EP0732032A1 - Processeur de donnees destine a assembler des paquets de transport de donnees - Google Patents

Processeur de donnees destine a assembler des paquets de transport de donnees

Info

Publication number
EP0732032A1
EP0732032A1 EP94905919A EP94905919A EP0732032A1 EP 0732032 A1 EP0732032 A1 EP 0732032A1 EP 94905919 A EP94905919 A EP 94905919A EP 94905919 A EP94905919 A EP 94905919A EP 0732032 A1 EP0732032 A1 EP 0732032A1
Authority
EP
European Patent Office
Prior art keywords
data
packet
word
words
header
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
EP94905919A
Other languages
German (de)
English (en)
Other versions
EP0732032A4 (fr
Inventor
Alfonse Anthony Acampora
Richard Michael Bunting
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
General Electric Co
Original Assignee
General Electric Co
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Family has litigation
First worldwide family litigation filed litigation Critical https://patents.darts-ip.com/?family=22237242&utm_source=google_patent&utm_medium=platform_link&utm_campaign=public_patent_search&patent=EP0732032(A1) "Global patent litigation dataset” by Darts-ip is licensed under a Creative Commons Attribution 4.0 International License.
Application filed by General Electric Co filed Critical General Electric Co
Publication of EP0732032A1 publication Critical patent/EP0732032A1/fr
Publication of EP0732032A4 publication Critical patent/EP0732032A4/fr
Withdrawn legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
    • H04N19/61—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding in combination with predictive coding
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/23—Processing of content or additional data; Elementary server operations; Server middleware
    • H04N21/234—Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/70—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by syntax aspects related to video coding, e.g. related to compression standards

