JP5703532B2 - Transcoding device - Google Patents

Description

  The present invention relates to a transcoding device that converts compressed data into compressed data of a different format.

  Once the compressed data is decoded, the process of changing the gradation, changing the image size, changing the frame rate, changing the presence / absence of interlace, changing the image aspect ratio, changing the crop, etc. is performed, and transcoding is performed again as compressed data ing. For example, when compressed AV data with high image quality is distributed via a portable device or the Internet, there is a problem in that an appropriate display cannot be performed because the data amount is large. Therefore, in such a case, transcoding for converting high-quality compressed AV data into low-quality compressed AV data for distribution is performed.

  In addition, the compression method is changed at the time of recompression (for example, MPEG2 is changed to MPEG4).

  In the past, if such transcoding was interrupted, once this was interrupted, the transcoded part had to be discarded and the transcoding had to be performed again from the beginning. This will be described below.

  For example, consider the case of transcoding MPEG2 data to MPEG4. FIG. 1A shows MPEG2 original data. I represents a compressed frame that can be decoded independently without referring to other frames (referred to as I frame). P represents a compressed frame that can be decoded by referring to the preceding I frame (referred to as a P frame). B represents a compressed frame that can be decoded by referring to both the preceding I or P frame and the subsequent I or P frame (referred to as a B frame). The symbols F1, F2,... Attached below I, P, B correspond to the frame numbers of the original data. Note that the frame numbers F1, F2,... In FIG. 1A are described for explanation, and the frame numbers are not described in the MPEG2 data.

  When the original data in FIG. 1A is decoded, the original data as shown in FIG. 1B is obtained. F1 represents the first frame, F2 represents the second frame, F3 represents the third frame,. As is clear from the comparison of the MPEG2 data in FIG. 1A and the original data in FIG. 1B, the MPEG2 data does not match the frame number at the time of reproduction and the data recording order. Since the B frame cannot be decoded unless the subsequent frame is referred to, such a recording order is used.

  For example, since the B frame of F1 in FIG. 1A cannot be decoded without referring to the I frame of F3, the I frame of F3 that is played back before the F frame B that is played back earlier during playback and later. Is recorded. The same applies to the B frame of F2.

  When the original data as shown in FIG. 1B is obtained by decoding, the original data is then encoded to obtain MPEG4 data as shown in FIG. 1C. In this way, transcoding from MPEG2 to MPEG4 is performed.

  Assume that the transcoding is interrupted when recording is performed up to the B frame of F4 in FIG. 1C during the above transcoding.

  After that, when transcoding is resumed, it is difficult to specify from which frame decoding / encoding should be started. The frame F4 recorded at the end of the transcoded and recorded MPEG4 data does not coincide with the last frame of the corresponding original data, and the frame number is not recorded in the MPEG4 data. Therefore, it becomes difficult to specify from which frame transcoding should be resumed when resuming.

  In order to solve such a problem, Patent Document 1 discloses that transcoding is performed in units of GOPs, which are a group of data, and transcoding is restarted from the beginning of the interrupted GOP.

JP 2001-094935 A

  However, in the method of Patent Document 1, it is difficult to use a B frame as compressed data after transcoding. The reason is as follows. For example, nine frames F3 to F11 shown in FIG. 1B constitute one GOP. Therefore, in the method of Patent Document 1, transcoding is performed for these nine frames. When transcoding is performed as usual, the last two frames of the compressed data become B frames (F10 and F11). These two B frames refer to the I frame (F12) which is the next GOP. However, reference cannot be made across GOPs in this way. For this reason, in the compressed data after transcoding, the B frame referring to the subsequent frame cannot be used. As a result, there is a problem that the compression rate after transcoding cannot be increased.

  Further, since the compression rate is not sacrificed, a method of outputting a frame scheduled to be output as a B frame so that there is no frame that refers to the next GOP at the time of interruption can be considered. However, when such interruption processing is performed, if the interruption is forcibly interrupted due to an unexpected power interruption or the like during the interruption processing, the P frame is not normally recorded and it is difficult to resume transcoding. There was a problem of becoming.

  An object of the present invention is to solve the above-described problems and provide an apparatus that can appropriately resume transcoding.

