JP2009005382A - Moving image playback apparatus and recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve the problem wherein random access in units of a GOP can not be performed, and after random access, audio/video synchronism is not kept. <P>SOLUTION: A plurality of GOPs are packetized and packet information is recorded in a management file for each packet. The packet information is constituted of: a starting position of the packet; an address of next packet information; a starting position of a GOP for each GOP; VPTS indicating a playback time of video; APTS indicating a playback time of audio; an ending position of coded data of an intra-frame coded frame (I ending position); and an ending position of coded data of a forward inter-frame prediction coded frame (P ending position). <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an apparatus and a recording medium for recording and reproducing compressed video and audio data on a disk medium or a semiconductor memory.

  In recent years, MPEG1 (ISO / IEC 11172) and MPEG2 (ISO / IEC 13818) systems are used in the fields of storage, communication, and broadcasting as moving picture encoding systems. In the MPEG system, intra-frame coding, forward inter-frame predictive coding, and bi-directional inter-frame predictive coding are switched for each frame.

  Hereinafter, a frame to be intra-coded is referred to as an I frame, a frame to be subjected to inter-frame prediction encoding in the forward direction is referred to as a P frame, and a frame to be subjected to inter-frame prediction encoding in both directions is referred to as a B frame.

  FIG. 19 shows an example of the arrangement of I frames, P frames, and B frames. In the MPEG system, as shown in FIG. 19, a plurality of frame groups including one or more I frames are referred to as “Group Of Pictures” (hereinafter abbreviated as GOP).

  When the encoded data compressed by the MPEG method is reproduced from the middle of the video sequence, it is necessary to decode from the I picture. This is because the P picture and B picture are inter-frame prediction encoded, so it is necessary to first decode the picture used for prediction. If this dependency is followed, all P pictures and B pictures are I pictures. Because it depends on.

  By the way, the MPEG system is variable length coding, and time and code amount are not proportional. That is, in the case of reproducing from the middle of encoded data by random access or the like, even if the reproduction start time is known, it is not known where the data at that time is recorded. For this reason, it is necessary to manage in advance which time data is recorded in which position on the medium.

  Japanese Patent Application Laid-Open No. 2004-228561 describes an apparatus that manages the position of a GOP that is an access unit in the MPEG system and efficiently performs random access. In Patent Document 1, n (n is an integer of 1 or more) GOPs are collectively recorded as a data unit (DUT) on an optical disk, and the recording position of the DUT is also recorded separately.

  FIG. 20 shows the configuration of the DUT used in Patent Document 1.

  The DUT 201 includes a DUT header 202, sub video data 203, audio data 204, and main video data 205. The sub video data 203 includes audio data 204 such as a sub-picture (caption data) corresponding to the time of the main video data 205. Is audio data corresponding to the time of the main video data 205, and video data compressed by the MPEG method is recorded in the main video data 205.

  Here, data other than the DUT header 202 has a different data length for each DUT. The length of each data constituting the DUT is determined so as to be an integral multiple of the physical sector length of the optical disc.

The sector number at which the head of the DUT is recorded is recorded separately in the management file of the disk. During random access or high-speed playback, the start position of the target DUT is searched from the management file information, jumped, and playback is started from the head of the GOP. To do.
Japanese Patent Laid-Open No. 7-284060

  However, the method described in Patent Document 1 (hereinafter referred to as a conventional example) has the following five problems.

  First, since only the start position of the DUT is managed in the conventional example, there is a problem that smooth display cannot be performed during high-speed playback at 2 to 3 times speed.

  During high-speed playback, reading of the I picture code from the compressed data is repeated.

  This situation will be described with reference to FIG.

  In FIG. 21, first, jump to the head of the DUT 211 and read the I picture 213 in the DUT 211. When the end of the I picture is detected, the process jumps to the top of the DUT 212, reads the I picture 214 in the DUT 212, and thereafter repeats the same operation.

  As an example, if the number of frames constituting a GOP is 15 and the DUT is composed of 4 GOPs, 1 DUT is equivalent to about 2 seconds of data, and one I picture of about 2 seconds of data is played back. repeat. In this case, when one I picture per second is reproduced, the speed is doubled, and when two I pictures are reproduced per second, the speed is quadrupled. In other words, only one image per second can be displayed at 2 × speed, and only two images per second can be displayed at 4 × speed, and smooth high-speed playback cannot be achieved.

  In the conventional example, if the number of GOPs constituting the DUT is reduced, more I pictures can be reproduced. For example, if the DUT is composed of 1 GOP, the playback speed is doubled when 4 I pictures are played per second. However, when the number of I pictures that can be read per unit time is larger, only a smaller number than the displayable number can be displayed.

  Secondly, the conventional example has a problem in that useless reading occurs during high-speed reproduction.

  As shown in FIG. 21, at the time of high-speed playback, first jump to the head of the DUT. The I picture data to be read is recorded in the main video data 205 of FIG. 20, but it is necessary to read the DUT header 202, the sub video data 203, and the audio data 204 so far.

  In high-speed playback, jumping to the next DUT is completed after reading of the I picture, but the end position of the I picture needs to be detected separately.

  FIG. 22 shows a block diagram of an apparatus for recording and reproducing video on a disc.

