JP4566821B2 - Apparatus having moving picture data playing function, moving picture playing program, and moving picture data playing method - Google Patents

Description

  The present invention relates to a technique for smoothly performing reproduction processing of a moving image file.

  Since moving image data has a large data capacity, it is often compressed and recorded as a moving image data file on a large-capacity recording apparatus having a movable part such as a hard disk. MPEG is widely used as a compression method for moving image data.

  MPEG moving image data is composed of three types of data: I picture, P picture, and B picture. An I picture is compressed by intraframe coding, and can restore an image by itself. The P picture is compressed by inter-frame prediction using motion compensation prediction based on past frame images. B pictures are compressed by inter-frame prediction using motion compensated prediction based on past or future frame images. When recording or transmitting MPEG moving image data, such I pictures, P pictures, and B pictures are often grouped into a GOP. In many cases, the GOP is further packetized.

  A moving image data reproducing device, editing device, transmitting device, etc. reads a GOP including a target frame from the hard disk and decodes it to restore the target frame.

JP 2003-153261 A

  In a device that reads and plays such a video data file (playback device, editing device, transmission device, etc.), it is important how to shorten the total read time from the hard disk that requires access time due to the movable parts. is there.

  In any apparatus, it is preferable that each frame is played back at the same time as the playback time. However, since processing time for decompressing compressed moving image data is required and access time to the hard disk is required, it may be difficult to maintain the playback speed. In particular, when reproducing in the reverse direction, access to the hard disk in the reverse direction occurs, which is a significant problem. Similarly, in the case of playback of special effects such as transitions, two moving image data files recorded at separate locations must be accessed alternately, and playback time delay is a particular problem. there were. Such a problem is not limited to MPEG moving image data but also to other encoded moving image data.

  Japanese Patent Application Laid-Open No. 2004-228561 discloses a device that has taken measures against this on the assumption that the reproduction time is delayed as described above. However, this does not directly solve the above problem.

  The present invention solves the above problems and provides a device that can quickly obtain a target frame even when encoded moving image data is recorded in a movable recording device. The purpose is to do.

(1) (2) An apparatus having a video playback function according to one aspect of the present invention is a video data file in which frames constituting a video are encoded and grouped, and a video that describes group information of the video data file record the index file, the editing configuration index file to identify the start frame and end frame order information of each video material which is part of the moving picture data file movable recording unit, recording means for accessing the edit configuration data file prediction means for predicting a plurality of frames required based on a part of the moving image data file corresponding to the predicted plurality of frames by the prediction means, reads from said movable recording unit, a frame group of the moving image data files stream cache processing for storing the stream cache recording portion of the non-movable for recording Decoding means, specifying means for specifying a required next target frame based on the editing configuration data file, based on the stream cache means stored moving picture data file, the target frame specified by the specifying means Frame restoring means.

  Prediction means is provided so that a part of the moving image data file corresponding to a plurality of frames is read from the movable recording unit and recorded in the non-movable recording unit, so the total time for acquiring the frame Can be reduced.

(3) In the moving image reproduction program according to one aspect of the present invention, the moving image data file recorded in the movable recording unit groups the encoded plural frames of moving images into group data, and arranges the group data in chronological order. And
The moving image index file describes the recording position of the representative point of each group data in the moving image data file as an access position, and represents the order of the representative frames included in each group data in the entire moving image data file. The stream cache processing means identifies group data including the predicted multiple frames based on the frame order information at both ends of the predicted multiple frames and the representative frame order information of the video index file, and Based on the access position of the group data, the group data is acquired from the movable recording unit.

Since the moving image index file is provided, group data including a plurality of expected frames can be quickly read from the movable recording unit.

(4) In the moving image reproduction program according to one aspect of the present invention , the moving image data file recorded in the movable recording unit is a packet obtained by grouping encoded plural frames of moving images into group data and packetizing the group data. Data is arranged in chronological order, and the moving image index file describes the recording position of the representative point of the packet including each group data in the moving image data file as the access position, and the representative frame included in each group data, Representative frame order information indicating the order of frames in the entire moving image data file is described , and the stream cache processing means a) includes frame order information at both ends of a plurality of predicted frames and representative frame order information in the moving image index file. Based on the expected multiple frames B) acquiring a packet including group data from the movable recording unit based on the access position of the packet including the group data, and c) restoring the group data from the packet to obtain a stream cache recording unit. It is characterized by recording.

Since the moving image index file is provided, group data including a plurality of expected frames can be quickly read from the movable recording unit.

(5) In the video reproduction program according to one aspect of the present invention , when the stream cache processing unit records the group data in the stream cache unit , the stream cache processing unit calculates individual frame order information of at least main picture data constituting the group data. The frame restoring means specifies the picture data necessary for restoring the target frame based on the individual frame information.

  Therefore, the frame restoration means can quickly restore the target frame.

(6) In the video reproduction program according to one aspect of the present invention , when the stream cache processing unit records the group data in the stream cache unit, the stream cache recording unit records at least the main picture data constituting the group data in the stream cache recording unit. The position is recorded in the stream cache recording unit as an individual access position, and the frame restoration means identifies picture data necessary for restoration of the target frame based on the individual frame information.

  Therefore, the frame restoration means can quickly restore the target frame.

(7) In the moving image reproduction program according to one aspect of the present invention , the frame restoration means records one or more frames restored to obtain the target frame in the decoded cache recording unit, and is newly required. If the target frame is recorded in the decoded cache recording unit, it is acquired from this, and if it is not recorded in the decoded cache recording unit, it is acquired from the stream cache recording unit and restored.

  Therefore, the frame restoration means can quickly obtain the most recently used frame.

