SMPTE ST 374M-2003 SMPTE STANDARD for Television - Mapping of Vertical Ancillary Data Packets and Extended Video Line Data into Video DIF Blocks of DV-Based 50 Mb s DIF Stream Form.pdf

1、 Table of contents 1 Scope 2 Normative references 3 General description 4 Division and control of the mapped space 5 Conversion and mapping process of VANC data packets 6. Extended Video Line compression and mapping coded pixels into the Compressed Macro Block data area 7 Mapping of VANC data and Ex

2、tended Video Line into the CMBs of video DIF Blocks Annex A Abbreviations Annex B Bits from SDI VANC Annex C Structure of the mapping process 1 Scope The purpose of this standard is to specify mapping of the extended video line data and the vertical ancillary data packets (VANC) present in the verti

3、cal blanking interval (VBI) space of the serial digital interface (SDI) into a 50 Mb/s digital interface format (DIF) structure defined in SMPTE 314M. Format of a VANC packet is defined in SMPTE 291M. Mapping of the extended video line data permits an increase of video aperture of 7.5 lines in the 5

4、25 system and an aperture increase of 9 lines in the 625 system. Mapping of VANC packets provides for carriage of metadata and data essence through the 50 Mb/s DV-based DIF stream structure. 2 Normative references The following standards contain provisions which, through reference in this text, cons

5、titute provisions of this standard. At the time of publication, the editions indicated were valid. All standards are subject to revision, and parties to agreements based on this standard are encouraged to investigate the possibility of applying the most recent editions of the standards indicated bel

6、ow: ANSI/SMPTE 259M-1997, Television 10-Bit 4:2:2 Component and 4fscComposite Digital Signals Serial Digital Interface SMPTE 291M-1998, Television Ancillary Data Packet and Space Formatting Page 1 of 19 pages SMPTE 374M-2003 Copyright 2002 by THE SOCIETY OF MOTION PICTURE AND TELEVISION ENGINEERS 59

7、5 W. Hartsdale Ave., White Plains, NY 10607 (914) 761-1100 Approved March 17, 2003 SMPTE STANDARD for Television Mapping of Vertical Ancillary Data Packets and Extended Video Line Data into Video DIF Blocks of DV-Based 50 Mb/s DIF Stream Format SMPTE 374M-2003 Page 2 of 19 pages SMPTE 314M-1999, Tel

8、evision Data Structure for DV-Based Audio, Data and Compressed Video 25 and 50 Mb/s SMPTE RP 168-2002, Definition of Vertical Interval Switching Point for Synchronous Video Switching 3 General description Metadata and data essence can be carried in the ancillary space of the SDI stream structures. T

9、hese data are formatted according SMPTE 291M as ancillary packets and located in VBI space as VANC packets. VANC packets are transformed into digital ancillary packets (DANC) and then mapped into available data space of the 50 Mb/s DV-based DIF stream. The space that is used for mapping data contain

10、ed in VANC packets and extended video line data is located in reserved space within a compressed macro block (CMB). Compressed macro blocks are located in multiple video DIF blocks of a DIF frame. Multiple DIF blocks of 80 bytes each form a DIF frame and are defined in the SMPTE 314M. Due to limited

11、 CMB data space, and to achieve maximum benefit of the available space by the user, management of the data space should be established (see note 1). A specific data management method is beyond the scope of this standard. NOTE 1 Designers should be aware that the size of VANC data can exceed the capa

12、city of the VAUX packs. It is the responsibility of the application to manage the mapping of VANC data into VAUX packs such that only complete VANC packets are mapped into VAUX DIF blocks. Where the VAUX capacity is exceeded, it is the responsibility of the application to manage this mapping in such

13、 a way as to minimize the effect on system performance of not mapping all of the VANC data. Mapping schemes This standard specifies three similar mapping schemes. The first scheme (see 7.1) maps only a single extended video line from the end of a frame. This is either line 525 in the 525-line system

14、 or line 623 in the 625-line system. The second mapping scheme (see 7.2) maps additional information as well as the vertical ancillary data packets contained in the last video line. The third mapping scheme (see 7.3) maps an additional 7.5 lines of the 525-line system or an additional 9 lines of the

15、 625-line system. These include the last line located at the end of a frame. i) Extended video line aperture The purpose of mapping the extended video line data is to expand the video aperture of the existing 50 Mb/s DV-based compressed signal. Video aperture of this compressed signal is nominally d

16、etermined by SMPTE 314M and equals 480 lines in the 525-line system and 576 lines in the 625-line system. The extended video line data mapping technique for expansion of the video aperture uses mild compression that is suitable for video signals and similar types of digital data signals, as is, for

17、example, teletext. The method of compressing this data is defined in this standard. ii) VANC packet mapping A VANC packet is located within VBI space of a SDI stream. The SDI interface operates as a 10-bit interface. However, the VANC data packets carry only 8-bit information and it is located in bi

18、ts b0 to b7 of the VANC packet. The upper two bits of the VANC packet (b8 and b9) can be derived during the reverse conversion process, when the DANC packet is converted back into a VANC packet, and placed back onto the SDI interface. Before the VANC packets are mapped into the video DIF blocks, the