Definitions

  • This invention is related to the field of digital signal processing, and more particularly to method and apparatus for processing a datastream including transport packets conveying coded MPEG-like data in a high definition television system, for example.
  • the codeword datastream is passed to a transport processor which packs the codeword data into transport cells each including header and packed data payload sections, and which provides output high priority and low priority datastreams.
  • a major function of the transport processor is to pack variable length codeword data, issued by a preceding priority processor, into packed data words.
  • An accumulation of packed words, called a data packet is prefaced with a transport header.
  • the transport packet format enhances resynchronization and signal recovery at a receiver, eg., after a disruption due to a transmission channel disturbance, by providing header data from which a receiver can determine re-entry points into the data stream on the occurrence of a loss or corruption of transmitted data.
  • Synchronization of data at an MPEG decoder also is facilitated by a Group of Pictures (GOP) beginning at a packet boundary.
  • GOP is a series of one or more pictures, or frames, intended to assist random access into a coded video bitstream sequence.
  • Resynchronization is also facilitated by responding to an intracoded I-frame picture start codeword, and by placing the picture start codeword at a packet boundary, e.g., in a system according to the MPEG standard.
  • variable length codewords e.g., MPEG codewords
  • incomplete data packets having less than a prescribed number of words are packed with no-operation "zero words" to produce a fixed length packet, and to define an inter-packet boundary at which a prescribed codeword may appear.
  • a special codeword is inserted into an MPEG datastream to signify the beginning of a Group of Pictures (GOP).
  • the PAF immediately precedes a Picture Start codeword for an intracoded "I" frame, which initiates a GOP.
  • the PAF indicates the imminent appearance of a Picture Start codeword, and encompasses one clock cycle during which certain "housekeeping" functions are performed before the Picture Start codeword appears at the beginning of the next packet. These housekeeping functions include, for example, resetting accumulators, checking header status, and generating a Last Word Indicator for the data packet under construction when the PAF appeared.
  • a data packet under construction is terminated, since a
  • GOP is intended to begin at a packet boundary. Such termination may result in an abbreviated packet containing less than a prescribed number of packed words.
  • the abbreviated data packet is filled with null (zeroed bits) words to make up a complete data packet with a prescribed number of words, and to define an inter- packet boundary at which a picture start codeword may appear.
  • Thirty 32-bit fixed length words form a data packet, which is prefaced by a 32-bit header.
  • FIGURE 1 is a block diagram of a portion of a video signal encoder including data word controller, data packer and data/header combiner apparatus according to the present invention.
  • FIGURES 2A, 2B and 2C show details of a word controller and data packer of FIGURE 1.
  • FIGURE 3 is a truth table associated with the operation of a word status controller shown in FIGURE 2A.
  • FIGURE 4 shows details of a packed data assembler network.
  • FIGURES 5 - 16 depict examples of last word generation.
  • FIGURE 17 shows details of a data and header combiner unit shown in FIGURE 1.
  • FIGURE 18 is a state diagram associated with the operation of a state controller shown in FIGURE 17.
  • FIGURE 19 is a block diagram of an HDTV encoding system including apparatus according to the present invention.
  • FIGURES 20A and 20B are pictorial representations of sequences of image fields/frames of encoded video signals.
  • FIGURE 21 is a pictorial representation of datablock generation developed by encoding/compression apparatus in the system of FIGURE 19.
  • FIGURE 22 is a generalized pictorial representation of the data format provided by the encoding/compression apparatus in the system of FIGURE 19.
  • FIGURE 1 is a block diagram of a data packer 12 and a packed data word controller 10 of a transport processor.
  • a major function of a transport processor is to pack variable length codeword data into fixed length (e.g., 32 bit) data words. Thirty accumulated data words constitute a data packet, which is eventually prefaced by a transport header.
  • Such a transport processor may be employed in a system for processing an MPEG-like compressed video signal, as will be discussed subsequently with regard to FIGURE 19. Additional aspects of
  • Controller 10 monitors the accumulation of Length data words relative to a Packet Alignment Flag (PAF) to ascertain the completion of 32-bit data words, assembled from a stream of variable length codewords, and the completion of 960-bit long data packets.
  • the Length data is a six bit parallel word coincident with a variable length codeword whose length it defines.
  • the binary value of the Length word indicates the number of bits in the coincident variable length codeword that actually represent the MPEG codeword to be transported.
  • Each variable length codeword appears on a 32 bit wide bus, with a variable number of valid bits (1 to 32) representing the MPEG codes.
  • the PAF is generated by input processor 14 so that the PAF appears one codeword immediately before an MPEG "I" (intracoded) frame Picture Start codeword at the beginning of a Group of Pictures.
  • the PAF may be generated by sensing the presence of the I frame Picture Start codeword by means of a digital comparator.
  • Unit 14 also includes a signal delay network for processing the Picture Start codeword and the PAF so that the PAF occurs in the codeword clock cycle immediately before the I frame Picture Start codeword.
  • the delay network also assures that the output signals applied to units 10 and 12 exhibit proper time synchronism.
  • Word Addresses are passed to data packer 12, which receives the variable length codewords for packing, to insure proper concatenation of the input variable length codewords.
  • Word Control signals are also presented to packer 12 to account for short words, to mark the last word in a packet, and to insure proper alignment of a sequence of thirty packed data words with an associated transport header.
