JP2007189570A - Picture alignment method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a picture alignment method capable of improving the image quality of an MPEG stream. <P>SOLUTION: In the picture alignment method, stuffing data 18 of video ES (Elementary Stream) are inserted after a B picture 17 immediately before an I picture and after a B picture immediately before a P picture when heads of the I picture and the P picture 16 are aligned immediately after a pack header with the video ES being an IBBP configuration in the case of multiplexing the video ES and an audio ES to generate an MPEG2-PS. Stuffing data are inserted in a video encoding part for encoding a video signal after finishing the encoding of the B picture immediately before the I picture or the P picture. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a picture alignment method suitable for use in creating a Moving Picture Experts Group (MPEG) stream.

  FIG. 4 is a diagram showing a data structure of MPEG2-PS (Program Stream). 4A shows an example of MPEG2-PS, and FIG. 4B shows an example of a video ES (Elementary Stream) in MPEG2-PS. In MPEG2-PS, the system layer is the highest layer, Below this, a video layer (video ES) and an audio layer (audio ES) are arranged.

  The video ES has a configuration in which an I picture, a B picture, and a P picture are repeatedly arranged in the order of, for example, IBBPBBPBBPBBPBB (IBBP configuration), and is divided in units of packs and stored in a system layer pack. Also, in MPEG2-PS, search and editing at the time of reproduction must be handled in units of pictures. Therefore, it is necessary to align pictures so that the top of the picture is aligned immediately after the header located at the top of the pack.

  FIG. 5 is a diagram for explaining a conventional picture alignment method. The example of FIG. 5 shows a case where the head of the P picture is aligned immediately after the pack header when the video ES has an IBBP configuration, and the audio ES is not shown.

  In the example of FIG. 5, the beginning of the fourth P picture is aligned immediately after the pack header by inserting padding PES (Packetized Elementary Stream) after the third B picture. As described above, conventionally, when the head of a picture is aligned immediately after the pack header, padding PES is inserted after the picture immediately before the picture to be aligned.

  Here, the MPEG2-PS system bit rate is Xa, of which the header of the system layer is Xb, the pad PES for picture alignment is Xc, the audio ES is Xd, and the video ES is Xe. Then, video ES minutes Xe is

It becomes. However, since the size of each picture of the video ES is variable depending on the picture, it is impossible to know in advance how large the pad PES for picture alignment will be, and after the video ES of each picture is generated Since it can be known for the first time, it is necessary to consider the maximum size that can be taken by the pad PES for picture alignment, and to estimate the bit rate in advance.

  Here, the maximum size that can be taken by the pad PES for picture alignment is when a pack that contains only the last 1 byte of the picture immediately before the picture to be aligned occurs. In this case, one pack Almost all of this becomes padding PES for picture alignment. For example, if the size of one pack is 2048 bytes, the size of the pack header is 14 bytes, and the size of the PES header is 18 bytes, the size of padding PES for picture alignment is

It becomes. Thus, for example, when the GOP (group of pictures) is an IBBP configuration, 1 GOP is 15 pictures, and 2 GOPs per second, the I picture and the P picture are aligned immediately after the pack header. The number of P pictures is 5 in 1 GOP and 10 in 1 second. Therefore, as the bit rate Xc of padding PES for picture alignment,

Therefore, the allocation of the bit rate Xe of the video ES is reduced by this amount.
JP 2004-079907 A

  Usually, in DVD-video or the like, about 10 Mbps is used as the system bit rate Xa. In this case, if the bit rate Xc of the pad PES for picture alignment is as shown in Equation 3, Although the image quality is not affected so much, if the system bit rate Xa is about 1 Mbps in the extremely narrow band portion of the long-time recording function or communication device, the bit rate Xc as shown in Equation 3 cannot be ignored. become.

  In view of the above, an object of the present invention is to provide a picture alignment method capable of improving the image quality of an MPEG stream.

  According to the picture align method of the present invention, when an elementary stream for each material is multiplexed to generate an MPEG stream, when the head of a predetermined picture is aligned at a predetermined position of a pack, the picture immediately before the predetermined picture After that, the stuffing data of the video ES is inserted.