(8) In the moving image reproduction program according to one aspect of the present invention , the frame restoration means starts from the movable recording unit when the target frame is not recorded in either the decoded cache recording unit or the stream cache recording unit. It is characterized by acquisition and restoration.

  Therefore, it is possible to reliably acquire the target frame.

(9) In the video reproduction program according to one aspect of the present invention , the stream cache processing unit or the prediction unit determines whether or not the predicted plurality of frames are in descending order in the expected reproduction order based on the frame order information. In the descending order, the correction is made in ascending order.

  Therefore, it is possible to quickly extract a plurality of predicted frames from the movable recording unit.

(12) A moving image reproduction method according to one aspect of the present invention is a moving recording unit that records a moving image data file in which frames constituting a moving image are encoded and has a movable unit for access, and a part of the moving image. An editing configuration data file recording unit that records an editing configuration data file that specifies start frame order information and end frame order information of each moving image material, and a stream cache recording unit that records a frame group of the moving image data file are accessible. A method of reproducing a moving image by a computer, wherein a plurality of frames required based on the editing configuration data file are predicted by a prediction unit, and a moving image data file corresponding to the predicted plurality of frames by the prediction unit is predicted by a stream cache processing unit A part of the The target frame is specified based on the editing configuration data file by the specifying unit, and the target frame is determined based on the moving image data file stored in the stream cache unit by the stream restoring unit. And the display unit displays the frame restored by the frame restoration means.

  Since a part of the moving image data file corresponding to a plurality of frames is collectively read from the movable recording unit and recorded on the non-movable recording unit, the total time for acquiring the frames can be reduced. .

  In the present invention, the “movable recording unit” refers to a recording device that requires a mechanical operation to access recorded contents, and includes a hard disk, a CD-ROM drive, a DVD drive, and the like. In the embodiment, the hard disk 2 corresponds to this.

  The “editing configuration data file recording unit” records at least editing configuration data, regardless of whether it is movable or non-movable. In the embodiment, the hard disk 2 and the memory 34 correspond to this.

  The “stream cache recording unit” is a unit that records at least a part of a moving image data file and is a non-movable type. Here, the “non-movable recording unit” refers to a recording device that essentially does not require a mechanical operation to access the recorded content, and is a concept including a RAM, a ROM, a USB memory, a memory card, and the like. In the embodiment, the memory 34 corresponds to this.

  In the embodiment, “prediction means” corresponds to step S102 in FIG.

  In the embodiment, “stream cache processing means” corresponds to step S103 in FIG.

  In the embodiment, “identifying means” corresponds to step S104 in FIG.

  “Frame restoration means” is a concept that includes not only a frame obtained by restoring an encoded frame but also a frame that is already restored and recorded. In the embodiment, step S105 in FIG. 16 corresponds to this.

  The “frame order information” is at least data that can clarify the composition order of frames, and is a concept including a frame number, a reproduction time from the start of a file of the frame, and the like.

  The “display unit” refers to a unit that displays the restored frame. In the embodiment, the display 32 corresponds to this.

  The “means for accessing the recording unit” is means for reading the content of the recording unit or recording the content in the recording unit. In the embodiment, steps S101 and S103 in FIG. 16, step S204 in FIG. Step S301 in FIG. 18, step S404 in FIG. 19, and the like correspond to this.

  The “program” is a concept that includes not only a program that can be directly executed by the CPU, but also a source format program, a compressed program, an encrypted program, and the like.

BEST MODE FOR CARRYING OUT THE INVENTION

1. Overall Configuration FIG. 1 is a functional block diagram of a moving image playback apparatus according to an embodiment of the present invention. In the movable recording unit 2, a moving image index file 4, a moving image data file 6, and an editing configuration data file 12 are recorded.

  The moving image data file 6 is obtained by grouping compressed moving images of a plurality of frames into group data, and arranging the group data in chronological order. The moving image index file 4 is provided corresponding to the moving image data file 6. In the moving image index file 4, group information is described for each group data. In the group information, the recording position of the representative point of each group data in the moving image data file 4 is the access position, and the order of the group start frame and group last frame included in each group data in the entire moving image data file 6 frame. The first last frame order information shown is described.

  The editing configuration data file 12 determines a moving image to be reproduced by specifying start frame order information and end frame order information of each moving image material that is a part of the moving image data file 6.

  The prediction means 8 refers to the editing configuration data file 12, and based on the frame number currently being processed and the playback mode (forward playback or reverse playback, etc.), a block of frames (predicted frames) ).

  The stream cache processing means 10 identifies group data including the predicted multiple frames based on the frame order information of the frames at both ends (first and last) of the predicted multiple frames and the order information of the representative frames in the index file 4.

Further, based on the access position of the group data described in the index file 4, the group data including the expected plural frames of the moving image data file 6 is accessed. The acquired group data 16 including the expected plural frames is recorded in the non-movable recording unit 14.

  At this time, the stream cache processing means 10 obtains the access position of the group data 16 from the index file 4, converts it into the access position (recording position) of the non-movable recording unit 14, and records it. Further, the access position (in the non-movable recording unit 14) of each frame included in the group data is calculated and recorded as the individual access position 20. Further, the first and last frame order information of the group data 16 is obtained from the index file 4, and based on this, the frame order information of each frame included in the group data is calculated and recorded as the individual frame order information 18. .

The specifying unit 22 refers to the editing configuration data file 12 and specifies the frame order information of the target frame to be processed next in the moving image data file 6. The frame restoring unit 24 decodes the target frame in the group data 16 based on the individual frame order information 18 and the individual access position 20 of the non-movable recording unit 14 and gives the decoded frame to the display unit 26. The display unit 26 displays the decoded frame.