  • Controller 10 tracks the completion of packets by accumulating the binary values of the Length words. Each value represents the number of valid bits in the associated codeword. A packet is complete when 960 bits have been accumulated. The starting point, or initialization, of this count is provided by the appearance of the PAF, which causes internal accumulators within controller 10 to be reset.
  • Packer 12 receives the variable length codewords (VLC) via a 32 bit parallel data bus.
  • VLC variable length codewords
  • the valid bits are packed into 32 bit words under the supervision of signals from controller 10. Concatenation is arranged to accommodate the eventual MPEG bit serial transmission order.
  • the packed data from unit 12 is sent at a variable word rate to an input FIFO data buffer 16 of data and header combiner 15.
  • Combiner 15 also receives a Data Write
  • Enable signal from packer 12 which enables valid data to be written to FIFO data buffer 16 in combiner 15.
  • a data packet is fully formed when thirty such words have been transmitted, unless a PAF forces a short packet.
  • a Last Word indicator produced by packer 12 marks the thirtieth word in a normal packet in this example, or the last word in a packet shortened by the appearance of the PAF.
  • Packed data words are transmitted to data/header combiner 15 whenever the packed data words are available.
  • transport headers are transmitted from a header generator 18 to an input FIFO header buffer 17 of combiner 15 whenever the headers are available.
  • Information utilized by header generator 18 to form headers is obtained from input processor 14 and word controller 10.
  • a Header Write Enable signal indicates that a header is available, and enables the headers to be written to FIFO
  • Combiner 15 prefaces each packed data payload with an appropriate header, and forwards the resulting transport packet, or block, to an output rate buffer as will be seen in FIGURE 19.
  • Combiner 15 also provides an output Data Ready signal to signify that a packed data word or transport header is ready to be forwarded.
  • a Header Indicator signal denotes that clock cycle in which the header is being forwarded. This signal acts as a transport packet boundary marker so that subsequent operations, such as Forward Error Correction (FEC), can be suitably applied to the transport cell.
  • FEC Forward Error Correction
  • Each header contains information related to the data in the data packet with which the header is associated.
  • the header information aids data assembly and synchronization at a receiver, and includes information such as service type (e.g., audio, video, data), frame type, frame number and slice number, for example.
  • service type e.g., audio, video, data
  • frame type e.g., frame number
  • slice number e.g., slice number
  • a header of this type and its processing are described in the context of an HDTV digital signal processing system employing MPEG signal coding in U.S. patent 5,168,356 - Acampora et al.
  • a data packet may contain less than thirty packed data words, from 1 to 29 words in this example.
  • the PAF provided by input processor 14, appears immediately before the Picture Start codeword of an intracoded I-frame at the beginning of a GOP, as will be discussed in connection with FIGURES 20-22.
  • the Picture Start codeword for an intracoded frame always starts a new packet, and the immediately preceding PAF signifies the end of a data packet and the beginning of a new packet. Packet alignment of this Picture Start codeword is instrumental for rapid acquisition of the data stream at a receiver.
  • PAF occurs during the formation of a fixed length word, an abbreviated data packet is formed.
  • the remaining bits in the packed word under construction are filled with a number of "zeroed bits" (from 1 to 31) in data packer 12.
  • the remaining words in the data packet will likewise be filled with "zeroed words” (from 1 to 29) in combiner 15, so that the transport packet size is maintained.
  • the need for such "zero word fill” is indicated by the appearance of a Last Word indicator before thirty data words have been transmitted to combiner 15.
  • the proper identification of the Last Word in a data packet is important.
  • the Last Word insures proper registration of a constructed packet with its associated transport header.
  • the Last Word also accounts for the filled packet at the MPEG Group Of
  • Last Word Pictures boundaries (i.e., intracoded frame), which is vital for re ⁇ synchronization at a television receiver/decoder such as after channel changing. Determination of the Last Word is not a trivial matter, and relies on specific knowledge of packet status such as when a packet is completed, and if completed, whether or not there is data segmentation into the next packet. There are conditions when the Last Word is the word formed in the current clock interval or the word to be formed in the next clock interval. Some specific examples of Last Word formation are as follows. In the absence of a PAF when a packet is complete the last word (the thirtieth word in this example) is the last word and is so marked by a Last Word Indicator. This is an example of a "true" Last Word.
  • a PAF may occur as a packet is completed with no bits segmenting into the next packet, i.e., the word ends exactly on a packet boundary.
  • the last word of a completed packet is marked as a Last Word because it is so in fact. This is another example of a "true" Last Word.
  • a PAF also may occur as a packet is completed with some bits segmenting into the first word of the next packet. In this case two successive Last Words are formed and marked accordingly.
  • the last word of the completed packet is marked as Last (a "true” Last Word), and the first word of the next packet is marked as Last because the PAF forces the packet abbreviation.
  • Last Word appearing before 30 words have been sent results in "zero word filling" to complete the packet.
  • Other examples of a Last Word follow.
  • a PAF may occur at a time when an . incomplete packet is under construction. If an internal word is completed with some bits segmenting into the next word, then the partial word becomes the Last Word. A particularly troublesome situation is presented when an incomplete packet is under construction and an internal word is completed with no bits segmenting into the next word.
  • the internal word may be conveyed to the data/header combiner before subsequent data (i.e., appearance of the PAF) indicates that this word was a last word, too late for it to be properly marked as such.
  • a zero word called a "pseudo" last word is generated and marked as the last word.