  According to the present invention, when aligning the beginning of a predetermined picture to a predetermined position of the pack, the stuffing data of the video ES is inserted after the picture immediately before the predetermined picture. Therefore, conventionally, padding PES is inserted. By doing so, the wasted portion can be allocated to the video ES. Therefore, the image quality of the MPEG stream can be improved. This effect becomes more prominent as the system bit rate of the MPEG stream is lower.

  1 to 3, in the case of generating MPEG2-PS according to an embodiment of the present invention, the video ES is an IBBP structure, and the head of the I picture and the P picture is immediately after the pack header. A description will be given taking the case of alignment as an example.

  FIG. 1 is a diagram for explaining an embodiment of the present invention in comparison with a conventional picture alignment method. 1A shows a part of MPEG2-PS when one embodiment of the present invention is applied, and FIG. 1B shows a part of MPEG2-PS when a conventional picture alignment method is applied.

  In FIG. 1, 10, 11, 12 are packs, 13, 14, 15 are pack headers, 16 is an I picture or P picture, 17 is the last part of the B picture immediately before the I picture or P picture 16, and 18 is Stuffing data of video ES, that is, undecoded data positioned in the video ES layer (specifically, zero_bit, zero_byte defined in 5.2.3 Definition of next_start_code () function in ISO 13818-2), 19 Is a padding PES for picture alignment.

  That is, in the case of generating MPEG2-PS by multiplexing video ES and audio ES, which are elementary streams by material, the video ES has an IBBP configuration, and the heads of the I picture and P picture are aligned immediately after the pack header. In conventional cases, padding PES for picture alignment is inserted after the B picture immediately before the I picture and after the B picture immediately before the P picture.

  On the other hand, an embodiment of the present invention inserts video ES stuffing data after the B picture immediately before the I picture and after the B picture immediately before the P picture. The head is aligned immediately after the pack header. The stuffing data is inserted after the encoding of the B picture immediately before the I picture or P picture is completed in the video encoding unit that encodes the video signal.

  In one embodiment of the present invention, as described above, the MPEG2-PS system bit rate is Xa, of which the system layer header is Xb, the picture alignment padding PES is Xc, and the audio If the ES part is Xd and the video ES part is Xe, the padding PES part Xc is assigned to the video ES part Xe.

It becomes. For example, if the system bit rate Xa is 1 Mbps, the system layer header portion Xb is 200 kbps, the picture alignment padding PES portion Xc is 160 kbps, and the audio ES portion Xd is 256 kbps. Applying an embodiment of the present invention,

It becomes. In contrast, when the conventional picture alignment method is applied,

It becomes. As described above, when the conventional picture alignment method is applied, only 384 kbps can be expected as the video ES portion Xe in the system bit rate Xa = 1 Mbps. However, when one embodiment of the present invention is applied, the video ES portion is used. 544kbps can be expected as Xe.

  However, when stuffing data is inserted, the bit rate Xe of the video ES is substantially lower than 544 kbps. However, when the bit rate Xe of the video ES is lowest, 2015 byte stuffing data is inserted for each picture alignment. In this case, the bit rate Xe of the video ES is substantially equal to Is 384 kbps, but otherwise it is higher than 384 kbps.

  On the other hand, in the conventional picture alignment method, the 160 kbps allocated to the pad PES for picture alignment does not become the video ES, so that the bit rate Xe of the video ES is higher than 384 kbps. Absent.

  FIG. 2 is a block circuit diagram showing a part of the MPEG2-PS creation apparatus. 2A shows an MPEG2-PS creation apparatus that implements an embodiment of the present invention, and FIG. 2B shows an MPEG2-PS apparatus that implements a conventional picture alignment method.

  In FIG. 2A, 20 is a video encoding unit that encodes a video signal and outputs a video ES, 21 is an audio encoding unit that encodes an audio signal and outputs an audio ES, and 22 is a video encoding unit 20. Is a multiplexing unit that multiplexes the video ES output by the audio encoding unit 21 and the audio ES output by the audio encoding unit 21 to create MPEG2-PS.

  2B, reference numeral 23 denotes a video encoding unit that encodes a video signal and outputs a video ES, 24 denotes an audio encoding unit that encodes an audio signal and outputs an audio ES, and 25 denotes a video encoding unit 23. A multiplexing unit 26 multiplexes the video ES output from the audio ES and the audio ES output from the audio encoding unit 24 to create an MPEG2-PS, and 26 is a padding PES insertion mechanism for picture alignment.