  In FIG. 22, the controller 229 controls the recording medium 231, and data from the head of the DIU is output from the recording medium 231. The output data is ECC (error correction) encoded and is decoded by the ECC unit 230.

  The decoded data is separated into a DUT header, sub-video data, audio data, and main video data by a multiplexing / separating circuit 227 and held in the memory 228. The DUT header is input to a system controller (not shown), and the sub video data is input to a sub video decoding circuit (not shown). The audio data is decoded by the audio codec 225, and the main video data is decoded by the video codec 226. Normally, video data is decoded by the video codec 226, so that the end of the I picture is also detected by the video codec 226.

  In FIG. 22, there is a processing time delay in the ECC unit 230 and the multiplexing / demultiplexing circuit 227 from when the data is read from the recording medium 231 to when the main video data is input to the video codec 226.

  Since the ECC unit 230 performs processing in a large unit such as 32 KB, a delay in the time for accumulating data for the unit occurs at least. Accordingly, when the video codec 226 detects the end of the I picture, the next unnecessary data has already been read.

  Third, in the conventional example, when recording is performed in real time on a recording medium, there is a problem that a large-capacity memory is required to configure the DUT.

  As shown in FIG. 20, in the conventional example, audio data is first recorded, and then main video data is recorded. That is, the main video data must be held in the memory until the DUT audio data is obtained.

For example, if the encoded data rate of the video is composed of 4 Mbps on average and the DUT is about 2 seconds,
4Mbit x 2 = 8Mbit
Of memory is required.

Furthermore, since the encoded data of the video has a variable length, the amount of code generated temporarily increases, so that it is necessary to have a memory capacity that is greater than the average code amount.
Fourthly, the conventional example has a problem that the encoded data cannot be processed in a unit smaller than that of the DUT.

  In the case of a disk medium, it is possible to perform program reproduction and editing in which the reproduction order is indicated by a pointer without changing the recorded data. In the conventional example, when the DUT is composed of a plurality of GOPs, since the GOP boundary is not known, the minimum editing unit becomes the DUT, and there is a problem that editing cannot be performed for each GOP which is an MPEG access unit.

  Fifth, since there is no information for synchronizing audio data and video data in the conventional example, there is a problem that synchronization between audio and video cannot be ensured.

  In the conventional example, the DUT is composed of encoded data corresponding to the time constituting the GOP. Here, the period of the video frame is 30000/1001 Hz, and encoding is also performed for each period. On the other hand, the sampling frequency of audio data is 44.1 KHz. In the case of the MPEG1 system, 1152 data is often used as an audio frame, and in the ATRAC system and ATRAC2 system, 1024 data is often encoded as an audio frame.

  The video frame is about 33 ms, and the audio frame is about 26 ms or about 23 ms, so that the time of the main video data and the audio data corresponding to the GOP cannot be completely matched. That is, the playback start time of the main video data and the audio data is completely the same for each DUT only in the first DUT, and the playback time of the main video and the playback time of the audio data are the same at the start point of the subsequent DUT. do not do.

  Since the time of the main video data and the audio data constituting the DUT is different, when the program reproduction in which the reproduction start point and the end point are designated by the pointer is continuously performed, the time error between the main video data and the audio data is increased. Accumulation causes problems that cannot be synchronized.

  The present invention has been made in view of such a problem, and is a moving image recording that realizes smooth high-speed reproduction by increasing the number of displayed images per unit time without reading useless data during high-speed reproduction. Reproduction method and apparatus, moving image recording / reproduction method and apparatus that do not require a large-capacity memory even in the case of real-time recording, and random access and editing in GOP units are possible. It is an object of the present invention to provide a moving image recording / reproducing method and apparatus capable of ensuring the synchronization of images.

  The present invention has taken the following measures in order to solve the problems.

  That is, the moving picture recording / reproducing apparatus according to the present invention reproduces multiplexed data from a recording medium in which multiplexed data obtained by multiplexing at least audio encoded data and video encoded data and management data are recorded as separate files. In the moving image reproducing apparatus, the management data includes at least position information on a recording medium of the multiplexed data and a reproduction time, and the position information is about the multiplexed data of a frame group composed of a plurality of frames. Including a start position of the frame group and an end position of a frame to be intra-coded, and the playback time includes a playback time of a first video frame of the frame group, and is randomly generated using the management data It is characterized by performing access.

  In addition, the multiplexed data is recorded on the recording medium while being divided into a predetermined length, and the multiplexed data of the frame group is recorded in the unit of the fixed length without excess or deficiency. Device.

  The recording medium according to the present invention is a recording medium on which multiplexed data obtained by multiplexing at least audio encoded data and video encoded data and management data are recorded as separate files, wherein the management data is at least multiplexed. The position information on the recording medium of the data and the reproduction time are included, and the position information includes the start position of the frame group and the frame encoded in the frame with respect to the multiplexed data of the frame group including a plurality of frames. Including the end position, the playback time includes the playback time of the first video frame of the frame group.

  According to the present invention, the position information for each GOP and the end position of the I picture or P picture are recorded in a management file recorded separately from the multiplexed data, and only necessary data is read from the recording medium during high-speed playback. This increases the number of images displayed during high-speed playback.

  In addition, since the end positions of both the I picture and the P picture are managed, many images can be displayed even at a high speed reproduction of 2 to 3 times, and the screen can be changed smoothly.