  • a pseudo last word is formed entirely of zero bits, in contrast to, for example, a segmented (incomplete) last word which is partially padded with zero bits.
  • FIGURE 2A shows details of controller 10 of FIGURE 1.
  • the controller includes an accumulator 20 in a feedback arrangement with a modulo 960 circuit 22.
  • a buffer register 23 is included in the feedback loop to hold the newly accumulated value at the end of each Length input cycle.
  • Input PAF and Length words are conveyed via an input register 24 to modulo unit 22 and accumulator 20, respectively.
  • the values of the Length words are successively accumulated by unit 20, and the feedback combination of accumulator 20 and modulo 960 unit 22 sets the length of a packet at 960 bits.
  • the accumulator output from register 23 represents the bit position within a packet and is conveyed to a packet status controller 25.
  • Packet status controller 25 also receives a PAF from input buffer register 24, and issues output signals required to create the writing instructions in word status controller 26.
  • a Packet Complete output signal is issued to word status controller 26 when the accumulator bit count is equal to or greater than 960.
  • An output Remnant signal is issued by controller 25 when the accumulator bit count is not on a word boundary (i.e., a bit count not equal to an integer multiple of 32).
  • a True Zero output signal is issued when the accumulator bit count is zero. This signal is significant in determining the correct formation of the last word only when a PAF is present.
  • a logic circuit for generating these signals is shown in FIGURE 2B, discussed below.
  • Accumulator 20 idles, and holds the last bit count value, when zero value Length words are received indicating the presence of zero length null codewords, i.e., no-operation (NO-OP) codewords.
  • NO-OP no-operation
  • One exception to this rule is that a PAF will always force the accumulator value to zero regardless of the bit count.
  • Another exception occurs when a packet is completed exactly on a packet boundary (i.e., the accumulator count is 960). In the next clock cycle, the accumulator count will be corrected to the binary value of the next Length word via modulo 960 unit 22. A packet is complete when the accumulator count is equal to or greater than 960.
  • the 10 bit accumulator outputs representing the accumulated lengths, are designated as 10 to 19.
  • a packet is completed if the accumulated lengths of the packed codewords is equal to or greater than 960. This condition is indicated when the four MSB bits of the accumulator, 16 to 19, are in a logic 1 state, as applied to an AND gate 30. True Zero is indicated when all ten accumulator bits are in a logic state 0, as applied to an OR gate 31. A "no remnant" condition is indicated when the five LSB accumulator bits, 10 to 14, are in a logic state 0, as applied to an OR gate 32.
  • the data packer Word Address is formed in response to the six LSB accumulator bits, 10 to 15. AND gate array 34 forces the Word Address to a logic state 0 (Word Address reset) when the Packet Alignment Flag (PAF) is encountered.
  • PAF Packet Alignment Flag
  • FIGURE 2C shows details of data packer 12 of FIGURE 1.
  • Variable length codewords are presented to a data shifter 35.
  • the shifter can be a barrel shifter such as Texas Instrument type 74AS8838.
  • an LSB subset of the length accumulator output emanates from the packet status controller 25 (see FIGURES 2A and 2B) and is presented to the data shifter as the Word Address.
  • the packed word is transferred to a holding register 36.
  • the availability of the packed data word is flagged by a Word Ready signal issued by register 36, so that the word can be transferred to a data assembly network 37.
  • Data assembler 37 uses control signals WENl, WEN2 and WZERO from packed word controller 10 (FIGURE 1) to issue the packed data word with its Data Write Enable and Last Word flag to a FIFO buffer contained within data and header combiner 15 of FIGURE 1.
  • the insertion of the next transport headers into the combined data stream follows the transmission of the last word in a packet such that the header for the next packet is inserted after the last word of the current packet.
  • the Accumulator Output (FIGURE 2A) is used by a Header Controller to indicate the position of certain codewords in the packet so that these positions can be described in entry point fields within the header.
  • the creation of Last Word indicators and flags enabling data words to be written to a FIFO buffer is facilitated by logic arrays associated with word status controller 26 and data assembly network 37.
  • Table 1 below shows the action states for creating the last word in response to logic array inputs PAF (Packet Alignment Flag), PC (Packet Complete), TZ (True Zero) and REM (Remnant, the word segmentation indicator).
  • PAF Packet Alignment Flag
  • PC Packet Complete
  • TZ Truste Zero
  • REM Remnant, the word segmentation indicator.
  • the formation of the output signals from data assembly network 37 is facilitated by output signals from controller 26, supplied via buffer register 28. These signals include a write enable signal W EN1 which signifies a last word appearing in the next clock cycle, a write enable signal W EN2 signifying a last word appearing in the current clock cycle, and a write zero signal W ZERO which creates a pseudo Last Word. This pseudo Last Word occurs when a PAF coincides with the packet formation residing at an internal codeword boundary of an incomplete packet. TABLE 1.
  • the data assembler network of FIGURE 4 comprises an output network of data packer 12 in FIGURE 1.
  • the data assembler comprises AND logic gates 42 and 44, an OR logic gate 46, and D-type flip-flops 43 and 45 arranged as shown.
  • Packed 32 bit wide Data Words are passed to a data FTFO via AND logic gates 42, and a Word Ready signal from preceding packer circuits is passed via AND logic gate 44, and becomes the Data Write Enable signal for data FIFO 16 in FIGURE 1.
  • Data write control signals W EN1, W EN2 and W ZERO from the packed word status controller (FIGURE 2A) are applied to flip-flops 43 and 45 and logic gate 46 as shown.
  • W EN2 indicates a Last Word flag associated with a current word
  • W EN1 indicates a Last Word flag associated with a word in the next clock cycle.
  • the W ZERO control indicates the formation of a pseudo Last Word (case 4 in Table 1) which is flagged as a last word by W EN1. In this case an all-zero word, referred to as the pseudo last word, is inserted into the Packed Data word stream, and is written to data FIFO 16.