  That is, in the conventional picture align method, since the pad PES for picture alignment is inserted, the multiplexing unit 25 needs the pad PES insertion mechanism 26 for picture align.

  On the other hand, in the embodiment of the present invention, the stuffing data of the video ES is inserted. This is because the video encoding unit 20 encodes the B picture immediately before the I picture or P picture. In the video ES that is performed after completion and output from the video encoding unit 20, since the heads of the I picture and the P picture are aligned, it is not necessary to provide the padding PES insertion mechanism 26 in the multiplexing unit 22.

  FIG. 3 is a diagram for explaining a control method of a VBV (Video Buffer Verifier) buffer when the embodiment of the present invention is implemented, and shows a change in the occupation amount of the VBV buffer. In one embodiment of the present invention, 2015 bytes of stuffing data may be output at the end of one picture. Therefore, for the VBV buffer, a value higher than the VBV buffer lower limit by 2015 bytes is used as the VBV buffer temporary lower limit, and the VBV buffer temporary limit is always set. Make normal rate control at a value above the lower limit.

  In this way, even if stuffing data with a maximum of 2015 bytes is output at the end of one picture, rate control can be prevented from failing.

  As described above, in one embodiment of the present invention, when video ES and audio ES are multiplexed to generate MPEG2-PS, the video ES has an IBBP configuration, and the beginning of the I picture and P picture is the pack header. When the alignment is performed immediately after the video picture, the stuffing data of the video ES is inserted after the B picture immediately before the I picture and after the B picture immediately before the P picture.

  Therefore, it is not necessary to allocate the fixed bit rate padding PES in the MPEG2-PS system bit rate, and this can be allocated to the video ES, so that the image quality of MPEG2-PS can be improved. it can. This effect becomes remarkable when the bit rate of MPEG2-PS is low.

  In the embodiment of the present invention, the case of generating MPEG2-PS has been described as an example. However, the present invention performs alignment of the beginning of a picture, such as MPEG4, by inserting stuffing data in the video ES layer. It can be applied to all MPEG streams that are possible.

  In the embodiment of the present invention, the case where the GOP has the IBBP configuration has been described as an example. However, the present invention can be applied to any configuration of the GOP, such as when the picture is aligned with IIIII or IPPP. can do.

  Further, in the embodiment of the present invention, when the GOP has the IBBP configuration, the case where the heads of the I picture and P picture are aligned immediately after the pack header has been described. The present invention can also be applied to the case where the head of the B picture is aligned immediately after the pack header, and can be applied as long as one picture has a bit rate of one pack or more as the system bit rate.

It is a figure for demonstrating one Embodiment of this invention compared with the conventional picture alignment method. It is a block circuit diagram which shows a part of MPEG2-PS production apparatus. It is a figure for demonstrating the control method of the VBV buffer in the case of implementing one Embodiment of this invention. It is a figure which shows the data structure of MPEG2-PS. It is a figure for demonstrating the conventional picture alignment method.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10, 11, 12 ... Pack 13, 14, 15 ... Pack header 16 ... I picture or P picture 17 ... Last part of B picture immediately before I picture or P picture 18 ... Stuffing data of video ES 19 ... Picture -Padding PES for alignment
DESCRIPTION OF SYMBOLS 20 ... Video encoding part 21 ... Audio encoding part 22 ... Multiplexing part 23 ... Video encoding part 24 ... Audio encoding part 25 ... Multiplexing part 26 ... Padding PES insertion mechanism

Claims (3)

  When an elementary stream for each material is multiplexed to generate an MPEG stream, if the head of a predetermined picture is aligned with a predetermined position of the pack, the stuffing of the video elementary stream is performed after the picture immediately before the predetermined picture. A picture alignment method characterized by inserting data.   2. The picture alignment method according to claim 1, wherein the stuffing data is inserted after the encoding of the picture immediately before the predetermined picture is completed in the video encoding unit. 3. A picture according to claim 1, wherein a value higher than the lower limit of the VBV buffer by the stuffing data is set as a temporary lower limit of the VBV buffer, and rate control is performed at a value equal to or higher than the temporary lower limit. Align method.

Images