  Further, since the recording position of multiplexed data to be read for high-speed reproduction can be known only by the management file, there is no need for a control signal from the video codec or the multiplexing / demultiplexing circuit, and there is an effect that the high-speed reproduction read control is simplified.

  Furthermore, since the position information is recorded in a management file separate from the multiplexed data and no additional information such as a header is added to the multiplexed data, all recording media recording speeds can be assigned to the multiplexed data. The image quality is improved.

  Furthermore, in the present invention, since the playback time of audio and video is recorded in the management file together with the position information for each GOP, random access is possible in GOP units, and there is a difference in playback time between audio and video. Can also be corrected, and synchronization can always be ensured with audio and video.

  Also, according to the present invention, a packet is composed of a plurality of GOPs, and both multiplexed data and management data are recorded for each packet, so that data can be easily deleted or edited in units of packets.

  Also, according to the present invention, since the multiplexed data of GOP is configured in the order of video encoded data and audio encoded data, the video encoded data output from the video codec can be recorded sequentially, and the encoded data There is an effect that it is not necessary to provide a large-capacity memory for buffering.

  In addition, since the deviation between the video frame and the audio frame constituting the multiplexed data of the GOP is below a certain threshold value, even if the multiplexed data is exchanged in GOP units or packet units in editing or the like, The deviation can be made below a certain threshold.

  Further, according to the present invention, since the correction is performed when the deviation amount between the reproduction time of the video frame and the reproduction time of the audio frame is large for each GOP boundary, the deviation amount does not accumulate.

  Further, the correction of the shift amount can be easily realized by freezing or skipping the video frame.

  Further, the correction of the shift amount is performed on the video frame, and the audio frame is reproduced continuously, so that there is an effect that an unpleasant sound is not generated.

  Further, according to the present invention, program reproduction can be expressed only by a pointer indicating the position of the management file. The management file is held in the controller, and there is an effect that program reproduction can be easily specified without reading multiplexed data.

  In addition, according to the present invention, since multiplexed data is configured for each video frame, not only video encoded data but also audio encoded data can be recorded sequentially, so that a buffering memory has a smaller capacity. It can be.

  Further, according to the present invention, since the boundary between the video encoded data and the audio encoded data is recorded in the management data, the circuit for detecting the boundary can be reduced.

  In addition, video encoded data and audio encoded data can be read in advance during high-speed reproduction, and high-speed reproduction while reproducing sound can be realized without reading unnecessary data.

  Further, according to the present invention, since the position information and time information for each GOP are multiplexed and recorded in the multiplexed data, the size of the management file can be greatly reduced.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
[First Embodiment]
FIG. 1 is a block diagram showing the configuration of the first embodiment of the moving image recording / playback apparatus of the present invention.
1 includes an audio codec 15, a video codec 16, a multiplexing / demultiplexing circuit 17, a memory 18, a controller 19, an ECC unit 20, and a disk 21. In the present invention, the multiplexing / demultiplexing circuit 17 and the controller are included. There are 19 operations.

The disk 21 refers to all rewritable recording media such as a magnetic disk, a magneto-optical disk, and a phase change disk, but is not limited to a disk and may be replaced with a semiconductor memory.
The operation during recording in the embodiment of FIG. 1 will be described.

  At the time of recording, audio data is input from the audio input terminal 11 and encoded by the audio codec 15. Video data is input from the video input terminal 13 and encoded by the video codec 16. Audio encoded data is output from the audio codec 15, and video encoded data is output from the video codec 16 and input to the multiplexing / demultiplexing circuit 17.

  The multiplexing / demultiplexing circuit 17 once outputs the encoded data to the memory 18, reads the encoded data from the memory 18 in the order described later, and outputs the encoded data to the ECC unit 20 as multiplexed data. The multiplexed data is error correction encoded by the ECC unit 20 and recorded on the disk 21. The controller 19 controls the recording medium.

The multiplexed data output from the multiplexing / separation circuit 17 will be described with reference to FIG.
FIG. 2A is a diagram illustrating a temporal relationship between a video frame constituting a GOP and an audio frame corresponding to the GOP.

  FIG. 2A shows an example in which 1 GOP is composed of 15 video frames, and the audio frame period is shorter than the video frame period. The 15 video frame period is the GOP period 22. When the video frame and audio frame periods are different, the GOP period 22 does not match even if a plurality of audio frames are collected. Therefore, a period in which the deviation from the beginning and the rear end of the GOP period 22 is less than one audio frame period is an audio period 23, and the encoded data of the video frame in the GOP period 22 and the audio frame in the audio period 23 are The encoded data is multiplexed and multiplexed.

  FIG. 2B shows an arrangement of video encoded data and audio encoded data in the multiplexed data.

  The multiplexed data is output for each GOP in the order of the video encoded data in the GOP period 22 and the audio encoded data in the audio period 23 shown in FIG. A plurality of GOPs are recorded on the disk in units of packets. FIG. 2B shows an example in which four GOPs are used as packets.

  FIG. 3 is a flowchart showing an operation of outputting the multiplexed data shown in FIG. 2B in the multiplexing / separating unit circuit 17.

  The multiplexing / demultiplexing circuit 17 once records the audio encoded data input from the audio codec 15 and the video encoded data input from the video codec 16 in the memory 18.