  • the Word Ready input signal to assembler gate 44 is provided by a holding register 36 (FIGURE 2C) to signify the availability of a packed 32-bit word.
  • FIGURES 5-16 Some of these examples show the effect of zero length NO-OP words coincident with, following and preceding the PAF.
  • FIGURES 5 and 6 illustrate variations of Case 5 in Table 1.
  • a packet is completed with segmentation into the next packet (i.e., accumulator bit value greater than 960).
  • FIGURE 6 a packet is completed exactly on a packet boundary (i.e., accumulator value equal to 960), with no segmentation or remnant into the next packet.
  • a Last Word flag occurs at the Packet Complete time. This occurrence is independent of True Zero and Remnant indications because a PAF is not present. Otherwise, True Zero and Remnant indications would have to be considered in the presence of a PAF.
  • FIGURES 7 and 8 illustrate Case 2 of Table 1.
  • a PAF occurs immediately after a packet is completed, with no segmentation, followed by a 32-bit Picture Start codeword.
  • FIGURE 8 is similar except that three intervening zero length no- operation (NO-OP) code words precede the Picture Start codeword.
  • NO-OP zero length no- operation
  • FIGURE 7 shows the more typical case where a 32 bit length Picture Start codeword follows immediately.
  • FIGURE 8 indicates that intervening NO-OP words are tolerated.
  • FIGURE 9 pertains to Case 6a of Table 1, where a PAF does not coincide with a Packet Complete signal. The PAF occurs after a packet is completed, with no segmentation and after NO-OP words, followed by a Picture Start codeword. In this case a Last Word indication is associated with the Packet Complete signal, but there is no Last Word indication associated with the PAF because a True Zero indication is involved in this case, due to the accumulator idling at zero.
  • FIGURES 10 and 11 illustrate Case 1 of Table 1.
  • a PAF occurs immediately after a packet is completed, with segmentation, followed by a Picture Start codeword.
  • FIGURE 11 is similar to FIGURE 10 except that intervening NO-OP words precede the Picture Start codeword.
  • Two Last Word indicators are required because of the segmentation remnant.
  • One Last Word indicator occurs at the Packet Complete interval, and the other occurs one clock interval after the PAF because of the segmentation.
  • FIGURES 12, 13 and 14 illustrate Case 3 of Table 1.
  • a PAF occurs sometime during the packet formation, but not on a word boundary (i.e., there is segmentation into the next word), and not coincident with a Packet Complete indication.
  • the Last Word signal then normally occurs at the next clock interval after the PAF as a result of the partially started word (due to segmentation).
  • the PAF occurs after a packet is completed, with segmentation and after some NO-OP words, followed by a Picture Start codeword.
  • FIGURE 13 a PAF occurs immediately as a word is completed, with segmentation, followed by a Picture Start codeword.
  • a PAF occurs after a word is completed, and after several codewords have caused segmentation.
  • FIGURES 15 and 16 illustrate Case 4 of Table 1, concerning the need to create a particular type of last word, the pseudo last word.
  • This case involves a PAF occurring immediately after (FIGURE 15) or some time after (FIGURE 16) a word is completed without segmentation, i.e., directly on a word boundary multiple of 32.
  • the premise in this case is that the completed word has been issued in advance of the knowledge (provided by the subsequent PAF) that it was the last word.
  • An all-zero pseudo last word is formed and issued. This is tolerable because MPEG allows any number of leading zeros before a start codeword, and the Picture Start codeword is guaranteed to be next by the occurrence of the PAF.
  • the balance of the packet would be filled with zeroed bit (null) words by the data/header combiner. Since one zero word has been issued and pseudo-marked as last, the combiner will issue one less word in this case.
  • a PAF occurs immediately as a word is completed (without segmentation), followed by a Picture Start codeword.
  • a word is completed (without segmentation) and followed by intervening NO-OP words. Afterwards a PAF occurs, followed by a Picture Start codeword.
  • FIGURE 17 shows additional details of Data/Header combiner 15 (FIGURE 1). Header components are written into a header FIFO 70, in response to a Header Write Enable signal, whenever the headers are produced by header generator 18. Similarly, packed data words are written into a data FIFO 72, in response to a Data Write Enable signal, whenever such words are produced by data packer 12. A Last Word indicator, produced in the data packing process, accompanies the last word in a packet whether or not such word is the thirtieth word. Header and data outputs of units 70 and 72 are multiplexed onto a common bus by means of multiplexer 76, and applied to an output register 78.
  • Register 78 provides a Data Ready signal, Packet Data and Headers, and a Header Indicator to rate buffers 713 and 714 as shown in FIGURE 19.
  • Multiplexer 76 is capable of issuing zero words on command, in response to an Issue Zero signal from FIFO state controller 74.
  • controller 74 which is a state machine. After a power-on or similar restart, controller 74 waits for a header to be available. An available header is presented to the output bus of multiplexer 76, along with a Data Ready indicator and a Header Indicator. Controller 74 will then serve data FIFO 72, extracting data whenever it is available, until a Last Word indicator appears. Each data transmitted is accompanied by a Data Ready indicator which is conveyed to output register 78. If 30 data words have been serviced after the Last Word indicator appears, controller 74 will then re-examine header FIFO 70 for available information.
  • controller 74 instructs multiplexer 76, by means of the Issue Zero command, to issue zero words for the balance of the packet. All such zero words are accompanied by a Data Ready indicator. Whenever there is no header or data to transmit, controller 74 instructs multiplexer 76 to issue zero words without any Data Ready indicator for the duration that data is unavailable.
  • FEC FEC coding and data interleaving subsequent to the rate buffers.
  • the FEC and interleaving processes require that the headers be conveyed first, i.e., a header will commence transmission to the rate buffer before the data packet described by the header.
  • the Empty Flag signals conveyed from header FIFO 70 and data FIFO 72 respectively indicate that no headers or data words are present, whereby state machine controller 74 idles. This condition is illustrated in the FIGURE 18 diagram for the "No Header" and "No Word" conditions of State 0 and State 1.