2. Hardware Configuration FIG. 2 shows a hardware configuration of the moving image playback apparatus. A display 32, a memory 34, a hard disk 2, a keyboard / mouse 36, and a CD-ROM drive 38 are connected to the CPU 30. The display 32 is for displaying the decoded moving image. The keyboard / mouse 36 is for inputting an operation for reproduction. The memory 34 is used as a work area for the CPU 30.

  The hard disk 2 stores an operating system 25 such as WINDOWS (trademark), a moving image data file 6, an editing configuration data file 12, a reproduction processing program 28, and the like. The reproduction processing program 28 reproduces the moving image data file 6 according to the description of the editing configuration data file 12 in cooperation with the operating system 25. The index creation program 5 creates an index file 4 corresponding to the moving image data file 6.

  The reproduction processing program 28 and the index creation program 5 are the ones recorded on the CD-ROM 40 and installed via the CD-ROM drive 38. The moving image data file 6 is recorded with moving image data from a digital video camera or the like. In this embodiment, an MPEG elementary stream is recorded as the moving image data file 6.

FIG. 3 shows the structure of the moving image data file 6. A plurality of GOPs are recorded in chronological order. There are cases where a sequence header SH is provided immediately before the GOP and cases where a sequence header SH is not provided. The sequence header SH records information such as a pixel aspect ratio necessary for reproduction. In each GOP, an I picture, a B picture, and a P picture are recorded. A start code 20 is recorded at the head of the GOP. Note that the GOP includes moving image data for ten or more frames.

3. Creation of index file As a preparation before reproduction, an index file 4 of the moving image data file 6 is created. 4 and 5 show flowcharts of the index creation program 5.

  First, the CPU 30 clears the GOP counter, the frame number counter, and the like. Next, in step S1, the CPU 30 reads the moving image data file 6 from the hard disk 2 from the top. In step S2, it is determined whether or not the end of the moving image data file 6 has been reached. If the end has not been reached, the CPU 30 searches for the beginning of the GOP. As shown in FIG. 3, a start code is present at the head of the GOP, and this is found. Note that the start code of the GOP starts with “000001B8”, so that it can be determined that this is the head of the GOP by detecting this.

  If the read portion is not the head of the GOP (step S3), the process returns to step S1 to further read the moving image data file 6. If the head of the GOP is found (step S3), the GOP counter is incremented (step S4). Next, the CPU 30 determines whether or not there is a sequence header SH immediately before the GOP (step S5). Since the start code indicating the start of the sequence header SH starts with “000001B3”, this can be detected.

  If there is a sequence header SH, the CPU 30 acquires the recording position of the start code of the SH header (the number of bytes from the beginning of the moving image data file (hereinafter the same)) and records it in the memory 34 as the access position ( Step S6). If there is no sequence header SH immediately before, the GOP start code recording position is acquired and stored in the memory 34 as an access position (step S7).

  Next, the CPU 30 acquires the time code (elapsed time from the beginning of the moving image) described in the start code of the GOP and the distinction between the closed GOP and the broken link, and stores them in the memory 34 (step S8). If the image data of the GOP can be restored without referring to the image data of the previous GOP, it is a closed GOP, and if it cannot be restored without referring to the image data of the previous GOP, it is a broken link Is recorded in the start code. The CPU 30 acquires this and records it in the memory 34.

  Next, the CPU 30 reads the moving image data file 6 and acquires the data of the GOP (step S9). The acquired GOP data is analyzed, and the number of frames is counted by a frame number counter (step S10). The CPU 30 analyzes the GOP as follows and acquires the number of frames included in the GOP.

  The CPU 30 finds the picture start code (00000100) in order from the top of the GOP. The picture start code is provided at the head of the I picture, B picture, and P picture. Therefore, the presence of an I picture, B picture, and P picture can be known by detecting the picture start code.

  For example, when an I picture, a B picture, and a P picture each correspond to one frame, one picture start code corresponds to one frame.

  In some cases, two I pictures (B picture, P picture) constitute one frame image. In such a case, two picture start codes correspond to one frame.

  As shown in FIG. 3, following the picture start code, “type” data indicating whether the picture is an I picture, a B picture, or a P picture, and whether one frame is constructed with the one picture (non-interlaced) ) “Distinction” data is recorded on whether one frame is constructed by two pictures (interlace). The CPU 30 refers to this “discrimination” data, and recognizes one frame with one picture start code in the case of non-interlace, and recognizes one frame with two picture start codes in the case of interlace. .

  On the other hand, the frame number counter counts the number of frames from the first frame of the moving image. The CPU 30 records the first frame number (count value of the frame number counter) of the GOP in the memory 34 as the first frame number.

  Next, the CPU 30 determines whether or not the frame has been analyzed to the end of the GOP (step S11). In this embodiment, the CPU 30 determines whether the frame has been analyzed to the end depending on whether the next GOP has appeared. If it is not the last, step S9 and subsequent steps are repeatedly executed. If the GOP is restored to the end, the count value of the frame number counter at that time is stored in the memory 34 as the last frame number (step S12).

  Next, the CPU 30 outputs the stored contents of the memory 34 to the index file 4 (step S13).

  FIG. 6 shows the structure of the index file 4. The total number of GOPs and the total number of frames are recorded after the processing for all the GOPs in the moving image data file 6 is completed. Here, the GOP ID (shown as GOP1 in the figure), the access position, the top frame number, the last frame number, the time code, and the type (closed GOP or broken link) are based on the recorded contents of the memory 34. To be recorded. The GOP ID is a serial number assigned to each GOP in order to identify the GOP.