19、y are converted (transformed) into a string of DANC packets. The DANC packets contain all necessary information present in the original VANC SMPTE 374M-2003 Page 3 of 19 pages packet, including VANC header packet information. The only information not directly mapped is the ADF (ancillary data flag)

20、bytes and bit 9 of the CS byte (check sum). The ancillary packet conversion/transform to a DANC packet is a direct conversion technique, fully transparent for data content, and defined in this document. 3.1 Data capacity and structure of the available data area Figure 1 shows organization of a 50 Mb

21、/s DIF frame as defined in SMPTE 314M. Every video DIF block that carries DV-based compressed video essence contains reserved data space used for mapped data. The data payload of a video DIF block is called the compressed macro block (CMB), as shown in figure 2. A compressed macro block contains two

22、 data areas, each occupying the space of two bytes, marked X0, X1. The 12-bits of each CMB data area (X0, X1) is a reserved data space, and used by the mapped data. The remaining lower 4 bits of the X0, X1 data area is set to code 0110. Capacity (see note 2) of the CMB data area for different televi

23、sion systems is calculated below, and represents the total available mapping space in a DIF frame: 525 /60 system (12 bits x 2 areas x 135 compressed macro blocks x 20 DIF sequences) / 8 bits = 8100 bytes 625 /50 system (12 bits x 2 areas x 135 compressed macro blocks x 24 DIF sequences) / 8 bits =

24、9720 bytes NOTE 2 The total data space available for mapping is allocated into two sections, where one section is dedicated to mapping of VANC packets and the other section is dedicated to mapping of extended video line data to achieve a larger video aperture. 4 Division and control of the mapped sp

25、ace (CMB data area) Figure 1 also shows VAUX DIF blocks located in each of the DIF sequences. DIF frame/stream control information is located in the payload space of these VAUX DIF blocks (3 VAUX DIF blocks are contained in each DIF sequence). This control information is called VAUX source pack (VS)

26、 and is defined in SMPTE 314M. The structure of a VAUX source pack is shown in table 1. While SMPTE 314M does not define the extension flag (EXT) control bits (EXT bits shown in SMPTE 314M are marked as “Reserved” bits for future use), these EXT bits are defined in this standard. The EXT control fla

27、g identifies which type of mapped information (VANC or extended video line video content) is located in the CMB data area as indicated in tables 2 and 3. SMPTE 374M-2003 Page 4 of 19 pages Figure 1 Organization of a DIF frame for 50-Mb/s DV-based compression QNOSTADC0Y0X0AC ACDC1Y1AC ACDC2CRACDC3CBA

28、CMSBLSB345 18. 19 32. 33 . 46 47 . 60 61 . 70 71 79Byte position numberX1Size in bytes 14 2 12 14 2 12 10 10MSBLSB4bits8bitsMAPPEDConstructof dataarea(X0,X1)0110AREAFigure 2 Construct of a 50-Mb/s DV-based compressed macro block (CMB) Compressed Macro Block CMB Data ID Byte position number 0 1 2 3 -

29、 - - - - - - - - - - - - - - - - - - - - - - - 79 Video DIF Block DIF sequence 1,0 Second channel DIF sequence n-1,0 DIF sequence 0,0 DIF sequence 1,1 DIF sequence n-1,1 DIF sequence 0,1 Subcode section Header section Audio default value is set to 1. 5.1 Line number generation A local television lin

30、e number counter shall create a LN number. This counter is reset at the beginning of each frame and corresponds to the start of a frame as defined in a relevant system standard. The counter increments value from 1 to n. For 525/60 system: LN0 - LN10 = 1,., 525 For 625/50 system: LN0 - LN10 = 1,., 62

31、5 5.2 Structure of a CMB data area for mapping DANC packet string Figure 4 shows the relationship between a CMB data area (X0, X1) of a video DIF block and the mapped string of bytes corresponding to a string of DANC Packets. The B of the Bm indicates a byte of the string, and UDW (255max) DID SDID

32、/DBN DC ADF b8 = even parity (b7-b0) 10 bits CS (b7- b0) (b7- b0) (b7- b0) (b7- b0) (b7-b0) CS8 MSB LSB b9 b0 b9 = not b8 8 bits LN4 LN3 LN2 LN1 LN0 LN10 LN9 LN8 LE LN6 LN5 LN7 MSB LSB b7 b0 UDW (255max) DID SDID /DBN DC CS (b7- b0) (b7- b0) (b7- b0) (b7- b0) (b7-b0) Res Res Res Res Res Res Res Res

33、Res Res Res Res SMPTE 374M-2003 Page 9 of 19 pages m indicates the order and byte number of the same string. Then, Bm, Bm+1, Bm+2, , show the order of the mapped data corresponding to the DANC Packet string. 5.2.1 String of DANC packets and its mapping The string of DANC packets with a length of S b