  • the Header/Data Select signal from controller 74 instructs multiplexer 76 to switch either the Header Output from unit 70 or the Data Output from unit 72 to the signal bus to the input of output register 78, when an associated Read Enable signal is conveyed to Header FIFO 70 or to Data FIFO 72, respectively.
  • Output buffer 78 to which zero words are added to an incomplete, abbreviated packet to produce a desired 30- word data packet, is considerably larger than preceding header buffer 70 and data buffer 72. These buffers advantageously continue to receive and process data without interruption. Such uninterrupted operation greatly simplifies timing and synchronizing functions, e.g., by eliminating clock stop/start synchronizing difficulties.
  • a start codeword is a specific re-synchronizing point in an MPEG- compatible data stream.
  • the start codeword appears at a packet boundary, the development of which is advantageously facilitated in the disclosed system by the use of zeroed-bit null words to complete a truncated data packet and define the packet boundary.
  • the MPEG standard tolerates any number of zeroed words before a start codeword, and a receiver/decoder ignores the zeroed-bit null words.
  • output buffer 78 is large and time resilient, and is therefore a convenient vehicle for performing the null word packing operation. In this regard it is noted that very little time is available (eg., one clock cycle) for null word packing between the appearance of a Packet Alignment Flag and a Picture Start codeword at a packet boundary.
  • FIGURE 19 illustrates an exemplary HDTV encoding system which may employ apparatus according to the invention in the transport processor section.
  • FIGURE 19 shows the system processing a single video input signal, but it is to be understood that luminance and chrominance components are processed separately, and that luminance motion vectors are used for generating compressed chrominance components.
  • the compressed luminance and chrominance components are interleaved to form macroblocks before codeword priority parsing. Additional information concerning the system of FIGURE 19 is found in U.S. patent 5,168,356-Acampora et al.
  • a sequence of image fields/frames shown in FIGURE 20A is applied to circuitry 705 which reorders the fields/frames according to FIGURE 20B.
  • the reordered sequence is applied to a compressor 710 which generates a compressed sequence of frames that are coded according to an MPEG-like format.
  • This format is hierarchical and is illustrated in abbreviated form in FIGURE 22.
  • the MPEG hierarchical format includes a plurality of layers each with respective header information. Nominally each header includes a start code, data related to the respective layer and provision for adding header extensions.
  • an I coded frame is one which is intraframe compressed such that only I frame compressed data is required to reproduce an image.
  • P coded frames are coded according to a forward motion compensated predictive method, where the P frame coded data is generated from the current frame and an I or P frame occurring before the current frame.
  • B coded frames are coded according to a bidirectionally motion compensated predictive method. The B coded frame data is generated from the current frame and from I and P frames occurring both before and after the current frame.
  • the coded output signal of the present system is segmented into groups of fields/frames, or groups of pictures (GOP) illustrated by the row of boxes L2 (FIGURE 22).
  • Each GOP (L2) includes a header followed by segments of picture data.
  • the GOP header includes data related to the horizontal and vertical picture size, the aspect ratio, the field/frame rate, the bit rate, etc.
  • the picture data (L3) corresponding to respective picture fields/frames includes a picture header followed by slice data (L4).
  • the picture header includes a field/frame number and a picture code type.
  • Each slice (L4) includes a slice header followed by a plurality of blocks of data MBi.
  • the slice header includes a group number and a quantization parameter.
  • Each block MBi represents a macroblock and includes a header followed by motion vectors and coded coefficients.
  • the MBi headers include a macroblock address, a macroblock type and a quantization parameter.
  • the coded coefficients are illustrated in layer L6.
  • Each macroblock includes 6 blocks, including four luminance blocks, one U chrominance block and one V chrominance block (see FIGURE 21).
  • a block represents a matrix of pixels, e.g., 8 x 8, over which a discrete cosine transform (DCT) is performed.
  • the four luminance blocks are a 2 x 2 matrix of contiguous luminance blocks representing, e.g., a 16 x 16 pixel matrix.
  • the chrominance (U and V) blocks represent the same total area as the four luminance blocks. That is, before compression the chrominance signal is subsampled by a factor of two horizontally and vertically relative to luminance.
  • a slice of data corresponds to data representing a rectangular portion of an image corresponding to an area represented by a contiguous group of macroblocks.
  • a frame may include a raster scan of 360 slices, 60 slices vertically by 6 slices horizontally.
  • the block coefficients are provided one block at a time by the DCT.
  • the DC coefficient occurs first, followed by respective DCT AC coefficients in the order of their relative importance.
  • An end-of-block code EOB is appended at the end of each successively occurring block of data.
  • data from compressor 710 is processed by a prioritizer 711 before being provided to a transport processor 712 which segments the data into high priority (HP) and standard priority (SP) components. These components are coupled via rate buffers 713 and 714 to respective forward error coding units 715 and 716.
  • the rate buffers temporarily store the packed data and headers for subsequent extraction by the FEC error coding networks.
  • a rate controller 718 cooperates with buffers 713, 714 to adjust the average data rate provided by compressor 710. Thereafter the signals are coupled to a transmission modem 717 where the HP and SP data quadrature amplitude modulate respective carriers within a standard 6 MHz NTSC television channel.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)
  • Color Television Systems (AREA)
  • Television Systems (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
  • Editing Of Facsimile Originals (AREA)