(1) (2) A transcoding device according to the present invention includes at least single compressed frame data that can be decoded independently, and reference compressed frame data that can be decoded by referring to the single compressed frame data, and A device that decodes the first compressed stream configured so that the order of the frame data is different from the order of the original frame data corresponding to the compressed frame data, and re-encodes the decoded stream into the second compressed stream. A decoding means for decoding the first compressed stream to obtain original frame data, and an encoding means for encoding the original frame data obtained by the decoding means to obtain and record a second compressed stream, The encoding means refers to the subsequent frame data in the second compressed stream. When the continuous portion of the reference compressed frame data is recorded, the last playback order in the playback order of each frame of the second compressed stream and the compressed frame data corresponding to the last playback order of the second compressed stream are The number of frames up to the last frame data of the continuous portion of the reference compressed frame data is recorded as log data, and when the processing is resumed after the processing of the decoding means and encoding means is stopped, the last reproduction of the recorded log data is performed It is characterized by further comprising a process restart instruction means for instructing to restart the process from the frame corresponding to the next order.

  Therefore, it is possible to resume the transcoding once interrupted and continue the transcoding.

(3) The transcoding device according to the present invention is characterized in that the encoding means records the log data including the earliest reproduction order in a group of frames including the frame of the last reproduction order. .

  Therefore, it is possible to easily find a group of frames including a frame to be resumed when resuming.

(4) In the transcoding device according to the present invention, the encoding means records the log data including the type of the frame in the last reproduction order, and the process restart instruction means records the last data in the log data. With reference to the type of frame in the reproduction order, an instruction is given to restart the processing from the single compressed frame data nearest to the already recorded second compressed stream.

  Therefore, it is possible to prevent the density of the single compression frame from increasing and the compression rate from decreasing.

  In the embodiment, “decoding means” corresponds to step S1.

  In the embodiment, “encoding means” corresponds to step S2.

  In the embodiment, “process resumption instruction means” corresponds to steps S10 to S15.

  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.

It is a figure for demonstrating transcoding. 1 is a functional block diagram of a transcoding device according to one embodiment. FIG. It is a figure for demonstrating the process of a transcoding apparatus. It is a figure which shows a hardware configuration. 3 is a flowchart of a transcode program 22. It is a figure which shows the example of log data. 3 is a flowchart of a transcode program 22. It is a figure for demonstrating the process of a transcoding apparatus.

1. 2 is a functional block diagram of a transcoding device according to an embodiment of the present invention. The decoding means 2 decodes the first compressed stream and restores the original frame data. The encoding unit 4 encodes the original frame data to generate a second compressed stream.

  When the encoding means 4 records the continuous portion of the reference compressed frame data referring to the subsequent frame data in the second compressed stream, the encoding means 4 and the second playback order in the playback order of each frame of the second compressed stream, The number of frames from the compressed frame data corresponding to the last reproduction order of the compressed stream to the last frame data of the continuous portion of the reference compressed frame data is recorded as log data 6. For example, when the B frame of F2 in FIG. 1C is recorded, the largest frame number (F3 in this case) and the number of frames from the F3 to the last frame F2 recorded (2 in this case) are recorded. That is, the log data 6 is not recorded when I of F3 and B of F1 are recorded, and the log data 6 is recorded when B of F2 is recorded. Therefore, the log data 6 is recorded with information up to the point when the recording of a set of frames starting with I or P, such as IBB and PBB, among the already recorded compressed frames is completed. .

  For example, as shown in FIG. 3C, it is assumed that transcoding is interrupted when a compressed frame is recorded up to B of F4. In this case, as shown in FIG. 3D, “F3” is recorded as the last reproduction frame number, and “2” is recorded as the number of frames from the frame to the final recording frame.

  When restarting transcoding, the process restart instruction means 8 reads the recorded log data 6. Based on the last reproduction frame number “F3” shown in the log data 6, the compressed data corresponding to the reproduction order is searched. For this purpose, for example, the process resumption instructing means 8 performs decoding from the beginning of the GOP of the recorded compressed frame after transcoding (F3 I in FIG. 3C) and searches for the compressed frame in the reproduction order.

  In this embodiment, the frame number (here, F1) of the frame with the earliest reproduction order is recorded in the frame in the GOP. Therefore, by decoding, it is possible to specify the compressed frame (I of F3 here) to be reproduced last.

  Therefore, the process resuming instruction unit 8 gives an instruction to the decoding unit 2 and the encoding unit 4 so as to resume transcoding from the frame number F4 next to the frame number F3.

  Further, the process resumption instructing means 8 reads out the number of frames (in this case 2) from the last reproduced frame F3 to the last recorded frame F2 indicated in the log data 6. Then, the remaining frames from the frame F3 to the last two compressed frames are left and the subsequent compressed frames are deleted. In the case of FIG. 3C, the frames F6 and F4 are deleted.

  In this way, the transcoded data is recorded as shown in FIG. 3E. When restarting, the first frame is recorded as an I frame.

2. Hardware Configuration FIG. 4 shows a hardware configuration of a transcoding device according to an embodiment. A memory 12, a hard disk 14, and a CD-ROM drive 16 are connected to the CPU 10.