  The flowchart in FIG. 3 shows an operation of reading the encoded data recorded in the memory 18.

  In FIG. 3, the encoded video data of the GOP period 22 shown in FIG. 2A is output in steps S32 to S35, and the encoded audio data of the audio period 23 is read in steps S36 to S39.

  First, in step S31, VPTS indicating the time of the video frame and APTS indicating the time of the audio frame are set to zero. In step S 32, the encoded video data is read from the memory 18 and output to the ECC unit 20.

  Next, in step S33, it is determined whether or not the video frame ends. Whether the video frame ends can be determined if the code amount of the video frame is known. For this purpose, a circuit for detecting the header of the video frame is added to the multiplexing / demultiplexing circuit 17 to detect a break of the video frame when writing the video encoded data to the memory 18, and the code amount for each video frame is calculated in advance. Alternatively, a circuit for calculating the code amount of the video frame is added to the video codec 16 so that the video codec is notified of the code amount for each video frame.

  If it is determined in step S33 that the video frame has ended, VPTS is increased by the video frame period in step S34. In step S35, it is determined whether or not the GOP is finished. If the GOP is not finished, steps S32 to S34 are repeated. Whether or not the GOP is ended may be determined by counting the number of times the video frame is ended and comparing whether or not it matches the number of video frames constituting the GOP. In the process up to step S35, all of the video encoded data in the GOP period is output. Since VPTS is set to 0 in step S31 and VPTS is incremented in step S34 every time video encoded data of one video frame is output, VPTS always indicates the time of video encoded data read from the memory 18. Become.

  If it is determined in step S35 that the GOP has ended, the encoded audio data is read from the memory 18 and output to the ECC unit 20 in step S36. Next, in step S37, it is determined whether or not the audio frame ends.

  In main audio encoding systems such as the MPEG1 system and the ATRAC system, the code amount of the audio frame is constant. Therefore, it is possible to determine whether or not the audio frame is ended by using the read data amount in step S36. If it is determined in step S37 that the audio frame has ended, APTS is increased by the audio frame period in step S38. In step S39, APTS and VPTS are compared. If APTS is small, steps S36 to S38 are repeated to read the next audio encoded data.

  At this time, all of the video encoded data in the GOP period is output, and VPTS indicates the beginning of the next GOP period. Therefore, the determination in step S39 is No. All the audio encoded data in the audio period is read. When APTS points outside the GOP period.

  FIG. 3 shows an example in which VPTS and APTS are calculated and obtained in the multiplexing / demultiplexing circuit 17, but VPTS may be input from the video codec 16 and APTS may be input from the audio codec 15.

  In steps S32 to S39, video encoded data corresponding to one GOP period and audio encoded data are read from the memory 18 in this order, and the multiplexed data shown in FIG. 2B is output. In step S40, it is determined whether or not the encoded data is finished, and steps S32 to S39 are repeated until the coded data is finished.

  In the conventional example, it is necessary to temporarily store all the video encoded data of the DUT in the memory, and a large-capacity memory is necessary. However, in this embodiment, since the video encoded data is sequentially read from the memory, Large memory is not required.

  The audio encoded data corresponding to the GOP needs to be stored in the memory. However, since the audio encoded data has a very small data amount compared to the video encoded data, the capacity is much smaller than that of the conventional example. It is only necessary to provide the memory.

  Note that the recording of the multiplexed data on the recording medium is performed for each fixed recording unit such as 2 Kbytes or 32 Kbytes. Therefore, data called stuffing may be added so as to be a recording unit for each video encoded data in the GOP period, for each multiplexed data in the GOP period, or for each packet. Stuffing is data that is all 0 or 1, and is ignored in the decoding process.

  In order to set the recording unit for each video encoded data in the GOP period, 0 or 1 is added to the recording unit after it is determined in step S35 that the GOP has ended, and the recording unit is set for each multiplexed data in the GOP period. In order to be set to No, in step S39, 0 or 1 is added to the recording unit after it is determined to be No, and in order to become a recording unit for each packet, the packet is determined after No is determined in step S39. It is sufficient to add 0 or 1 to the recording unit in the case of the end of the packet.

  In the present embodiment, since the packet is composed of four GOPs, the number of GOPs is integrated, and when the number of GOPs is a multiple of 4, it can be determined that the packet has ended.

  Next, the operation of the controller 19 will be described. The controller 19 manages the multiplexed data recorded on the disc 21 as a file, and performs disc recording control and playback control.

  On the disc 21, a management file is recorded as information on the entire disc in an area different from the area for recording the multiplexed file. The management file is read from the disk 21 when the moving image recording / reproducing apparatus is activated or when a disk medium is inserted into the apparatus, and is stored in the controller 19. The controller 19 controls the disk so that multiplexed data is recorded in an empty area of the disk during recording, and updates the contents of the management file as needed.

  FIG. 4 shows an example of the management file. The management file includes “disc information 41” such as a disc title, “file information” such as recording position information of the file (that is, multiplexed data) and information indicating a connection when the file is divided and recorded in a discontinuous area. 42 "," free area information 43 "which is disk free information, and" packet information 44 "which records information for each packet.

  As shown in FIG. 2B, in this embodiment, a group of four GOPs is called a packet, and management is performed in units of packets. There is one “packet information” per packet, and at the time of recording, the “packet information” of the recorded packet is added each time a packet is recorded on the disk 21.