Abstract

Dans un système de traitement de signaux de télévision numériques, un mot codé spécial, code d'alignement de paquets (PAF) (14), est inséré dans un train binaire de mots codés en MPEG, afin d'indiquer la présence d'un groupe d'images (GOP). Le PAF précède immédiatement un mot codé de départ d'image pour une imgae 'I', ce qui déclenche un GOP. Un paquet de données (12) en cours d'élaboration au moment de l'apparition d'un PAF (14), est terminé, étant donné qu'un GOP est censé commencer au niveau d'une limite de paquet. Cet arrêt peut produire un paquet abrégé, dont le nombre de mots codés est inférieur à celui qui est nécessaire pour constituer un paquet de données. Le dernier mot de chaque paquet est désigné en tant que tel, afin de faciliter la combinaison consécutive de paquets de données (12) avec des en-têtes respectives (18). Un paquet de données incomplet est rempli par des mots nuls (bits mis à zéro), afin de créer un paquet de données complet contenant un nombre déterminé de mots.
EP94905919A 1993-11-30 1993-11-30 Processeur de donnees destine a assembler des paquets de transport de donnees Withdrawn EP0732032A4 (fr)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/US1993/011583 WO1995015650A1 (fr) 1993-11-30 1993-11-30 Processeur de donnees destine a assembler des paquets de transport de donnees

Publications (2)

Publication Number Publication Date
EP0732032A1 true EP0732032A1 (fr) 1996-09-18
EP0732032A4 EP0732032A4 (fr) 1998-12-30

Family

ID=22237242

Family Applications (1)