  Next, the CPU 30 returns to step S1 and performs the same processing for the next GOP. Thereby, the access position, the top frame number, the last frame number, the time code, and the type are recorded for GOP2 in FIG.

  When this process is repeated and the end of the moving image file is reached (step S2), the CPU 30 records the total number of frames and the total number of GOPs in the index file 4 (step S15). In this way, the index file 4 is generated.

  In this embodiment, in order to associate the moving image data file 6 with the index file 4 created for the moving image data file 6, both files have the same file name and different extensions. For example, when the moving image data file is aaa.mpg, the index file is aaa.imv. Thereby, it can be easily determined that the moving image data file having the same file name is associated with the index file.

  In the above-described embodiment, the case where the moving image data file 6 has the elementary stream (ES) format has been described. As the moving image data file 6, there is a PES format in which elementary streams are packetized as shown in FIG. A video stream start code is attached to each packet of the PES, and data obtained by subdividing the ES is described as content. For example, the packet P1 records data of the α part of ES. Further, the packet P2 records the data of the β part of ES.

  Flow charts of processing for generating an index file for such a PES format moving image data file 6 are shown in FIGS. Steps similar to those in FIGS. 4 and 5 are given the same reference numerals.

  In order to determine whether or not the CPU 30 is at the head of the GOP, it is necessary to first solve the packet. Therefore, step S25 for solving a packet is executed before step S3.

  In FIG. 4, the recording position of the start code of SH or GOP is acquired and recorded in the index file (steps S6 and S7). However, in FIG. 9, the recording position of the video stream start code of the packet including the SH or GOP start code is acquired and recorded in the index file (steps S65 and S75). For example, when the SH of the ES in FIG. 7 is found, the head position of the packet P2 including this is recorded as the access position. Thus, by recording the recording position of the start code of the packet in the index file, it is possible to know which packet should be accessed to obtain a desired frame without solving the packet at the time of editing.

  In step S85, the PTS (time information) recorded in the start code of the packet including the SH or GOP start code is acquired and recorded in the index file.

  In addition, before counting the number of frames in step S10, a process of solving the packet and acquiring GOP data is performed (step S95).

  The data structure of the index file 4 generated by the above processing is shown in FIG. Basically, it is the same as in FIG. 6 except that the packet position is recorded as the access position.

  Note that a PS (program stream) including both PES of audio data and PES of moving image data can be created in the same manner as described above with an index file for only PES of moving image data.

  In the above embodiment, the case where the moving image data file 6 is in the form of packetized elementary stream (PES) has been described. As shown in FIG. 11, the moving picture data file 6 also has a transport stream (TS) format in which PES is further packetized. Each packet of the TS is assigned a packet ID, and data obtained by subdividing PES is described as the content.

  Flow charts of a program for generating an index file for such a PES format moving image data file 4 are shown in FIGS. Steps similar to those in FIGS. 4, 5, 8, and 9 are given the same reference numerals.

  In order to determine whether or not it is the head of the GOP, the CPU 30 first needs to solve the TS packet to obtain the PES packet (step S255), and further solve the PES packet (step S25). Therefore, before step S3, step S255 for solving the TS packet and step S25 for solving the PES packet are executed.

  In FIG. 8, the recording position of the video stream start code of the PES packet including the SH or GOP start code is acquired and recorded in the index file (steps S65 and S75). However, in FIG. 12, the recording position of the TS packet including the video stream start code of the PES packet including the SH or GOP start code is acquired and recorded in the index file (steps S655 and S755). For example, when the ES SH in FIG. 11 is found, the head position of the PES packet P2 including this is found, and the head position of the TS packet γ including this head position is recorded as the access position. By recording the recording position of the start code of the TS packet in the index file in this way, it is possible to know which TS packet should be accessed to obtain a desired frame without solving the packet at the time of editing.

  In step S855, the PTS (time information) recorded in the PES in the TS packet at the access position is acquired and recorded in the index file.

  Also, before counting the number of frames in step S10, processing is performed to obtain the PES packet by solving the TS packet, and further obtaining the GOP data by solving the PES packet (steps S955 and S95).

  The data structure of the index file 4 generated by the above processing is shown in FIG. This is basically the same as FIGS. 6 and 10 except that the position of the TS packet is recorded as the access position.

  In each of the above embodiments, the case where the moving image data file 4 is ES, PES (PS), or TS has been described. However, as shown in FIG. 15, it may be determined whether the acquired moving image data file 6 corresponds to any of the above (step S50), and an appropriate process may be selected to create the index file 4. .

When there are a plurality of moving image data files 6, an index file 4 is generated for each moving image data file 6.

4). Reproduction Process Next, the reproduction process will be described. A flowchart of the reproduction processing program is shown in FIG. Here, a case where the moving image data file 6 has a structure as shown in FIG. 3 (ES) will be described.

  In step S <b> 101, the CPU 30 reads the editing configuration data file 12 from the hard disk 2 and records it in the memory 34. The structure of the editing configuration data file 12 is shown in FIG. As shown in FIG. 21, the editing configuration data file 12 is a file that defines clips (moving image materials) C1, C2, and C3 from the moving image data file 6 and defines edited moving images.

  In FIG. 20, timeline information 200 shows information for each clip (moving picture material). The information of each clip includes AVI file name 202, IN point information 204, OUT point information 206, start position time code 208 on the timeline, audio level information 210, audio video link information 212, and playback speed information 214. included.