  The “packet information” includes “packet start position”, “next address”, and “GOP start position”, “VPTS”, “APTS”, “I end position”, and “P end position” for each GOP.

  FIG. 5 is a flowchart showing an operation for updating the “packet information” of the management file shown in FIG. 4 at the time of recording.

  This operation is the same as the operation of constructing a packet by the multiplexing / demultiplexing circuit 17 shown in FIG. 3 except that “start position of packet”, “start position of GOP”, “VPTS”, “APTS”, “I end position”. The operation of outputting “P end position” is added, and the same operation parts as in FIG.

  The controller 19 takes in the position information input from the multiplexing / demultiplexing circuit 17 and adds “packet information” for each packet. For this reason, the controller 19 also manages an empty area for writing “packet information” on the management file. “Next address” of “packet information” indicates the address of “packet information” of the packet following the current packet, and is given by the controller 19.

  In this embodiment, “packet start position”, “I end position”, and “P end position” (hereinafter collectively referred to as position information) are relative positions from the beginning of the multiplexed data. Is 32K bytes, the first 32K bytes are 0, the next 32K bytes are 1, and the value is incremented every 32K bytes thereafter. The file system (not shown) performs conversion from the relative position to the physical address of the disk. The multiplexing / demultiplexing circuit 17 counts the amount of data to be recorded in the memory 18 and the amount of data to be output to the ECC unit 20 and uses it for calculation of position information.

  In FIG. 5, first, “packet start position” is output to the controller 19 in step S51. In step S 52, “GOP start position” and “VPTS” are output to the controller 19. Next, after reading of encoded data of one video frame is completed in S32 to S33, it is determined whether or not the I picture ends (step S53).

  If the I picture ends, the “I end position” is output to the controller 19 (step S54). If the I picture does not end, it is determined whether the first P picture ends in the GOP (step S55). . If the end of the first P picture in the GOP, the “P end position” is output to the controller 19 (step S56). When the output of the GOP video encoded data is completed in the loop of steps S32 to S35, APTS is output in step S57, and after the GOP audio encoded data is read in S36 to S39, it is determined whether or not the packet is ended. (Step S58).

  If it is not the end of the packet, S52 to S39 are repeatedly executed, and processing for each GOP is performed until the packet ends. In the case of the end of the packet, it is determined in step S40 whether or not the end of the encoded data, and if it is not the end of the encoded data, steps S51 to S58 are repeatedly executed to repeat the process for each packet.

  Step S51 is executed once for each packet, and S52, S54, S56, and S57 are executed once for each GOP, and information necessary to configure packet information is output from the multiplexing / demultiplexing circuit 17 to the controller 19. The

  This embodiment is characterized in that location information for each packet and location information for each GOP are recorded in the management file, and the contents of the management file are not limited to the example shown in FIG. For example, management information for each multiplexed data may be added to the management file, or a packet code amount or a code amount for each GOP may be added to the packet information. Further, in the example of FIG. 4, packet information for multiplexed data of the entire disk is managed collectively, but may be managed separately for each multiplexed data.

  Furthermore, the management file shown in FIG. 4 is an example. For example, the “P end position” may be two or more.

Next, the normal playback operation in the present embodiment will be described.
During normal reproduction, the controller 19 controls to read a file (that is, multiplexed data) to be reproduced on the disk 21. The position of the file to be reproduced on the disc is recorded in “file information 42” in FIG.

  The data read from the disk is subjected to error correction decoding by the ECC unit 20 and becomes multiplexed data. The multiplexed data is separated into audio encoded data and video encoded data by a multiplexing / separating circuit 17 and once recorded in the memory 18. Audio encoded data is output from the memory 18 to the audio codec 15 and decoded. The encoded video data is output to the video codec 16 and decoded.

  In the disk medium, the recorded multiplexing is not changed, and it is possible to reproduce data that is not continuous by random access as if it were continuous data. In this case, a process for ensuring synchronization between audio and video is necessary.

FIG. 6 shows an example in which two GOPs are reproduced as if they were continuous data.
As shown in FIG. 2A, generally, a GOP period that is a period of a video frame constituting a GOP and an audio period that is an audio frame period corresponding to the GOP do not match. In FIG. 6, the multiplexed data composed of the GOP period 61 and the audio period 62 shown in (a) and the multiplexed data composed of the GOP period 63 and the audio period 64 shown in (b) are continuously provided. The case of reproduction is shown in (c). When the audio period is longer than the GOP period as shown in FIGS. 6A and 6B, when the GOP is continuously reproduced, the error in the reproduction time between the audio frame and the video frame is accumulated as shown in FIG. 6C. There is a problem to do.

  Therefore, in this embodiment, the reproduction time is corrected using “VPTS” and “APTS” in the packet information of the management file.

  For example, when a video frame precedes an audio frame as shown in FIG. 6C, a freeze period 65 for interrupting video decoding by one frame is set to ensure synchronization. This state is shown in FIG. Conversely, if the audio frame precedes the video frame, a skip period for skipping one frame of the video is set.