The hard disk 14 stores an operating system 20, a transcode program 22, and MPEG2 data 24 to be transcoded. The transcode program 22 performs its function in cooperation with the operating system 20. The operating system 20 and the transcode program 22 are the ones recorded on the CD-ROM 18 and installed on the hard disk 14 via the CD-ROM drive 16.

3. Transcode Processing Here, a case where MPEG2 data is transcoded to MPEG4 data will be described.

  FIG. 5 shows a flowchart of the transcode program 22 recorded on the hard disk 14. The CPU 10 reads and decodes the MPEG2 data recorded on the hard disk 14 and generates an original frame (step S1). At this time, the CPU 10 counts the frame number of the generated original frame and associates it with each original frame. The generated original frame and frame number are temporarily recorded in the memory 12 or the hard disk 14. For example, the MPEG2 data shown in FIG. 3A is read to obtain the original frame shown in FIG. 3B.

  Subsequently, the CPU 10 encodes the original frame obtained by decoding into MPEG4. When the encoding for the set of original frames F1, F2, and F3 is completed, the CPU 10 records MPEG4 data on the hard disk 14 (step S3). For example, an I frame of F3, a B frame of F1, and a B frame of F2 shown in FIG. 3C are recorded. However, frame numbers are generally not recorded in compressed video files. Therefore, the frame number is not recorded in this embodiment.

  When a set of MPEG4 data is recorded on the hard disk 14, the CPU 10 records log data on the hard disk 14 (step S4). An example of log data is shown in FIG.

The “maximum frame number” is the largest frame number among the already recorded MPEG4 data. Here, “3” is recorded as the largest frame number corresponding to F3. “The number of frames until the last recording frame” is the number of frames from the position where the frame having the highest number is recorded to the frame recorded at the last position. Here, the number of frames “2” from the I frame of F3 in FIG. 3C to the B frame of F2 is recorded. “Frame type” is the type of frame having the largest frame number. Here, “ I ” is recorded. The “number of bytes from the beginning of the GOP including the maximum frame number” is the number of bytes from the beginning of the compressed data at the beginning of the GOP including the frame having the maximum frame number in the compressed data after transcoding. Here, for example, 10 bytes are recorded. The “first frame number of GOP” is the frame number of the first frame of the GOP. Here, “1” is recorded.

  When the recording of the log data is finished, the CPU 10 determines whether or not the transcoding is finished (step S5). If not completed, step S1 and the subsequent steps are repeatedly executed. As a result, MPEG2 data is converted into MPEG4 data and recorded on the hard disk 14. At the same time, log data is recorded one after another as shown in FIG.

  If it is determined in step S5 that transcoding has ended, the CPU 10 ends the process.

  As described above, transcoding is executed. Note that there may be an arbitrary interruption during transcoding or an unexpected error. In this case, a process for restarting transcoding will be described below. FIG. 7 shows a flowchart when the transcode program 22 resumes transcoding.

  Here, as shown in FIG. 8C, a case will be described as an example in which the recording is interrupted due to an unexpected error at the time of recording up to P of F20010 and B of F20008. In FIG. 8, F20010 is abbreviated as F10. F20010 P, F20008 B, and F20009 B are written immediately after the decoding and encoding process up to the frame F20010 is completed. Therefore, normally, these sets of F20010 P, F20008 B, and F20009 B are written together. However, when the process is interrupted due to an unexpected error, as shown in FIG. 8C, a situation may occur in which the process ends when recording is performed up to P of F20010 and B of F20008.

  At the time of the interruption as described above, since P of F20010, B of F20008, and B of F20009 are not recorded, log data for these is not recorded. That is, the log data is in a state where recording is performed for P of F20007, B of F20005, and B of F20006.

  First, the CPU 10 reads the last log data recorded on the hard disk 14 (step S10). Next, the maximum frame number is acquired from the log data in FIG. Here, “20007” is acquired. Next, the CPU 10 acquires the number of bytes from the beginning of the file up to the beginning of the GOP from the log data. Here, “120000000” bytes are acquired. Further, the top frame number of the GOP is acquired (step S11). Here, “20002” is acquired.

  The CPU 10 decodes frame by frame from the beginning of the GOP of MPEG4 data that has already been transcoded and recorded, and identifies the compressed frame having the maximum frame number (step S12). That is, P of F20007 in FIG. 8C is specified. At this time, the head of the GOP can be easily found based on the acquired number of bytes. Furthermore, since the first frame number of the GOP is known, a compressed frame corresponding to the maximum number of frames can be easily found.

