1、SMPTE STANDARD for Television - SMPTE 362M-2002 Object Data Format for the Exchange of DV-Based Audio, Data and Compressed Video using ATM Common Layer over Asynchronous Transfer Mode (ATM) AALType 1 1 Scope This standard defines the object data format for the exchange of DV-based audio, data, and c
2、ompressed video (which data structure is defined in SMPTE 314M) over the common layer defined in SMPTE 354M. The standard covers the transmission of audio, compressed video, and subcode data with DV-based 25-Mb/s and 5O-Mb/s data structures for 525/60 and 625/50 systems and the transmission of the h
3、eader data of serial data transport interface defined in SMPTE 305.2M. 2 Normative references The following standards contain provisions which, through reference in this text, constitute provisions of this standard. At the time of publication, the editions indicated were valid. All standards are sub
4、ject to revi- sion, and patties to agreements based on this standard are encouraged to investigate the possibii of applying the most recent edition of the standards indicated below. SMPTE 305.2M-2000, Television - Serial Data Transport Interface (SDTI) SMPTE 314M-1999, Television - Data Structure fo
5、r DV-Based Audio, Data and Compressed Video - 25 and 50 Mb/s SMPTE 354M, Television - ATM Common Layer for Transport of Packetized Audio, Video and Data over Asynchronous Transfer Mode using ATM Adapta- tion Layer Type 1 3 Common layer As defined in SMPTE 354M, the common layer is the next higher la
6、yer of theAAL1 and provides transmission Page 1 of 14 pages packets of applications. The common layer consists of two layers: the lower layer is the SYNC layer and the higher layer is the container layer. 3.1 SYNC layer TheSYNC stream block (SSB), which is defined in the SYNC layer, includes contain
7、ers. The SSB has one or more containers, and the transmission order of the SSB is defined in SMPTE 354M. 3.2 Container layer The container layer defines the structure of the con- tainer in the SSB. Each container is intended to hold a video frame (optionally two fields) worth of packet- zed audio, v
8、ideo, and data. 3.2.1 Container format The container has two basic modes: the simple mode and the extended header mode. The simple mode supports four objects of several restricted types. The extended header mode supports up to 16 objects. 3.2.2 Object classification system As described in SMPTE 354M
9、, the container system provides a mechanism for defining the bit and byte packing of audio-video data types. Object classifi- cation is represented in the header by the object class word. This word is composed of two key elements: the type byte and the index byte object. The type byte is a coarse cl
10、assification of the data type, and the index type is the key into the table (which is defined in figure 6 of SMPTE 354M ) which represents data types. Copyright02002 by THESOCIRYOF MOTION PICTURE AND TELNISION ENGINEERS 595 W. Hartcdale Ave., White Plains, NY 10m7 (914) 761-1100 Approved January 31,
11、2002 SMPTE 362M-2002 Stream Stream Time CDS Packet Descriptor Stamp Length 3.2.3 Compressed AV stream Video program is defined as an object class 20h and compressed AV stream (dv-based) is defined as table index 0001h of the video program object in SMPTE 354M. The format of a compressed AV stream is
12、 based on the subcontainer format shown in figure 1. The subcontainer is composed of a 12-byte stream header and multiple/single compressed data stream (CDS) packets. 3.2.3.2.1 Packet type The packet type is 16-bit data and indicates the CDS packet structure. The packet type is defined as follows: -
13、 Bit 31: The existence indication bit of the packet time stamp word. O - indicates that the packet time stamp word exists. 1 - indicates that the packet time stamp word does not exist. - Bit 30: The existence indication bit of the byte count word. 3.2.3.1 Stream header The stream header, which is sh
14、own in figure 1, is composed as follows: Stream descriptor: 32 bits (4 bytes) Stream time stamp: 32 bits (4 bytes) CDS packet length: 32 bits (4 bytes) 3.2.3.2 Stream descriptor The stream descriptor (which is shown in figure 2 ) is composed of the following: Packet type: 16 bits (2 bytes) Stream ty
15、pe: 16 bits (2 bytes) O - indicates that the byte count word exists. 1 - indicates that the byte count word does not exist. - Bit 29: The existence indication bit of the FEC (forward error correction) word. O - indicates that the FEC word exists. 1 - indicates that the FEC word does not exist. - Bit
16、 28: The indication bit that indicates multiple packets or single packet. O - indicates that the subcontainer is composed of 1 - indicates that the subcontainer is composed of a single CDS packet. multiple CDS packets. 31 16 15 O I I Packet Type Stream Type I Figure 2 - Stream descriptor Page 2 of 1
17、4 pages SMPTE 362M-2002 - Bit 27: The indication bit that indicates that the stream time stamp word is valid or invalid. O - indicates that the stream time stamp word is 1 - indicates that the stream time stamp word is valid. invalid. - Bit 26: The indication bit that indicatesfixed-packet size or v
18、ariable-packet size. O - indicates that the CDS packet size is fixed. 1 - indicates that the CDS packet size is variable and the CDS packet length word is invalid. - Bit 25 and Bit 24: Reserved for future purposes (default value is hex 00). - Bit 23 through Bit 16: Indicates the byte length of FEC.
