SMPTE ST 351M-2000 Transporting MPEG-2 Recording Information Through High-Definition Digital Interfaces.pdf

1、 STD-SMPTE 35LM-ENGL 2000 8357401 0004275 9T9 SMPTE STANDARD SMPTE 351 M-2000 for Television - Transporting MPEG-2 Recoding Information through High-Definition Digital Interfaces Contents Scope Normative references General Definitions MPEG-2 recoding information Macroblock rate information MPEG-2 r

2、ecoding information transport mechanism Picture rate information Scope This standard specifies an embedded transport mechanism for the MPEG-2 recoding data set as defined in SMPTE 327M for the representation of MPEG-2 recoding information on a SMPTE 274M interface and subsequently upon a SMPTE 292M

3、bit-serial digital interface. The recoding data set is derived from an ISO/IEC 1381 8-112 compliant MPEG bitstream during the decoding process, as described in the ISOAEC 13818-112 standards. For the minimum operation of this standard, the MPEG-2 recoding data set shall be spatially and temporally a

4、ligned to each decoded macroblock mapped into a SMPTE 274W292M interface. This standard specifies the spatially and temporally aligned transport of the MPEG-2 recoding data set within the active picture area on SMPTE 274W292M interfaces for equipment that complies with the ISO/IEC 13818-112 standard

5、s, including 4:2:2PQHL and MPQHL for 60- and 50-Hz interlaced and 60-, 30-, 25-, and 24-Hz progressive video standards. Page 1 of 11 pages 2 Normative references The following standards contain provisions which, through reference in this text, constitute provisions of this standard. All standards ar

6、esubject to revision and parties to agreements based on this standard are encouraged to investigate the possibility of applying the most recent edition of the standards indicated below. SMPTE 274M-1998, Television - 1920 x 1080 Scan- ning and Analog and Parallel Digital Interfaces for Multiple Pictu

7、re Rates SMPTE 292M-1998, Television - Bit-Serial Digital Interface for High-Def inition Television Systems SMPTE 327M-2000, Television - MPEG-2 Video Recoding Data Set ITU-T H.222.0 with amend 1-2, ISOAEC 13818-1 (1 996-04), Information Technology- Generic Coding of Moving Pictures and Associated A

8、udio Information: Systems ITU-T H.262 with amend lhrr 1 and amend 2/corr 2, ISOAEC 1381 8-2 (1 996-05), Information Technology - Generic Coding of Moving Pictures and Associated Audio Information: Video 3 General The principal application of this standard is to pre- serve the quality of the video si

9、gnal when cascading MPEG-2 decoders and coders for editing or transcod- ing purposes by feeding forward previous coding decisions. Copyright O Zoo0 by THE SOCIETY OF MOTION PICTURE AND TELEVISION ENGINEERS 595 W. Harlsdale Ave., White Pbms, NY 10607 (9141 761-1100 Approved October 23,2000 SMPTE 351M

10、2000 This transport mechanism for the MPEG-2 recoding data set permits the simutaneous processing of both the video and the MPEG-2 recoding data set and, consequently, of the MPEG bitstream. This allows lossless cascading, frame accurate editing, and logo/caption insertion to be performed. The info

11、rmation contained in the MPEG-2 recoding data set is defined in SMPTE 327M. This recoding information shall be temporally locked to the decoded (or partially decoded) video to the nearest MPEG-2 frame or field depending on the picture structure of the coded MPEG-2 bitstream. It shall also be spatial

12、ly locked with the decoded video to the nearest MPEG-2 macroblock within the decoded frame/field. To accrue the full benefits of the recoding information when cascading via a digital baseband interface, the following recommendations shall be adhered to: - The transport mechanism shall preserve at le

13、ast the 8 most significant bits of active video. The mechanism outlined uses the least significant bit of each 10-bit chrominance sample to transmit the data through the SMPTE 274M or 292M interface. NOTE - Users should be aware that some equipment using SMPTE274W292M interfaces may aiter or use the

14、 1 th bit of the video samplesfor other (nonstandardized) purposes. Where this occurs, the recoder will be unable to benefii from the MPEG-2 recoding information. - It is also necessary for the recoding information to be aligned with the decoded MPEG macroblocks in the decoded pictures, both spatial

15、ly and tempo- rally. This standard is based on producing a SMPTE 274M/292M compliant output to cover HD-MPEG bitstreams up to and including 4:2:2PQHL and MPQHLfor 1920x 108060(59.94) or502:1,1920 x 1080 30 (29.93, 25 or 24 (23.98) 1:1, and 1280 x 720 60 (59.94) 1 :1 systems. 4 Definitions 4.1 aligne

16、d: As applied usually to recoding in- formation arranged so that information relating to a macroblock (or other region) is embedded within that region of the decoded signal. 4.2 bitstream: An ordered series of bits con- forming to ISOAEC 13818-2. 4.3 bottom field: One of two fields that comprise a f