  Next, the CPU 10 acquires “the number of frames up to the final recording frame” (step S13). Here, “2” is acquired. The CPU 10 deletes the compressed frame after the frame added with “the number of frames until the final recording frame” from the compressed frame having the specified maximum frame number (step S14). Here, the data from P of F20007 in FIG. 8C to B of F20006, which is two after, is left, and the compressed data after that (P of F20010, B of F20008) is deleted. Thereby, the recorded compressed data (MPEG4 data) becomes as shown in FIG. 8D. This is to delete compressed data that is not recorded as a complete set.

  Subsequently, the CPU 10 sets the frame number next to the maximum frame number as the frame number to be resumed (step S15). Therefore, the CPU 10 decodes the MPEG2 data and searches for a frame corresponding to the restart frame number. Thereafter, transcoding is executed according to the flowchart of FIG. Thereby, as shown in FIG. 8E, recording is resumed from the I frame of F20010. As described above, transcoding can be resumed.

3. Other embodiments
(1) In the above embodiment, depending on which part is interrupted, the place to become the P frame becomes the I frame, and the density of the I frame increases and the compression degree decreases. This is because the first compressed frame recorded at the time of restart is an I frame. In order to deal with this problem, control may be performed so that transcoding is resumed from the latest I frame in the compressed data (MPEG4) with reference to log data. In this case, in step S14, the recorded compressed frames equal to or less than the I frame are deleted.

(2) In the above embodiment, the case where MPEG2 data is transcoded to MPEG4 data has been described. However, once the compressed data is decoded, a process such as changing the gradation, changing the image size, changing the frame rate, changing the interlaced presence, changing the image aspect ratio, changing the crop, etc. is performed again. The code can be applied in general.

(3) In the above embodiment, the frame number is used as the playback order, but PTS or the like may be used as the playback order.

Claims (6)

At least single compressed frame data that can be decoded independently, and reference compressed frame data that can be decoded by referring to the single compressed frame data, and the order of each compressed frame data is an original frame corresponding to these compressed frame data A device that decodes a first compressed stream configured differently from the data order and re-encodes the decoded stream into a second compressed stream,
Decoding means for decoding the first compressed stream to obtain original frame data;
Encoding means that encodes the original frame data obtained by the decoding means to obtain and record a second compressed stream;
When the encoding means records the continuous portion of the reference compressed frame data that refers to the subsequent frame data in the second compressed stream, the encoding means, and the second playback order of the frames in the second compressed stream, Recording the number of frames from the compressed frame data corresponding to the last playback order of the compressed stream to the last frame data of the continuous portion of the reference compressed frame data as log data,
A process resumption instruction means for instructing to resume the process from the frame corresponding to the next order of the reproduction order of the last recorded log data when the process is resumed after the processes of the decoding means and the encoding means are stopped. A transcoding device characterized by that. At least single compressed frame data that can be decoded independently, and reference compressed frame data that can be decoded by referring to the single compressed frame data, and the order of each compressed frame data is an original frame corresponding to these compressed frame data A transcoding program for realizing, by a computer, a device that decodes a first compressed stream configured differently from the data order and re-encodes the decoded stream into a second compressed stream. ,
Decoding means for decoding the first compressed stream to obtain original frame data;
A program for functioning as an encoding unit that encodes the original frame data obtained by the decoding unit to obtain and record a second compressed stream,
When the encoding means records the continuous portion of the reference compressed frame data that refers to the subsequent frame data in the second compressed stream, the encoding means, and the second playback order of the frames in the second compressed stream, Recording the number of frames from the compressed frame data corresponding to the last playback order of the compressed stream to the last frame data of the continuous portion of the reference compressed frame data as log data,
The program restarts the computer from a frame corresponding to the next sequence of the last reproduction order of the recorded log data when the processing is resumed after the decoding unit and the encoding unit stop processing. A transcode program that functions as a process resumption instruction means for instructing Oite the equipment of claim 1,
Said encoding means, said as log data, most including behind playback order of the frame, equipment, characterized in that the recording including the earliest reproduction order in the frame of the loaf. In 請 Motomeko 2 of the program,
Said encoding means, said as the log data, including the rearmost reproduction order of the frame, characterized and to Help program to be recorded including the earliest reproduction order in the frame of the loaf. Oite the equipment of claim 1 or 3,
The encoding means records the log data including the type of the frame in the last reproduction order,
The process resumption instructing unit resumes the process from the single compressed frame data nearest to the second compressed stream already recorded with reference to the type of the frame in the last reproduction order in the log data. equipment characterized in that it instructs. In program according to claim 2 or 4,
The encoding means records the log data including the type of the frame in the last reproduction order,
The process resumption instructing unit resumes the process from the single compressed frame data nearest to the second compressed stream already recorded with reference to the type of the frame in the last reproduction order in the log data. program characterized by instructing.

Images