  The AVI file name 202 is the file name of the moving image data file 6. The IN point information 204 is a position on the main moving image data file 6 where the clip starts. The OUT point information 206 is a position on the moving image data file 6 where the clip ends. A start position time code 208 on the timeline indicates a reproduction timing when reproducing the head of the clip. That is, the reproduction timing of the first frame of the clip when reproducing the edited moving image is represented by the elapsed time from the reproduction start. Audio level information 210 indicates the playback level of audio information. The audio video link information 212 is information for synchronizing the moving image and the audio. The playback speed information 214 indicates the playback speed.

  FIG. 21 shows the relationship among the moving image data file 6, the desired edited moving image, and the editing configuration data. As shown in the drawing, when a part of the moving image data file 6 is extracted and combined as clips C1, C2, and C3 to make an edited moving image, the editing configuration data is configured as shown in the drawing. The start point a1 (frame number in the moving image data file 6) and end point a2 (frame number in the moving image data file 6) of the clip C1 are described in the first clip description portion of the timeline information of the editing configuration data 15. Further, the frame number b1 in the edited moving image is recorded in the time code description portion at the start position on the timeline. The same applies to the clips C2 and C3.

  Note that the edited moving image is specified by the moving image data file 6 and the editing configuration data file 12, and is not recorded independently as an edited moving image data file. By doing so, it is possible to perform editing without modifying the moving image data file 6 only by correcting the editing configuration data file 12.

  Next, the CPU 30 predicts necessary plural frames to be cached based on the editing configuration data file 12 recorded in the memory 34 (step S102). A detailed flowchart of this prediction processing is shown in FIG.

  The CPU 30 creates a list of necessary plural frames based on the editing configuration data file 12 (step S201). For example, it is assumed that the description of the IN point and the OUT point of the timeline information 200 of the editing configuration data file 12 is as shown in FIG. According to the configuration data file 12, it can be seen that in the normal playback mode, playback is performed in the order of the clips CL1, CL2, CL3, and CL4. Note that the clip CL3 is a clip in which the frame advances in the reverse direction in the normal playback mode.

  Here, it is assumed that the clip CL1 (frame numbers 1 to 10) has already been cached. In this case, the CPU 30 determines the frame range to be cached next as follows. Here, it is assumed that normal reproduction is instructed. Also in the case of backward reproduction or the like, the prediction direction is only reversed, and prediction can be performed by the same method.

  The number of frames to be cached is predetermined, for example, 20 frames. Based on the timeline information of the editing configuration data file 12, the CPU 30 sets 20 frames after the already cached frames as necessary plural frames. In the case of FIG. 22, the frames 15 to 25 of the clip CL2 and the frames 30 to 20 of the clip CL3 are set as a plurality of necessary frames. The CPU 30 records this in the memory 34 as a range list as shown in FIG. 23A (step S201). In FIG. 23A, two items are provided in the range list. When the size of the target clip is large, there is one item.

  Next, when the IN point is larger than the OUT point in each item of the range list, the CPU 30 switches the IN point and the OUT point (step S202). Therefore, the range list is rewritten as shown in FIG. 23B.

  The CPU 30 rearranges the items in the range list so that the frame numbers of the IN points are in ascending order. This is to eliminate the head seek in the returning direction when accessing the hard disk 2 later. In the range list of FIG. 23B, since the order is already in ascending order, no rearrangement is performed.

  Next, the CPU 30 refers to the index file 4 based on the IN point and the OUT point for each item in the range list, identifies the GOP of the moving image data file 6 including the necessary multiple frames, and sets the access position. Obtain (step S204). For example, if the contents of the index file 4 are as shown in FIG. 24, GOP1 and GOP2 are specified for the first item (frame numbers 15 to 25), and GOP2 is specified for the second item (frame numbers 20 to 30). Is identified. That is, for the first item, it is understood that it is sufficient to read from the access position of GOP1 to the last position of GOP2 (immediately before the access position of GOP3). As for the second item, it is understood that it is sufficient to read from the access position of GOP2 to the last position of GOP2 (immediately before the access position of GOP3).

  Subsequently, the CPU 30 collects continuous or overlapping access ranges into one (step S205). In the above case, since the access ranges for GOP2 are overlapped, these are collected and used as the access range from the access position of GOP1 to the last position of GOP2 (immediately before the access position of GOP3). In addition, when the access range of two items is continuous, it is set as one continuous access range. This is to increase the access speed.

Next, the CPU 30 determines whether or not the access range of the necessary plural frames calculated as described above has already been cached in the stream cache recording unit (memory 34). If there is a cached range, the part is removed from the access range of the necessary plural frames (step S206). For example, in the above example, if GOP1 is cached in the stream cache recording unit ( memory 34 ), only GOP2 is set as a necessary plurality of frames. That is, the access range is from the access position of GOP2 to the last position of GOP2 (immediately before the access position of GOP3). As described above, a GOP including necessary plural frames (necessary plural GOP) is predicted, and an access range is acquired.

  Returning to FIG. 16, in step S103, the CPU 30 reads the predicted plurality of necessary GOPs from the moving image data file 6 of the hard disk 2 and records them in the stream cache recording unit (memory 34). A detailed flowchart of this stream cache processing is shown in FIG.

  The CPU 30 accesses the moving image data file 6 on the hard disk 2 based on the access position at the head of the access range calculated in the prediction process, and reads out the necessary GOP (step S301).

  The CPU 30 records the contents of the read GOP in the stream cache recording unit (step S302). Here, the contents of the GOP are recorded as they are without being decoded. The CPU 30 acquires the frame number of the first frame of the GOP based on the index file 4 at the time of the above recording (may be acquired in step S102). Further, at the time of the recording, the presence of an I picture, P picture, and B picture constituting the GOP is determined, the frame number of each frame is individually calculated, and the individual frame number is calculated (step S303).

