarly to the above, in order to decode the P5 picture, the reference picture I2 picture must be decoded. Since there is no reference picture in the I2 picture, it can be decoded by itself.

  Therefore, as shown in FIG. 5, in order to decode the P5 picture, the I2 picture is decoded, and then the P5 picture is decoded. Each decoded picture is referred to for decoding the next picture, so it must be temporarily stored. The I2 picture is stored in FM0, and a P5 picture is used using the I2 picture stored in FM0. Is decoded and stored in FM2. The P5 picture is transferred to the video output memory 134 (Vout) when the display of the B6 picture is finished, and the reproduction of the P5 picture is started from that point. As can be seen from FIG. 5, when the decoding of the P5 picture is completed, the display time of the B6 picture has not ended yet, so a time gap occurs between the display of the B6 picture and the display of the P5 picture. There is nothing. Similarly, the P5 picture is stored in Vout for about 167 milliseconds from this point.

  Next, similarly, the B4 picture, which is the previous picture, is decoded and reproduced. This timing is shown across FIGS. 5 to 6 for convenience of illustration.

  Similarly to the above, in order to decode the B4 picture, the reference picture I2 picture and the P5 picture must be decoded, and in order to decode the P5 picture, the reference picture I2 picture must be decoded. In addition, since there is no reference picture in the I2 picture, it can be decoded by itself.

  Therefore, as shown in FIGS. 5 to 6, in order to decode the B4 picture, the I2 picture and the P5 picture are sequentially decoded, and then the B4 picture is decoded. Each decoded picture is referred to for decoding the next picture, so it must be temporarily stored. The I2 picture is stored in FM0, and a P5 picture is used using the I2 picture stored in FM0. Is decoded and stored in FM1, and the B4 picture is decoded and stored in FM3 using the I2 picture stored in FM0 and the P5 picture stored in FM1. The B4 picture is transferred to the video output memory 134 (Vout) when the display of the P5 picture is finished, and the reproduction of the B4 picture is started from that point. As can be seen from FIG. 5 to FIG. 6, when the decoding of the B4 picture is completed, the display time of the P5 picture has not ended yet. Therefore, there is no temporal difference between the display of the P5 picture and the display of the B4 picture. There is no gap. Similarly, the B4 picture is stored in Vout for about 167 milliseconds from this point.

  Next, similarly, the B3 picture, which is the previous picture, is decoded and reproduced.

  Similarly to the above, in order to decode the B3 picture, the reference picture I2 picture and the P5 picture must be decoded, and in order to decode the P5 picture, the reference picture I2 picture must be decoded. In addition, since there is no reference picture in the I2 picture, it can be decoded by itself.

  Therefore, as shown in FIG. 6, in order to decode the B3 picture, the decoding is performed in order of the I2 picture and the P5 picture, and then the B3 picture is decoded. Each decoded picture is referred to for decoding the next picture, so it must be temporarily stored. The I2 picture is stored in FM0, and a P5 picture is used using the I2 picture stored in FM0. Is decoded and stored in FM1, and the B3 picture is decoded and stored in FM2 using the I2 picture stored in FM0 and the P5 picture stored in FM1. When the display of the B4 picture is finished, the B3 picture is transferred to the video output memory 134 (Vout), and playback of the B3 picture is started from that point. As can be seen from FIG. 6, when the decoding of the B3 picture is completed, the display time of the B4 picture has not yet ended, so that a time gap is generated between the display of the B4 picture and the display of the B3 picture. There is nothing. Similarly, the B3 picture is stored in Vout for about 167 milliseconds from this point.

  Next, similarly, I2 picture, which is the previous picture, is decoded and reproduced. Since there is no reference picture in the I2 picture, it can be decoded by itself.

  Therefore, as shown in FIG. 6, the I2 picture is decoded and stored in FM3. When the display of the B3 picture is completed, the I2 picture is transferred to the video output memory 134 (Vout), and the reproduction of the I2 picture is started from that point. As can be seen from FIG. 6, when the decoding of the I2 picture is completed, the display time of the B3 picture has not ended yet, so a time gap occurs between the display of the B3 picture and the display of the I2 picture. There is nothing. Similarly, the I2 picture is stored in Vout for about 167 milliseconds from this point.

  Next, similarly, the B1 picture, which is the previous picture, is decoded and reproduced. Similarly to the above, in order to decode the B1 picture, the reference picture P11 picture and I2 picture must be decoded. Moreover, since the P11 picture at this time is the P11 picture of GOP (−1), which is the previous GOP, it is necessary to perform GOP configuration analysis of GOP (−1).