  A method for setting a video freeze period and a skip period will be described with reference to FIG. FIG. 7 shows the reproduction time at the GOP boundary between video encoded data and audio encoded data that are continuously reproduced. FIG. 7A shows a video frame to be decoded, and FIGS. 7B and 7C show an audio frame to be decoded. A video GOP period 71 starts from time 1 at the time of recording and playback, and is an audio period at the time of recording. 72 indicates that it starts from time 2 but starts from time 3 as an audio period 3 during reproduction.

  As shown in FIGS. 6B and 6C, the relationship between the video frame and the audio frame at the time of reproduction may not match the relationship between the video frame and the audio frame at the time of recording. In FIG. 7, the time 1 is recorded as “VPTS” and the time 2 is recorded as “APTS” in the management file at the time of recording. In the present embodiment, a video freeze period and a skip period are set according to the magnitude of the “deviation amount”.

  A method of obtaining the deviation amount 74 in FIG. 7 will be described with reference to FIG.

  FIG. 8 shows an example in which reproduction is started from the beginning of the multiplexed data, GOP 81 in the middle is reproduced, jumping to GOP 83 and reproduction from GOP 83 is performed.

  In FIG. 8, “VPTS” and “APTS” of GOP 82 are VPTS 82 and APTS 82, respectively, and “VPTS” and “APTS” of GOP 83 are VPTS 83 and APTS 83, respectively. Here, when (VPTS82-APTS82) and (VPTS83-APTS83) do not match, the playback time of the audio frame is shifted in the playback of GOP83. Assuming that the amount of deviation that delays reproduction shown in FIG. 7C is a positive value, the amount of deviation in the case of FIG. 8 is (APTS82−VPTS82) − (APTS83−VPTS83).

  In FIG. 9, when the positive shift amount is equal to or longer than the video frame period, the video decoding is stopped for one frame period to freeze the video output, and the negative shift amount is equal to or longer than the audio frame period. An operation for skipping video decoding by one frame is shown in a flowchart.

In FIG. 9, first, the deviation amount is set to 0 in step S91. If the deviation amount is equal to or longer than the video frame period in step S92, the audio decoding is delayed. Therefore, in step S93, the video decoding is stopped by one frame, the video output is frozen, and the video frame is calculated from the deviation amount in step S94. Subtract period. If the shift amount is less than the video frame period in step S92, it is determined in step S95 whether the shift amount is smaller than the negative audio frame period.
If Yes in step S95, the video decoding is delayed, so in step S96 the video decoding is skipped and the video is advanced by one frame. In step S96, the first I picture of the GOP is reproduced normally, and the next B picture is skipped. This is because if the I picture is skipped, the entire GOP cannot be decoded.

  In addition, the B picture has an effect that the code amount is small and skipping can be easily performed. If there is no B picture in the GOP, the last P picture of the GOP is skipped. In step S97, the video frame period is added to the shift amount. Next, audio data and video data are decoded until GOP is completed in steps S98 and S99. When the GOP decoding is completed, a deviation amount at the head of the next GOP is calculated in step S100.

  In step S100, [next recorded GOP] is a GOP that was continuous at the time of recording, and [next GOP to be reproduced] indicates a GOP to be reproduced next by random access. In the example of FIG. 8, [Next recorded GOP] is GOP 82, and [Next GOP to be reproduced] is GOP 83. [Next-recorded GOP] and [Next-playback GOP] differ only when the current GOP and the next-playback GOP are not recorded continuously, and are the same when played back in the order recorded. It indicates GOP and the amount of deviation does not change.

  Steps S92 to S100 are repeatedly executed until the encoded data is completed in step S101. When the absolute value of the deviation amount is large at the head of the GOP, video freeze or video skip is performed so as to correct the deviation amount. And video sync.

  Next, the high-speed playback operation in this embodiment will be described.

  In the high-speed playback operation, random access and I-picture reading are repeated as shown in FIG.

  In the present embodiment, “GOP start position” to “I end position” in the “packet information” of the management file is the video encoded data of the I picture. Since the “I end position” is not managed in the conventional embodiment, it is necessary to separately detect the end of the I picture, and there is a time difference between the data reading from the disk and the end of the I picture being detected. Although unnecessary data has been read, in this embodiment, since data to be read is known in advance, useless data is not read, and the number of displayed images during high-speed playback is increased.

  Whether to read out the I picture of all GOPs during thin playback or thin out may be determined based on the speed with the number of displayed images during fast playback.

  When more images can be displayed than all I pictures, a smoother high-speed reproduction can be realized by displaying P pictures in addition to I pictures. Data read when displaying an I picture and a P picture during high-speed playback will be described with reference to FIG.

FIG. 10 shows an example of encoding with the GOP configuration shown in FIG. 19, and the encoded data of the GOP is I, B, B, P,. . . Are recorded in this order. In general, the amount of codes per video frame is the largest for I pictures, the largest for P pictures, and the smallest for B pictures. In many cases, the code amount of a B picture is considerably smaller than that of an I picture or P picture.
Therefore, when reading I and P pictures in the same GOP, rather than randomly accessing each of the I and P pictures, reading from the start position of the I picture to the end position of the P picture and discarding unnecessary B pictures Is efficient.

  Therefore, in this embodiment, when reading out the I picture and the P picture, the “packet information” is read from “GOP start position” to “P end position”, and the video codec 16 reproduces the I picture 101, which is unnecessary. The sign of the B pictures 102 and 103 to be discarded is discarded, and then the P picture 104 is reproduced.