  Further, the CPU 30 acquires the access position of the GOP based on the index file 4 at the time of recording (may be acquired in step S102). This access position is converted into an access position in the stream cache recording unit (memory 34). Further, at the time of recording, the existence of an I picture, P picture, and B picture constituting the GOP is determined, and the access position of each frame in the stream cache recording unit (memory 34) is calculated individually, and the individual access position is calculated. (Step S303).

  The CPU 30 records the individual frame number, the individual access position, and the picture type (discrimination between P, B, and I pictures) calculated as described above in the memory 34 as cache index data. FIG. 25 shows an example of data in the cache index file. In this embodiment, individual frame numbers and individual access positions are calculated and recorded for all frames, but this may be performed only for main frames.

  Next, the CPU 30 determines whether or not all necessary multiple GOPs have been processed (step S304). In the example of FIG. 23, the necessary plurality of GOPs are finally grouped into one continuous group, but depending on the case, they may be divided into a plurality of groups. In such a case, the above processing is repeated for all necessary plural GOPs.

  Returning to FIG. 16, in step S <b> 104, the CPU 30 specifies a frame to be decoded next based on the editing configuration data file 12 and sets it as a target frame. That is, the frame number of the target frame is specified. Since the CPU 30 records which frame has been decoded and transferred to the reproduction buffer, the CPU 30 can specify the frame to be decoded next.

  Next, the CPU 30 performs a process of decoding and restoring the target frame (step S105). A detailed flowchart of this frame restoration process is shown in FIG.

  The CPU 30 determines whether or not the target frame is recorded in the decoded cache (memory 34) (step S401). If it is recorded in the decoded cache, it is read out and the target frame is acquired (step S402).

  If not recorded in the decoded cache, the CPU 30 searches for the target frame from the stream cache recording unit (step S403). The CPU 30 determines whether there is a target frame by determining whether there is a frame number of the target frame based on the individual frame number described in the cache index file (see FIG. 25) of the memory 34. Can do. If the target frame is in the stream cache recording unit, the CPU 30 acquires the type of the target frame described in the cache index file.

  If the target frame is an I picture, the stream cache recording unit is accessed based on the individual access position of the target frame, data is acquired, and the target frame is decoded (step S404).

  If the target frame is not an I picture, the individual access position of the I picture immediately before the target frame is acquired from the cache index file, the stream cache recording unit is accessed, data is acquired, and the I frame is decoded. The subsequent frame is decoded and this is continued until the target frame. In this way, the target frame is decoded (step S404). The frame decoded in step S404 is recorded in the decoded cache (memory 34) (step S405).

  In step S403, there is a case where the target frame does not exist in the stream cache recording unit (when prediction is lost). In this case, the CPU 30 refers to the index file 4 of the hard disk 2 and specifies the GOP for obtaining the target frame. This can be determined based on the frame number of the target frame, the first frame number of the GOP, and the last frame number. The access position of the identified GOP is acquired (step S406).

  Next, the CPU 30 accesses the moving image data file 6 on the hard disk 2 based on the access position, and acquires the GOP. A target frame is obtained by decoding from the first frame of the GOP (step S407). Note that the frame decoded in step S407 is recorded in the decoded cache (memory 34) (step S405).

  Returning to FIG. 16, the CPU 30 outputs the target frame obtained as described above to a reproduction buffer (not shown) (step S106). Thereby, the display 32 displays the frame recorded in the reproduction buffer.

  The CPU 30 records in the memory 34 which frame has been output to the reproduction buffer. When the CPU 30 outputs all necessary plural frames among the frames included in the GOP recorded in the stream cache recording unit to the reproduction buffer, the CPU 30 erases the contents of the stream cache recording unit.

  When there is no more content in the stream cache recording unit (step S107), the process returns to step S102 to record the next necessary plurality of frames in the stream cache recording unit.

  Two or more stream cache recording units may be provided. When all the necessary plural frames of one stream cache recording unit are output to the reproduction buffer, the contents are erased and processing is performed so as to accumulate the next necessary plural frames. During this time, the target frame is decoded according to the recording contents of the other stream cache recording unit. In this case, the processes in steps S102 and S103 and the processes in steps S104 to S106 in the flowchart of FIG. 16 are executed in parallel.

  The case where the moving image data file has a structure as shown in FIG. 3 (ES) has been described above. Next, a case where the moving image data file has a structure as shown in FIG. 7 (PES, PS) will be described. The basic processing flow is the same as in FIG.

  However, the prediction process, the stream cache process, and the frame restoration process are slightly different as shown in FIGS. 26, 27, and 28, respectively. Parts that are the same as those in FIGS. 17, 18, and 19 are denoted by the same reference numerals.

  In the prediction process of FIG. 26, in step S2041, the access position of the PES packet including the necessary plural frames is acquired from the index file (see FIG. 10).

  In the stream cache process of FIG. 27, a PES packet including a necessary plurality of GOPs is accessed in step S3011, and the PES packet is acquired and the GOP is restored in step S3012. In this way, at the stage of the stream cache process, the GOP is collectively restored and cached to speed up the process.

  In the frame restoration process of FIG. 28, in step S4061, the index file 4 (FIG. 10) is referred to, and the access position of the PES packet for obtaining the target frame is acquired. In step S4062, a PES packet is acquired, and a GOP is restored based on the PES format packet.

  The case where the moving image data file has the structure as shown in FIG. 7 (PES, PS) has been described above. Next, a case where the moving image data file has a structure as shown in FIG. 11 (TS) will be described. The basic processing flow is the same as in FIG.