  Therefore, as shown in FIGS. 6 to 7, when the DeMux of the GOP (± 0) including the decoding of the I2 picture is completed, the DeMux of the GOP (−1) and the configuration analysis are executed. As described above, this configuration analysis takes approximately 160 milliseconds because it is executed in parallel with one MOP (-1) DeMux. However, the configuration analysis is executed in parallel with the DeMux of the GOP (-1) is only one example, and it is not always necessary to execute the configuration analysis in parallel with the DeMux of the GOP (-1). It may be executed before and after, may be executed after some time, or may be executed at any other time.

  When the DeMux of GOP (-1) and the configuration analysis are completed, the P11 picture of GOP (-1) is decoded. In order to decode the P11 picture, the reference picture P8 picture must be decoded, and in order to decode the P8 picture, the reference picture P5 picture must be decoded, and the P5 picture can be decoded. The I2 picture that is the reference picture must be decoded. Since there is no reference picture in the I2 picture, it can be decoded by itself.

  Therefore, as shown in FIG. 7, in order to decode the P11 picture, the decoding is performed in order of the I2 picture, the P5 picture, and the P8 picture, and then the P11 picture is decoded. Each decoded picture is referred to for decoding the next picture, so it must be temporarily stored. The I2 picture is stored in FM0, and a P5 picture is used using the I2 picture stored in FM0. Is decoded and stored in FM1, P8 picture is decoded using P5 picture stored in FM1 and stored again in FM0, and P11 picture is decoded using P8 picture stored in FM0. And store it in FM1 again.

  Then, returning to decoding of GOP (± 0) again, the I2 picture is decoded and stored in FM0 in order to decode the B1 picture. The B1 picture of GOP (± 0) is decoded and stored in FM2 with reference to the I2 picture stored in FM0 and the previously decoded P11 picture of GOP (-1) stored in FM1 To do.

  However, as shown in FIG. 7, since the GOP (-1) DeMux has been executed twice at this point, the display of the I2 picture has already ended, and the display of the I2 picture and the display of the B1 picture In the meantime, a time gap (discontinuity of the picture) occurs.

  This is a problem when the conventional method does not use an additional frame buffer.

  Therefore, in the present invention, as shown in FIG. 8, when it is necessary to analyze the structure of the previous GOP (−1) for decoding the B1 picture, the decoding of the I2 picture is completed. The DeMux of the GOP (± 0) is aborted (interrupted), and the configuration analysis with the DeMux of the GOP (−1) is executed immediately after that point. This requires about 160 milliseconds. As described above, next, in order to decode the P11 picture of GOP (−1), decoding is performed in order of I2 picture, P5 picture, P8 picture, and P11 picture. ) DeMux is terminated, and the process returns to GOP (± 0) decoding again. In order to decode the B1 picture, the I2 picture is decoded and stored in FM0. The I2 picture stored in FM0 and the previously decoded GOP ( -1) With reference to the P11 picture stored in FM1, the BOP picture of GOP (± 0) is decoded and stored in FM2.

  In this way, when it is necessary to analyze the configuration of the previous GOP GOP (−1) for decoding the B1 picture, the decoding of the necessary picture is completed and the DeMux of the GOP is terminated. As shown in FIG. 8, by immediately starting the structure analysis or decoding of the picture with the DeGux of the next GOP, the display time of the I2 picture is not yet finished when the decoding of the B1 picture is completed. The processing can be executed so that there is no time gap between the display of the B1 picture and the display of the B1 picture.

  Even in the decoding of the next B0 picture, it is necessary to refer to the P11 picture of GOP (−1), which is the previous GOP. Therefore, simultaneously with the decoding of the B1 picture of GOP (± 0), GOP ( (± 0) DeMux is terminated, and decoding of GOP (−1) DeMux and I2 picture, P5 picture, P8 picture, and P11 picture is started immediately.

  When the decoding of the P11 picture is completed, the DeMux of the GOP (-1) is interrupted at the same time as the FM1 is stored, the DeMux of the GOP (± 0) is returned again, the I2 picture is decoded and stored in the FM0, and this FM0 is stored. Using the stored I2 picture and the P11 picture of GOP (−1) previously decoded and stored in FM1, the B0 picture of GOP (± 0) is decoded and stored in FM3.

  In this way, as shown in FIGS. 8 to 9, when the decoding of the B0 picture is completed, the display time of the B1 picture has not yet ended, and there is a time between the display of the B1 picture and the display of the B0 picture. It is possible to execute the processing so that a general gap does not occur.

  Summarizing the processing described above, rather than storing and storing pictures that are required for reference, it is possible to reduce the required memory capacity without the need for a memory buffer for storage, by repeatedly decoding the same picture as needed. However, in this method, when a reference picture exists across two GOPs such as a B1 picture and a B0 picture, the two GOPs must be DeMuxed. Therefore, in order to avoid this, the DeMux of the GOP is stopped immediately after the decoding of the necessary picture is completed, and the DeMux of the next GOP is started.