19、3.2.3.2.2 Stream type The stream type is 16-bit data. The definition of the stream typedepends on the compressed data stream. 3.2.3.5.1 CDS packet time stamp The CDS packet time stamp is a 32-bit word and exists when B31 of the packet type is set to O. The CDS packet time stamp is used for recoverin
20、g the relative CDS packet timing from end to end. The reference clock for the packet time stamp is the same frequency as that for the object time stamp in the container header. The synchronization scheme using this word is described later. 3.2.3.5.2 Byte count The byte count is a 32-bit word and exi
21、sts when 830 of the packet type is set to O. The byte count indicates the byte length of the variable CDS packet payload size or the valid byte length in the fixed CDS packet payload. If the packet time stamp word does not exist, the byte count word is located at the head word of the CDS packet. 3.2
22、.3.5.3 CDS packet payload The CDS packet payload format depends on the compressed data stream. 3.2.3.5.4 FEC (forward error correction) 3.2.3.3 Stream time stamp The stream timestamp is composed of a 32-bit packet length word and indicates the point of time at which the compressed AV stream object i
23、s multiplexed into the container. The stream time stamp is used for recovering the relative stream object time from end to end and is valid when B27 of the packet type is set to O. The synchronization scheme using this word is described later. 3.2.3.4 CDS packet length The CDS packet length is compo
24、sed of a 32-bit packet length word and indicates the byte length of the CDS packet multiplexed immediately after the stream header. The CDS packet length word is valid when 827 of the packet type is set to O. 3.2.3.5 CDS packet format The CDS packet format is composed of a 32-bit packet time stamp w
25、ord, a 32-bit byte count word, an N byte CDS packet payload, and M bytes FEC. The CDS packet format is shown in figure 3. The forward error correction is M byte code and exists when 829 of the packet type is set to O. The FEC applies to the first byte through the last byte of the CDS packet. 31 O 1
26、CDS Packet Time Stamp I I Byte Count I FEC Figure 3 - CDS packet format Page 3 of 14 pages SMPTE 3621111-2002 4 50 bytes b 3.2.3.6 Synchronization scheme In a frame/field compression system, such as DV- based audio, data and compressed video, if only the frame/field frequency is recovered accurately
27、, it is not necessary to recover the relative CDS packet timing accurately but to recover the CDS packet within the frame/field in which they were generated by the sender. Therefore, the stream time stamp (STS) or the CDS packet time stamp (CTS) is not always neces- sary. 4 Object for SDTI informaon
28、 and DVhsed audio, data and compressed video 4.1 Defined objects Two kinds of objects, the object for SDTI information and the object for DV-based audio, data, and com- pressed video are defined and shown in figure 4. 4.2 Object format for SDTI information The object for SDTI information (SDTI heade
29、r is de- fined in SMPTE 305.2M) is treated as the ancillary data type object and is shown in figure 5. The line number information in the object for SDTI information shall be the same as the start line number of each video frame in the SDTI channel unit. The object for SDTI information shall corresp
30、ond to data which are BO, B1, B2, B3, B4, B5, B6, and B7 (8 bits) of each word numbered from 3 to 52 in the SDTI header (which is defined in SMPTE 305.2M). The data length is 10 bits; BO through B9. B9 is the most significant bit (MSB). 4.3 Object format for DV-based audio, data and compressed video
31、 DV-based audio, data, and compressed video data are treated as the CDS packet payload of the video program object. I I I I 1 I I 1 I I I I I I I I I Object for SDTI information I Object for DV-based audio, data and Compressed video Figure 4 - Two kinds of defined objects Figure 5 - Object for SDTI
32、information Page 4 of 14 pages SMPTE 362M-2002 - Bit 26 = O: The CDS packet size is fixed. 4.3.1 Stream header 4.3.1.1.3 CDS packet length FEC I 4.3.1.1 Stream descriptor 4.3.1.1.1 Packet type - Bit 31 = 1 : The packet time stamp word does not exist. - Bit 30 = 1 : The byte count word does not exist
33、. - Bit 29 = 0 or 1 According to the error scheme the DV-based stream transport adopts. When the FEC word exists, the packet length shall have the value hex O0 O0 O0 A4 (164 bytes). When the FEC word does not exist, the packet length shall have the value hex O0 O0 O0 AO (160 bytes). 4.3.1.1.4 CDS pa
34、cket format According to the packet type, the CDS packet format of the DV-based CompreSSion stream is show in figure 6. - Bit 28 = 1 : The subcontainer is composed of multiple CDS packets. - Bit 27 =O or 1 : According to the relationship to other objects in the container. CDS Packet Payload m - Bit