  At the time of high-speed reproduction, the controller 19 performs control so as to reproduce the encoded data from the I picture or the I picture to the first P picture, and otherwise performs the same operation as the normal operation.

  Next, a description will be given of program reproduction in which reproduction start points and end points of multiplexed data recorded on the disk 21 are designated by pointers and reproduction editing is performed without changing the recorded multiplexed data.

FIG. 11 shows an example of program reproduction. FIG. 11 shows an example in which only GOPs 111 to 112 and GOPs 114 to 115 are selected and reproduced from the multiplexed data. The management file address indicating the reproduction start point and end point is recorded separately from the multiplexed data as a user program. In FIG. 11, the user program has two sets of start address and end address, start address 1 indicates the start position of GOP 111, and end address 1 indicates the end position of GOP 112. The start address 2 indicates the start position of the GOP 113, and the end address 2 indicates the end position of the GOP 114. Since the user program indicates the address of the management file, the start address 1 indicates the address where the “GOP head position” of the packet information of the packet including the GOP 111 is recorded, and the end address 1 indicates the “GOP of the packet information of the packet including the GOP 113. “Start position” indicates an address to be recorded. In the present embodiment, since the end position of the GOP is not managed, the position immediately before the “GOP start position” of the next GOP is set as the end position of the GOP.
Also during program reproduction, the controller 19 controls to read the designated multiplexed data from the disk 21, and otherwise performs the same operation as the normal operation.

  In the present embodiment, when deleting multiplexed data, the multiplexed data itself is deleted, and packet information corresponding to the multiplexed data to be deleted is also deleted. The packet information part that has been deleted to become a free area is managed using a separate table for managing the free area, or the packet information part is rewritten so as to fill the free area part every time the multiplexed data is deleted.

  In the present embodiment, there is no information indicating the boundary between audio encoded data and video video encoded data constituting multiplexed data. The boundary between the encoded audio data and the encoded video data is, for example, when video decoding is performed, and when the GOP decoding ends, the boundary between the encoded video data and the encoded audio data is used, or at the beginning of each encoded data. It can be recognized by adding a header indicating the type of encoded data.

  [Modification of First Embodiment] In the first embodiment, as shown in FIG. 2B, video encoded data and audio encoded data for each GOP are recorded together. As shown in FIG. 12, GOP video encoded data and audio encoded data may be divided and recorded.

  In order to configure the multiplexed data shown in FIG. 2B, it is necessary to store all audio data in the GOP period in the memory 18, but in the case of configuring the multiplexed data shown in FIG. The encoded audio data can be sequentially output, and the capacity of the memory 18 can be further reduced.

  FIG. 13 is a flowchart showing an operation of outputting the multiplexed data shown in FIG. 12 in the multiplexing / separating unit circuit 17. In the flowchart of FIG. 12 and the flowchart of FIG. 3 showing the operation of forming the multiplexed data of FIG. 2B, in FIG. 3, the video encoded data is read for each GOP in steps S32 to S35, and steps S36 to S39 are performed. In FIG. 13, video encoded data of one video frame is read in steps S32 to S34, and one video frame is supported in steps S36 to S39. The audio encoded data is read out, and steps S32 to S34 and steps S36 to S39 are repeatedly executed until the GOP is completed, but the processes executed in each step are the same.

  In this embodiment, encoded data of video frames is not input in the order of video frames to be reproduced. For example, in the case of the GOP shown in FIG. 19, the GOP starts from the B picture, but the encoded data is input from the I picture after the B picture. Therefore, an audio frame corresponding to a video frame is a correspondence in the order of encoded data.

[Second Embodiment]
In the first embodiment, the boundary between the encoded audio data and the encoded video data constituting the multiplexed data is not known unless the multiplexed data is input. In the present embodiment, information indicating the boundary is included in the management file so that the boundary between the encoded audio data and the encoded video data can be recognized in advance.

  In the present embodiment, for example, video frames and audio frames can be selected and read out even during high-speed reproduction, and high-speed reproduction while reproducing sound is possible.

  The configuration of the management file in this embodiment is shown in FIG.

  In the management file of FIG. 14 and the management file of the first embodiment shown in FIG. 4, the management file of FIG. 14 includes “video code amount” indicating the GOP video encoded data amount in the packet information. Is different.

  In this embodiment, as shown in FIG. 2B, since the GOP video encoded data is recorded from the head of the GOP, the audio encoded data is calculated from the “GOP start position” and the “GOP code amount”. You know the starting position.

  FIG. 15 is a flowchart showing an operation for updating the “packet information” of the management file shown in FIG. 14 during recording. This operation is a flowchart of the operation of updating the “packet information” of the management file shown in FIG. 5. In FIG. 5, the operation that output “APTS” in step S35 is replaced with “video code amount” and “APTS”. The operation is the same as that shown in the flowchart of FIG. 5 except for step 151. The “video code amount” is calculated by counting the code amount when reading the video encoded data in step S32.

[Third Embodiment]
In the first and second embodiments, “GOP start position”, “VPTS”, “APTS”, “I end position”, and “P end position” for each GOP are recorded in the management file of the entire disk. There is a problem that the data amount of the management file increases as the recording time increases. In the third embodiment, part of the GOP position information is recorded in the multiplexed data to reduce the data amount of the management file.