  In the above description, reverse decoding and playback of B0 picture from P11 picture which is all pictures included in GOP (± 0) which is one GOP is normal, without causing a time gap (discontinuity or delay). Can be terminated.

  Next, the P11 picture, which is a picture included in the previous GOP (-1), moves from the P11 picture to the decoding and playback of the B0 picture. This initial state is shown in FIG. Since it is the same as (± 0), the description is omitted.

  The present invention does not increase the capacity of the decoded image memory buffer, and does not add decoding means, video / audio separation means, etc., and can reproduce the video / audio stream in the reverse direction without interrupting the reproduced video. It becomes possible and its industrial applicability is extremely large.

Block diagram of the media player 100 according to the first embodiment of the present invention The figure which shows an example of the picture structure of GOP The figure which shows the relationship of the picture used (reference) at the time of encoding or decoding The figure which shows the timing and time passage at the time of 1/5 low speed reverse direction reproduction (the 1) The figure which shows the timing and time passage at the time of 1/5 low speed reverse direction reproduction (the 2) The figure which shows the time and the passage of time at the time of 1/5 low speed reverse playback (the 3-past example) The figure which shows the time and the passage of time at the time of 1/5 low speed reverse playback (the 4-conventional example) The figure which shows the timing and the passage of time at the time of 1/5 low speed reverse playback (the 5-example of this invention) The figure which shows the timing at the time of 1/5 low speed reverse direction reproduction | regeneration, and passage of time (the 6-example of this invention)

Explanation of symbols

100 Media Player 102 Host Control Unit 104 Video / Audio Stream Recording Unit 106 Playback Control Unit 108 Video / Audio Separation Unit 122 Audio Decoder 124 Video Decoder 126 Audio Memory Buffer 128 Video Memory Buffer 132 Audio Output Memory 134 Video Output Memory

Claims (11)

A method for backward reproduction of a video stream comprising an image composition group including an independent encoded image, a forward reference encoded image, and a bidirectional reference encoded image,
During the time between the decoding of one image belonging to a certain image composition group and the decoding of other images belonging to the same image composition group, either the video / audio separation process or the image structure analysis process of the next image composition group Or do both,
Reverse playback method. The one image is an independently encoded image;
The reverse reproduction method according to claim 1. The first image of the image group is an independently encoded image;
A second image of the image composition group is a bi-directional reference encoded image that refers to the independent encoded image and an image of the previous image composition group;
The second image is an image that is reproduced first when reproducing in the forward direction than the first image,
When the decoding for reproducing the first image is completed, the video / audio separation processing of the image composition group including the first image is interrupted,
Starting video / audio separation processing and image configuration analysis processing of the previous image configuration group after the interruption;
The reverse reproduction method according to claim 1. The video stream is an MPEG2 encoded video stream,
The image composition group is a GOP,
The independent coded image is an I picture;
The forward reference encoded image is a P picture;
The bi-directional reference encoded image is a B picture.
The backward reproduction method according to claim 3. The first image is an I2 picture;
The second image is a B0 picture or a B1 picture;
The backward reproduction method according to claim 4. The first image is an I5 picture,
The second image is any one of B0 to B4 pictures;
The backward reproduction method according to claim 4. A memory buffer for four images as a decoded image memory buffer;
The backward reproduction method according to claim 3. The reverse reproduction is reverse slow reproduction;
The backward reproduction method according to any one of claims 1 to 7. A playback device for a video stream comprising an image configuration group including an independent encoded image, a forward reference encoded image, and a bidirectional reference encoded image,
The first image of the image group is an independently encoded image;
A second image of the image composition group is a bi-directional reference encoded image that refers to the independent encoded image and an image of the previous image composition group;
The second image is an image that is reproduced first when reproducing in the forward direction than the first image,
During reverse playback of the video stream,
When the decoding for reproducing the first image is completed, the video / audio separation processing of the image composition group including the first image is interrupted,
Starting video / audio separation processing and image configuration analysis processing of the previous image configuration group after the interruption;
Video stream playback device. A computer-executable backward reproduction program for backward reproduction of a video stream including an image configuration group including an independent encoded image, a forward reference encoded image, and a bidirectional reference encoded image,
The first image of the image group is an independently encoded image;
A second image of the image composition group is a bi-directional reference encoded image that refers to the independent encoded image and an image of the previous image composition group;
The second image is an image that is reproduced first when reproducing in the forward direction than the first image,
When the decoding for reproducing the first image is completed, the video / audio separation processing of the image composition group including the first image is interrupted,
Starting video / audio separation processing and image configuration analysis processing of the previous image configuration group after the interruption;
Reverse playback program. An information recording medium readable by a computer,
An information recording medium on which the backward reproduction program according to claim 10 is recorded.

Images