  However, the prediction process, the stream cache process, and the frame restoration process are slightly different as shown in FIGS. 29, 30, and 31, respectively. Parts that are the same as those in FIGS. 17, 18, and 19 are denoted by the same reference numerals.

  In the prediction process of FIG. 29, in step S2045, the access position of the TS packet including the necessary plural frames is acquired from the index file (see FIG. 14).

  In the stream cache process of FIG. 30, a TS packet including a plurality of necessary GOPs is accessed in step S3015. In step S3016, the TS packet is acquired to restore the PES packet, and further the GOP is restored. In this way, at the stage of the stream cache process, the GOP is collectively restored and cached to speed up the process.

  In the frame restoration process of FIG. 31, in step S4065, the index file 4 (FIG. 14) is referred to, and the access position of the TS packet for obtaining the target frame is acquired. In step S4066, a TS packet is acquired, a PES packet is restored based on the TS packet, and a GOP is restored.

  In the above embodiment, the case where there is one clip to be reproduced at the same time has been described. However, as shown in FIG. 32, the present invention can also be applied to a transition in which two frames are decoded and one frame is generated based on these frames. In FIG. 32, portions T1 and T2 are transitions.

  In this case, independent frame cache recording units are provided for the two clips to be transitioned. For example, in the transition T1, two frames of the clip Clip001 of the moving image data file X and the frame of the clip Clip002 of the moving image data file Y must be synthesized. When both frames are read from the hard disk 2, the moving image data file X and the moving image data file Y must be accessed alternately for each frame. Therefore, the processing speed is extremely slow.

  In this embodiment, since the frame cache recording unit is provided in the memory 34, the processing speed can be improved.

Since such a transition is described in the transition information 215 of the editing configuration data 12 as shown in FIG. 33, the CPU 30 can make a prediction.

5). Other Embodiments In the above-described embodiment, the playback device has been described as an example. However, the present invention can be similarly applied to a device having a playback function, such as a video editing device or a video transmission device. In the moving image editing apparatus, editing is performed by rewriting the description of the editing configuration data file 6 based on the operation input of the operator.

  In the above embodiment, access is speeded up using an index file. However, the present invention can be implemented without using an index file. In this case, since the predicting unit 8 and the stream cache processing unit 10 cannot refer to the index file 4, the moving image data file 6 is directly accessed to search for desired data. In this case, the stream cache processing unit 10 may generate a cache index file when the relevant part of the moving image data file 6 is read and recorded in the non-movable recording unit 14.

  In the above embodiment, as shown in FIGS. 6, 10, and 14, in the index file, for each GOP, the head frame and the last frame are set as the representative frames, and the head frame number and the last frame number are recorded. Yes. However, the head frame number may be recorded using the head frame as a representative frame. In this case, the final frame number is not recorded, but if the CPU 130 acquires the first frame number of the next GOP arranged in time series in steps S402 and S503, the CPU 130 recognizes the range of frame numbers included in the GOP. Can do.

  In the above-described embodiment, data in the MPEG format has been described. However, the present invention can be similarly applied to moving image data in other compression formats that use bidirectional predictive coding.

  In the above embodiment, the head position is used as a representative point of the group data or packet. However, any location may be used as a representative point as long as it can be easily accessed. For example, when accessing from behind the moving image data file, the end position may be used as the representative point.

  In the above-described embodiment, each function is achieved by using the CPU, but all or a part thereof may be configured by hardware.

FIG. 1 is a functional block diagram of a moving image playback apparatus according to an embodiment of the present invention. It is a figure which shows the hardware constitutions of a moving image reproducing device. It is a figure which shows the structure of the main (proxy) moving image data file (ES). It is a flowchart of an index file creation process. It is a flowchart of an index file creation process. It is a figure which shows the structure of an index file. It is a figure which shows the structure of a moving image data file (PES). It is a flowchart of an index file creation process. It is a flowchart of an index file creation process. It is a figure which shows the structure of an index file. It is a figure which shows the structure of a moving image data file (TS). It is a flowchart of an index file creation process. It is a flowchart of an index file creation process. It is a figure which shows the structure of a moving image index file. It is a flowchart of an index file creation process. It is a flowchart of a reproduction | regeneration processing program. It is a flowchart of a prediction process. It is a flowchart of a stream cache process. It is a flowchart of a frame restoration process. It is a figure which shows the structure of edit structure data. It is a figure which shows the relationship between a moving image data file and edit structure data. It is a figure which shows a part of edit structure data. It is a figure which shows a list. It is an example of data of an index file. It is a figure which shows the example of data of a cache index file. It is a flowchart of a prediction process. It is a flowchart of a stream cache process. It is a flowchart of a restoration process of a frame. It is a flowchart of a prediction process. It is a flowchart of a stream cache process. It is a flowchart of a frame restoration process. It is a figure for demonstrating a transition. It is a figure which shows the detail of the transition information of edit structure data.