  FIG. 16 shows the structure of multiplexed data in the third embodiment of the present invention.

  FIG. 16 is obtained by adding GOP management data to the head of a GOP in addition to the multiplexed data configuration shown in FIG. In the GOP management data, “VPTS”, “APTS”, “I end position”, and “P end position” recorded in the management file in the first embodiment are recorded. The configuration of the management file in this case is shown in FIG.

  FIG. 18 is a flowchart showing an operation for updating the “packet information” of the management file shown in FIG. 17 during recording. This operation is performed by updating “GOP start position”, “VPTS”, “APTS”, “I end position”, and “P end position” in the flowchart of the operation for updating the “packet information” of the management file shown in FIG. The operation part to be output is changed to steps S181 to S186, and the same operation part as in FIG.

  In FIG. 18, “packet start position” is output to the controller 19 at the head of the packet (step S51). Next, “GOP start position”, “VPTS”, and “APTS” are output to the ECC unit 20 in step S181. Next, when the input of the encoded data of the I picture from the video codec 16 is completed, the “I end position” is output to the ECC unit 20 (step S183), and the input of the encoded data of the first P picture of the GOP is completed. , “P end position” is output to the ECC unit 20 (step S185).

  Steps S181 to S185 are executed at the head of the GOP, and position information and time information for each GOP are output when the input of the encoded data of the P picture from the video encoder 16 is completed. Video encoded data and audio encoded data are output.

It is a block diagram of an embodiment of a moving image recording / reproducing apparatus of the present invention. It is a figure which shows the structure of the multiplexed data in 1st Embodiment. It is a flowchart which shows the operation | movement which outputs the multiplexed data in 1st Embodiment. It is a figure which shows the structure of the management file in 1st Embodiment. It is a flowchart which shows the operation | movement which comprises the management file in 1st Embodiment. It is a figure explaining a mode that the gap of an audio frame and a video frame accumulates. It is a figure explaining the deviation | shift amount of an audio frame and a video frame. It is a figure explaining the deviation | shift amount of the audio frame after a random access, and a video frame. It is a flowchart explaining the operation | movement which synchronizes an audio frame and a video frame. It is a figure explaining the data read from a disk at the time of high-speed reproduction. It is a figure explaining a user program. It is a figure which shows the structure of the other multiplexed data in 1st Embodiment. It is a flowchart which shows the operation | movement which outputs the other multiplexed data in 1st Embodiment. It is a figure which shows the structure of the management file in 2nd Embodiment. It is a flowchart which shows the operation | movement which comprises the management file in 2nd Embodiment. It is a figure which shows the structure of the multiplexed data in 3rd Embodiment. It is a figure which shows the structure of the management file in 3rd Embodiment. It is a flowchart which shows the operation | movement which comprises the management file in 3rd Embodiment. It is a figure explaining GOP. It is a figure explaining DUT in a prior art example. It is a figure explaining the operation | movement at the time of the high speed reproduction in a prior art example. It is a block diagram of the conventional moving image encoder.

Explanation of symbols

11 Audio input terminal 12 Audio output terminal 13 Video input terminal 14 Video output terminal 15 Audio codec 16 Video codec 17 Multiplexing / demultiplexing circuit 18 Memory 19 Controller 20 ECC unit 21 Disk 22 GOP period 23 Audio period 41 Disk information 42 File information 43 Free space information 44 Packet information 61 GOP period 62 Audio period 63 GOP period 64 Audio period 65 Freeze period 71 GOP period 72 Audio period 73 Audio period 74 Deviation 81 GOP
82 GOP
83 GOP
84 GOP
101 I picture 102 B picture 103 B picture 104 P picture 111 GOP
112 GOP
113 GOP
114 GOP
115 GOP
116 GOP
201 DUT
202 DUT header 203 Sub-picture data 204 Audio data 205 Main audio data 211 DUT
212 DUT
213 I picture encoded data 214 I picture encoded data 221 Audio input terminal 222 Audio output terminal 223 Video input terminal 224 Video output terminal 225 Audio codec 226 Video codec 227 Multiplexing / separating circuit 228 Memory 229 Controller 230 ECC unit 231 Recording media

Claims (3)

A moving picture reproducing apparatus for reproducing multiplexed data from a recording medium in which multiplexed data and management data in which at least audio encoded data and video encoded data are multiplexed are recorded as separate files,
The management data includes at least position information and reproduction time on the recording medium of multiplexed data,
The position information includes, for multiplexed data of a frame group composed of a plurality of frames, a start position of the frame group and an end position of a frame to be intra-coded,
The playback time includes the playback time of the first video frame of the frame group,
A moving picture reproducing apparatus, wherein random access is performed using the management data. The multiplexed data is recorded on the recording medium by dividing it into a fixed length, and the multiplexed data of the frame group is recorded in units of the fixed length without excess or deficiency,
The moving image reproducing apparatus according to claim 1. A recording medium in which multiplexed data obtained by multiplexing at least audio encoded data and video encoded data and management data are recorded as separate files,
The management data includes at least position information and reproduction time on the recording medium of multiplexed data,
The position information includes, for multiplexed data of a frame group composed of a plurality of frames, a start position of the frame group and an end position of a frame to be intra-coded,
The playback time includes the playback time of the first video frame of the frame group;
A recording medium characterized by the above.