34、ytes is numbered from byte 0 to byte (S -1). This string is mapped into the 2nd CMB data area in the sequence and order of the byte sequence number Bm. The LSB of the DANC string is mapped in the order shown in Figure 4. The total size S of all mapped DANC packets should be smaller, or up to the dat

35、a capacity, available in CMB data area assigned for this purpose. If there is an excess of DANC packet data, management of the mapping process may be required as noted in clause 3 (see note 1 on page 2). Figure 4 Relation between CMB data area and a byte number of the DANC packet string 5.2.2 Interl

36、eaving of a DANC packet string into DIF sequences The process of mapping of the DANC packet string starts by mapping byte B0 through B2 (sequence of Bm bytes) into the first video DIF block of the DIF sequence 0 belonging to the first channel of a DIF frame (see figure 1 and table 6 or 7). The subse

37、quent Bm bytes B3 through B5 of the string are mapped into first video DIF block of the DIF sequence 0 belonging to the second channel of a same DIF frame (see figure 1 and table 6 or 7). This process continues until there are no more Bm bytes present. Any unused space within the DANC area shall be

38、filled with FFh. 5.2.3 Placement of VANC packets during data recovery VANC packets recovered during the inverse DANC to VANC process shall be placed into the VBI space of a SDI interface. The placement of VANC packets into the VBI space shall be accomplished by following means: a) If LE is set to ze

39、ro: The placement of the VANC packets into the VBI ancillary space starts one line after the vertical switching point line as defined in RP 168. The insertion of VANC packets continues to the last line of the VBI before the start of the first active video line in a field. The placement of packets sh

40、all conform to SMPTE 291M rules requiring that ancillary packets are contiguous and left justified. A single video line in VBI space can carry multiple VANC packets as long as space is available. NOTE In this case the recovered VANC data packets are not located at the same place on the interface whe

41、re they were originally located. b7 b0 Where: m = 0, 3, 6, B m+2 Bm 8 bits LSB MSB X0 0110 X1 0110B m+2 B m+1 CMB data area b0 - LSB b0 - LSBb0 - LSBb4b7 - MSB b7 - MSB b3b7 - MSBSMPTE 374M-2003 Page 10 of 19 pages b) If LE is set to one: The placement of the packets within the VBI ancillary space s

42、hall be determined by the line number LN indicated in the header of the DANC packet. Recovered VANC data packets are located on the same horizontal line of the interface where they were originally found. The placement of packets shall conform to SMPTE 291M rules requiring that ancillary packets are

43、contiguous and left justified. 6 Extended video line compression and mapping of coded pixels into the CMB data area 6.1 Extended video line pixel and a coded pixel The value of each pixel of the extended video line is described as Pix j, l, k. During a mild compression process, each pixel Pix j, l,

44、k is compressed from an 8-bit value into a 6-bit value by adopting modified PCM. This method adds prediction and adaptive shift quantization to the normal PCM coding. The value of the coded (compressed) pixel (CP) is expressed as CP j, l, k. The symbol j is used for identification of the individual

45、signal components regardless of a non-coded or coded form (j=0 for Y; j=1 for CB; j=2 for CR) of the pixel. The symbol l represents the line number and symbol k identifies the horizontal position in a line. 6.2 Pixel coding (compression) process The compression process is shown below, where all the

46、data operations are twos complement. The described decoding section shows a typical process of a decoder. /* Line number */ line5258 = 20, 21, 22, 263, 282, 283, 284, 525 ; line6259 = 19, 20, 21, 22, 331, 332, 333, 334, 623 ; /* Pixel position */ start_p13 = 0, 0, 0 ; start_p23 = 360, 180, 180 ; end

47、_p3 = 720, 360, 360 ; Encoding if ( 525/60system ) e = 8 else e = 9; if ( 525/60system ) line = line525 else line = line625; color = 3; for ( i = 0; i 3; /* MSB 5 bits */ if ( Pu 0 ) Sf = Pu-1; else Sf = Pu; SI = Pix j,linei,k - Sf; SP = Pr - Sf; /* limitter */ SMPTE 374M-2003 Page 11 of 19 pages if

48、 ( SI 112 ) SI = 112; SR = SP /* LSB 2 bits */ if ( SR = 10b ) RO = 2; else if ( SR = 01b ) RO = 1; else if ( SR = 00b ) RO = 0; else RO = -1; /* 8 bits to 6 bits */ PQ = ( SP + 1 ) 2; IQ = ( SI - RO + 2 ) 2; E = SI - SP; if ( E 1; E3 = E2 - 1; else Q4 = IQ + 3; E2 = E 1; E3 = E2 + 1; Q1 = PQ + E3;

49、Q0 = PQ + E; P4 = ( IQ 4) CP j,linei,k = Q4; Pr = check_limit ( P4 + Sf ); else if ( | E3 | 1) CP j,linei,k = Q1; Pr = check_limit ( P1 + Sf ); else CP j,linei,k = Q0; Pr = check_limit ( P0 + Sf ); Decoding if ( 525/60system ) e = 8 else e = 9; if ( 525/60system ) line = line525 else line = line625; color = 3; for ( i = 0; i 3; /* MSB 5 bits */ if ( Pu 0 )