Explanation of symbols

DESCRIPTION OF SYMBOLS 2 ... Movable recording part 4 ... Index file 6 ... Movie data file 8 ... Prediction means 10 ... Stream cache processing means
12 ... Editing composition data file
14 ... Non-movable recording unit 16 ... Group data 18 ... Individual frame order information 20 ... Individual access position 22 ... Specific means 24 ... Frame restoration means 26 ... Display unit

Claims (8)

Movie data file in which frames constituting a movie are encoded and grouped, movie index file describing group information of the movie data file, start frame and end frame of each movie material which is a part of the movie data file Recording means for recording and accessing the edit configuration data file specifying the order information in the movable recording unit, and
Predicting means for predicting a plurality of frames required based on the editing configuration data file;
A part of the moving image data file corresponding to the predicted plurality of frames by the predicting unit is read from the movable recording unit and stored in a non-movable stream cache recording unit for recording a frame group of the moving image data file. Stream cache processing means;
Specifying means for specifying the next required target frame based on the editing configuration data file;
A device having a video playback function, comprising: a frame restoration unit that decodes a target frame specified by the specifying unit based on a video data file stored in the stream cache unit;
The stream cache processing means, when recording the group data in the stream cache unit, calculates the individual frame order information of at least the main picture data constituting the group data and records it in the stream cache unit,
The frame restoration means specifies picture data necessary for restoration of the target frame based on the individual frame order information,
When recording the group data in the stream cache unit, the stream cache processing unit records the recording position in the stream cache recording unit of at least main picture data constituting the group data as an individual access position in the stream cache recording unit. ,
The frame restoration means is an apparatus having a moving picture reproduction function for specifying an access position of picture data necessary for restoration of a target frame based on the individual access position . Computer
Movie data file in which frames constituting a movie are encoded and grouped, movie index file describing group information of the movie data file, start frame and end frame of each movie material which is a part of the movie data file Recording means for recording and accessing an edit configuration data file specifying the order information in the movable recording unit,
Prediction means for predicting a plurality of necessary frames based on the editing configuration data file,
A part of the moving image data file corresponding to the predicted plurality of frames by the predicting unit is read from the movable recording unit and stored in a non-movable stream cache recording unit for recording a frame group of the moving image data file. Stream cache processing means,
A specifying means for specifying a next required target frame based on the editing configuration data file;
Frame restoring means for decoding the target frame specified by the specifying means based on the moving image data file stored in the stream cache means;
A video playback program to function as
The stream cache processing means, when recording the group data in the stream cache unit, calculates the individual frame order information of at least the main picture data constituting the group data and records it in the stream cache unit,
The frame restoration means specifies picture data necessary for restoration of the target frame based on the individual frame order information,
When recording the group data in the stream cache unit, the stream cache processing unit records the recording position in the stream cache recording unit of at least main picture data constituting the group data as an individual access position in the stream cache recording unit. ,
The frame restoration means is a moving picture reproduction program for specifying an access position of picture data necessary for restoration of a target frame based on the individual access position . In the program of Claim 2,
The moving image data file recorded in the movable recording unit is a group data obtained by grouping encoded moving images of a plurality of frames, and the group data is arranged in time series,
The moving image index file describes the recording position of the representative point of each group data in the moving image data file as an access position, and represents the order of the representative frames included in each group data in the entire moving image data file. Is a description of information,
The stream cache processing means identifies group data including the predicted multiple frames based on the frame order information at both ends of the predicted multiple frames and the representative frame order information of the moving image index file, and based on the access position of the group data A moving image reproduction program for acquiring group data from the movable recording unit. In the program of Claim 2,
The moving image data file recorded in the movable recording unit is a group data obtained by grouping encoded frames of moving images, and packet data obtained by packetizing the group data is arranged in time series,
The moving image index file describes the recording position of the representative point of the packet including each group data in the moving image data file as the access position, and indicates the order of the representative frame included in each group data in the frame of the entire moving image data file. It describes the representative frame order information,
The stream cache processing means a) specifies group data including a plurality of predicted frames based on frame order information at both ends of the predicted plurality of frames and representative frame order information of the moving image index file, and b) includes the group data. A moving image reproducing program that acquires a packet including group data from the movable recording unit based on the access position of the packet, and c) restores the group data from the packet and records it in the stream cache recording unit. In the program in any one of Claims 2-4,
The frame restoration means records one or more frames restored to obtain the target frame in the decoded cache recording unit, and the newly required target frame is recorded in the decoded cache recording unit. For example, a moving image reproduction program that is acquired from this and is acquired from the stream cache recording unit and restored if it is not recorded in the decoded cache recording unit. In the program of Claim 5,
The frame restoration means is a moving image reproduction program for obtaining and restoring from the movable recording unit when the target frame is not recorded in either the decoded cache recording unit or the stream cache recording unit. In the program in any one of Claims 2-6,
The stream cache processing unit or the prediction unit determines whether or not the predicted plurality of frames are in descending order in the expected reproduction order based on the frame order information, and corrects in ascending order if descending order. Record a video data file that encodes the frames that make up a video, and specify a movable recording unit with a movable unit for access, and start frame order information and end frame order information for each video material that is part of the video A method of reproducing a moving image by a computer accessible to an editing configuration data file recording unit that records the editing configuration data file and a stream cache recording unit for recording a frame group of the moving image data file,
Predicting means to predict the required multiple frames based on the editing configuration data file,
The stream cache processing unit reads out a part of the moving image data file corresponding to the predicted plurality of frames by the prediction unit from the movable recording unit, and stores it in the stream cache recording unit.
By the specifying means, the next necessary target frame is specified based on the editing configuration data file,
A video reproduction method characterized by decoding the target frame based on a video data file stored in a stream cache means by a stream restoration means,
When recording the group data in the stream cache unit by the stream cache processing unit, the individual frame order information of at least main picture data constituting the group data is calculated and recorded in the stream cache unit,
Based on the individual frame order information, the frame restoration means identifies picture data necessary for restoration of the target frame,
When recording the group data in the stream cache unit by the stream cache processing means, the recording position in the stream cache recording unit of at least main picture data constituting the group data is recorded in the stream cache recording unit as an individual access position. ,
A moving image reproduction method for specifying an access position of picture data necessary for restoring a target frame based on the individual access position by the frame restoration means.

Images