17、rame. Each line of a bottom field is spatially located immediately below the corresponding line of the top field. 4.4 cascading: This describes the process where video that has once been coded (com- pressed) is subsequently decoded and coded once more. This cascaded step could carry on for any numbe

18、r of generations. 4.5 chrominance: The color-difference samples defined in ITU-R BT.709-4. 4.6 coding: This is the process by which an un- compressed video sequence is compressed to a bitstream that conforms to ISOAEC 1381 8-2. 4.7 column: A vertical column of macroblocks spanning the full height of

19、 the decoded picture (columns are numbered from left to right starting at zero). 4.8 CRC: Cyclic redundancy check. A class of error detecting codes used in this standard to detect errors when the recoding information and the decoded pictures are being edited or proc- essed in a SMPTE 274M or 292M st

20、ream by external equipment. 4.9 decoded video, decoded picture: Output video, generated by an ISOAEC 1381 8-2 compli- ant decoder that complies with the SMPTE 274M or 292M interface standard. 4.10 decoder: A compressed bitstream decoder that complies with ISOAEC 1381 8-2. 4.11 DTS: Decoding time sta

21、mp (see ISO/IEC 1381 8-1). 4.12 embedded: As applied usually to the re- coding information conveyed within a digital video signal so as to be capable of being passed through digital video equipment. 4.13 macroblock Defined in ISOAEC 13818-2 as a block of 16 x 16 luminance pixels. 4.14 macroblock rat

22、e information: T h is c o r r e- sponds to the coding information from the ISO/IEC 1381 8-2 bitstream that relates to the individual macroblocks as defined in this standard. Page 2 of 11 pages 4.15 MPEG profildlevel: As defined in ISOAEC 1381 8-2. 4.16 MPEG-2 recoding information: This com- prises t

23、he elements defined in SMPTE 327M and the additional information described in this standard required for the practical transport and use of the MPEG-2 recoding data set. SMPTE 351 M-2000 These additional elements are required to allow the video and the MPEG-2 recoding information to be simultaneousl

24、y edited using SMPTE 274W292M based equipment. The MPEG-2 recoding information is subsequently aligned within the macroblock as described in figure 1. 6 Macroblock rate information 4.17 picture: As defined in ISO/IEC 13818-2. 4.18 picture rate information This corresponds to the coding information f

25、rom the ISOAEC 1381 8-2 bitstream. that relates to the whole pic- ture as defined in this standard. 4.19 PTS: Presentation time stamp (see ISOAEC 13818-1). 4.20 recoding data set: The set of information defined in SMPTE 327M. 4.21 stripe: A horizontal row of macroblocks spanning the full width of th

26、e decoded picture (stripes are numbered from top to bottom start- ing at zero). 4.22 sufficient: As applied usually to the re- coding data set containing the necessary infor- mation to enable transparent recoding (in a mathematical sense) of the video signal. 4.23 top field: One of two fields that c

27、omprise a frame. Each line of a top field is spatially located immediately above the corresponding line of the bottom field. 4.24 transcoding: A conversion within the MPEG-2 stream domain, such as bitrate chang- ing or changing the group of pictures (GOP) structure. 4.25 video: A signal conforming t

28、o SMPTE 274M in this standard. 5 MPEG-2 recoding information MPEG-2 recoding information comprises: - The elements defined in SMPTE 327M. - Additional elements from both macroblock and picture rate information as described in this standard. The contents of the MPEG-2 recoding data set relat- ing to

29、each rnacroblock shall be spatially as well as temporally aligned with the decoded video pixels re- lating to that macroblock. The elements in table 1 are at picture rate in the MPEG-2 stream, but are inserted into MPEG-2 recod- ing information at both picture and macroblock level to create the alig

30、nment required to allow editing of the resultant SMPTE 274W292M signal. The elements in table 2 are required in the transport mechanism to synchronize the video information and the MPEG-2 recoding information. Where reference is made to elements in ISOAEC 13818-2, these are indicated in the text in

31、italics. 7 MPEG-2 recoding information trans- port mechanism This comprises 256 bits of data per macroblock that shall be placed bit by bit onto the least significant bit of each 1 W chrominance sample in the decoded video. The decoded 4:2:2 component video macroblock cor- responds to a matrix of 16

32、 x 16 luminance compo- nents and two matrices of 8 x 16 coiordiference components. Hence, in total there are 256 pixels of luminance and 256 pixels of chrominance in each macroblock. In SMPTE 292M video, the multiplexed order for these pixels is Cb,Y,Cr,Y. To increase the resilience of the detection

33、 of edits in the macroblock formatted recoding information, a form of parity scrambling is applied to the data before being placed in the least significant bit of the 10-bit chrominance samples. The parity scrambling scheme used is shown in figure 4, where the bits represent the bits of 10-bit video