35、23 through bit 16: When the 4-byte FEC word exists, the length of FEC shall have the binary value 4.3.1.1.5 CDS packet payload 04. When the FEC word does not exist, the length of FEC shall have the binary value OO. The CDS packet payload format for the DV-based com- pression stream is shown in figur
36、e 7. The DIF block ID 4.3.1.1 2 Stream type (IDO-2) and DIF block data are specified in SMPTE 314M.ACDSpacket includestwoDIF blockdata. Inthe 525/60 system, one compressed AV stream object is composed of 750 CDS packets (1500 DIF blocks) for 25-Mb/s structure or 1500 CDS packets (3000 DIF - Bit 15 t
37、hrough bit 8: Resewed for future purposes (default value is O000 OOOOb). blocks) for 50-Mb/s structure. - Bit 6 through bit 4: These three bits indicate the DIF structure with the following values: 31 Bit6 Bit5 Bit4 O o O O 1 1 1 1 O : Reserved 1 :Reserved O :Reserved 1 : 25-Mb/s structure O : Reser
38、ved 1 : 50-Mb/s structure O :Reserved 1 : Reserved O IDO I ID1 I ID2 DIF block data IDO ID1 ID2 DIF block data - Bit 3 through bit O: Resewed for future purposes (default value is OOOOb). Figure 7 - CDS packet payload format Page 5 of 14 pages CMPTE 3621111-2002 120,000 bytes (525/60) 12 bytes 144,0
39、00 bytes (625/50) 14 b 4.3.1.1.6 FEC (forward error correction) Stream Header DIF Frame data L The CDS packet data are optionally protected by the forward error correction (FEC). The FEC is a 4-byte code and is located immediately after the CDS packet payload when 829 of the packet type is set to O.
40、 The FEC applies to the first byte through the last byte of the CDS packet. A4-byte Reed Solomon error correc- tion shall be used as the method of FEC. 240,000 bytes (52560) 12 bytes 288,000 bytes (625/50) 14 b FEC is not necessary when only the ATM AAL type 1 is used as the transmission means betwe
41、en end to end. However, for example, when the FC-AV trans- mission means with FEC option is used together with the ATM AALtype 1 transmission means between end to end, FEC is used to make the gateway between the FC-AV network and the ATM network simpler. 4.3.2 Object data format When FEC does not ex
42、ist, the object as the com- pressed AV stream for DV-based audio, data, and compressed video data are shown in figures 8 and 9. The object for DV-based audio, data, and compressed video data for 25 Mb/s is as follows: (figure 8): DIFframe data (DIF blocks): 1,500 DIF blocks (120,000 bytes) (525/60)
43、1,800 DIF blocks (144,000 bytes) (625150) The object for DV-based audio, data, and compressed video data for 50 Mb/s is as follows (figure 9): DIFframe data (DIF blocks): 3,000 DIF blocks (240,000 bytes) (525160) 3,600 DIF blocks (288,000 bytes) (625/50) 4.3.3 DIF frame data The DIF frame data shall
44、 consist of 1,500 DIF blocks (525/60) or 1,800 DIF blocks (625/50), and each DIF block consists of 80 bytes. The DIF block is specified in SMPTE 314M. The transmission order within one compressed video frame for 25 Mb/s and 50 Mb/s structures is shown in figures 10 and 11. Stream Header DIF Frame da
45、ta I Figure 9 - Example of object for DV-based audio, data and compressed video for 50 Mbls Page 6 of 14 pages SMPTE 362M-2002 Structure of a DIF block ID DIF Frame Data DIF Sequence Data i I DF Frame Data (Data in one video frame) I structure Of a Sequence Audio that is, F800h, in the following cas
46、es: When the packet time stamp word does not exist, bit 31 is set to 1. When the byte count word does not exist, bit 30 is set to i. When the FEC word does not exist, bit 29 is set to 1. When the subcontainer is composed of multiple CDS packets, bit 28 is set to 1. When the stream time stamp word is
47、 invalid, bit 27 is set to 1. When the CDS packet size is fixed, bit 26 is set to O. Bit 25 and bit 24 are oob, and bit 23 through bit 16 are all set to O. Page 9 of 14 pages SMPTE 362M-2002 00 h Object 3 Object type defined 21 The value of the stream type in the stream descriptor be- comes O000 O00
48、0 O01 l OOOOb (0030h) in the following cases: When bit 15 through bit 8 are set to default value, When bit 7 is set to default value, ob. When 25-Mb/s DV-based 525/60, bit 6 through bit 4 are set to 011b. When bit 3 through bit O are set to default value OOOOb. Therefore, the value of the stream des
49、criptor is F800 0030h. The stream time stamp is set to O000 ooooh because it does not exist. The CDS packet length, when ECC does not exist, becomes the same as the length of two DIF blocks, 160 bytes (O000 O0 Ah by tes). 0000 oooob. OOh OOh OOh The value of the container header is shown in figure A.2. The value of the sequence number is defined by the user. The value of the clip ID is set to O000 ooooh as the default value (null). The value of the container time stamp is set to O000 O000 O000 ooooh as the default value (null). The value of the video frame rate is set to 051, as is 29.