34、 (Le., bits 2 through 9 correspond to the 8-bit video). Page 3 of 11 page6 Line Frame RH Coding Coding O 1 2 3 4 5 O 1 2 6 7 3 8 9 4 10 11 5 12 13 6 7 14 15 Figure 1- Macroblock format of the MPEG-2 recoding information Page 4 of 11 pages Parameter to pf i e I df i rs t rep e atf i rs tf i el d 422-

35、420-c hroma picrate-element picrate-element31:16 picrate_elementl5:0 q-scaletype I reserved I I These blocks have no meaning and all the bits shall be set to O. No. bits Definition 1-bit flag I-bit flag Set to a value equal to top-field-first held in the original bitstream and also indicates, along

36、with repeat-firstfield, the temporal align- ment of the recoding information with its associated video. Set to a value equal to repeat-firstfield held in the original bit- stream. The contents of the recoding information of the first field must also be repeated, as indicated by this flag. A value of

37、 O indicates the bitstream was 4:2:0 and chrominance up- sampling was performed to output 4:2:2 video. A value of 1 indicates the bitstream was 4:2:2 and no chrominance filtering was performed. Represents part of the picture rate information for the present picture and has its content dispersed with

38、in the decoded picture. This is defined in further detail in clause 8. Set to a value equal to q-scale-type held in the original bitstream. 1 -bit flag 16-bit ui 16-bit ui 1 -bit flag I Parameter srib-sync-code rolling-srib-mb-ref srib-crc srib_crc31 :I 61 srib_crc15:0 Table 2 - Additional transport

39、 elements No. bits 5-bit flag 1 -bit flag 16-bit Ui 16-bit Ui 16-bit U D ef i n it i on I This is a fixed bit-string 11 11 1 which shall be used to indicate the left alignment of the first row of each macroblock. This flag shall be set to the value 1 when picture-structure is set to frame picture st

40、ructure (value 11) and indicates that the MPEG-2 recoding information is distributed over 16 frame lines. If picture-structure is not set to frame (value 11) then the flag is set to O and the recoding infor- mation is distributed over 16 field lines. This mechanism ensures that the recoding informat

41、ion remains spatially and temporally locked to the corresponding pixels in the decoded video frames/fields. The distribu- tion of the recoding information for frame pictures and field pictures is further shown in figures 2 and 3. This is a 16-bit modulo 65521 rolling reference value. This rolling re

42、ference increments on every macroblock of the transmitted picture. The count shall be continuous across transmitted picture boundaries. This value shall be initialized at start-up to a number selectable between O and 65520 inclusive. This is to allow systems of decoders to be built with unique recod

43、ing information identifiers. A 32-bit CRC to allow error detection for each macroblock of the MPEG-2 recoding information. A model for the operation of this 32-bit CRC is defined in annex A of ISO/IEC 1381 8-1. The CRC is calculated on the 224 data bits of the macroblock of MPEG-2 recoding informati

44、on taken in raster order, using the following generator polynomial: x32 + x26 + x23 + x2* + Xl6 + XI2 + x” + XO + x8 + x7 + x5 + x4 + x2 + x + 1. ) The largest 16-bit derived prime number was chosen to ensure that there is a minimum coherence between the position of the macroblock and its equivalent

45、 address in subsequent pictures. Page 5 of 11 pages STDmSMPTE 35LM-ENGL 2UUO 8357L1UL UOOLi280 2bb D SMPTE 351 M-2000 top feld O 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 tap field O 1 2 3 4 5 6 7 9 10 11 12 13 14 15 a battom feld 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Stripe O Column O Stripe 1 Column O O

46、dd lines of macroblack format fl Re-Coding Information ven lines of macroblock format ReCoding Information bdtorn field 0 1 2 3 4 5 6 7 9 10 11 12 13 14 15 a Figure 2 - Recoding information in frame coded pictures (fr-fl-srib=l) Macroblock format RaCoding Information for Top Field Stripe O Column O

47、I I I l Macroblock format Recoding Information for Bottom Field Figure 3 - Recoding information in field coded pictures (fr-fi-sribd) Page 6 of 11 pages Consider the first few video samples in a stream as being CbO, YO, CrO, Y1, Cbl, Y2, crl, Y3, etc. The panty for scrambling the macroblock formatte

48、d recd- ing information inserted into each chrominance sam- ple is derived by combining the parity of the chrominance sample with the parity of the subsequent luminance sample in the video steam. For example, the parity used to scramble the macro- block formatted recoding information for sample CbO

49、above is the combined parity of CbO and YO, the scrambling of data for CAI is taken from the combined parity of CIO and Y1, etc. A detailed example of the macroblock formatted recoding information bits encoded into the SMPTE 292M samples for the first four luminance and asso- ciated chrominance samples is shown in table 3. 8 Picture fate information This clause defines the picture rate information re- quired in the recoding information, and table 4 shows the elements which must be included. SMPTE 351M-2W The picture rate information must all pass without error in o