Application Number Title Priority Date Filing Date
EP94905919A Withdrawn EP0732032A4 (fr) 1993-11-30 1993-11-30 Processeur de donnees destine a assembler des paquets de transport de donnees

Country Status (8)

Country Link
EP (1) EP0732032A4 (fr)
JP (1) JP4357594B2 (fr)
KR (1) KR100327684B1 (fr)
CN (2) CN1099808C (fr)
AU (1) AU5983794A (fr)
MY (2) MY166757A (fr)
TW (1) TW243578B (fr)
WO (1) WO1995015650A1 (fr)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100821835B1 (ko) * 2001-06-30 2008-04-11 주식회사 케이티 멀티미디어 서버 시스템에서의 실시간 영상정보 정렬 방법
SE0303085D0 (sv) * 2003-11-18 2003-11-18 Scalado Ab Method for creating a compressed digital image representation and image representation format
JP4754648B2 (ja) * 2010-03-16 2011-08-24 ゼネラル・エレクトリック・カンパニイ データストリームを処理する方法
JP4648489B2 (ja) * 2010-03-17 2011-03-09 ゼネラル・エレクトリック・カンパニイ データストリームを処理する方法
CN108966000B (zh) * 2018-07-17 2021-01-29 北京世纪好未来教育科技有限公司 播放方法及其装置、介质、终端

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0441168A2 (fr) * 1990-02-06 1991-08-14 ALCATEL ITALIA Società per Azioni Système, structure de paquet et dispositif pour le traitement d'information de sortie d'un codeur de signal
WO1991013516A2 (fr) * 1990-02-26 1991-09-05 Alcatel Bell-Sdt S.A. Procede de codage de signaux d'image

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB8929152D0 (en) * 1989-12-22 1990-02-28 Gen Electric A digital augmentation system for actv-ii
US5122875A (en) * 1991-02-27 1992-06-16 General Electric Company An HDTV compression system
US5289276A (en) * 1992-06-19 1994-02-22 General Electric Company Method and apparatus for conveying compressed video data over a noisy communication channel

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0441168A2 (fr) * 1990-02-06 1991-08-14 ALCATEL ITALIA Società per Azioni Système, structure de paquet et dispositif pour le traitement d'information de sortie d'un codeur de signal
WO1991013516A2 (fr) * 1990-02-26 1991-09-05 Alcatel Bell-Sdt S.A. Procede de codage de signaux d'image

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
JOSEPH K ET AL: "MPEG++: A ROBUST COMPRESSION AND TRANSPORT SYSTEM FOR DIGITAL HDTV" SIGNAL PROCESSING. IMAGE COMMUNICATION, vol. 4, no. 4 / 05, 1 August 1992, pages 307-323, XP000293751 *
MASAHIRO WADA ET AL: "STRUCTURE PACKING OF CODED VIDEO SIGNALS FOR ATM NETWORKS" ELECTRONICS & COMMUNICATIONS IN JAPAN, PART I - COMMUNICATIONS, vol. 75, no. 5, 1 May 1992, pages 23-33, XP000311465 *
See also references of WO9515650A1 *

Also Published As

Publication number Publication date
TW243578B (en) 1995-03-21
MY118734A (en) 2005-01-31
EP0732032A4 (fr) 1998-12-30
CN1111876A (zh) 1995-11-15
JPH09507975A (ja) 1997-08-12
KR960706751A (ko) 1996-12-09
CN1434634A (zh) 2003-08-06
CN1196331C (zh) 2005-04-06
AU5983794A (en) 1995-06-19
KR100327684B1 (ko) 2002-07-27
JP4357594B2 (ja) 2009-11-04
WO1995015650A1 (fr) 1995-06-08
CN1099808C (zh) 2003-01-22
MY166757A (en) 2018-07-20

Similar Documents

Publication Publication Date Title
US5796743A (en) Data word indicator in a system for assembling transport data packets
US5287178A (en) Reset control network for a video signal encoder
JP3708974B2 (ja) コード化ビデオ信号の処理装置
EP0844792B1 (fr) Procédé d'agencement de données video comprimées pour la transmission sur une voie de communication bruyante
US5231486A (en) Data separation processing in a dual channel digital high definition television system
US5784110A (en) Data processor for assembling transport data packets
US5767912A (en) Datastream packet alignment indicator in a system for assembling transport data packets
WO1995015650A1 (fr) Processeur de donnees destine a assembler des paquets de transport de donnees
EP0732029B1 (fr) Indicateur d'alignement de paquets de flux de donnees dans un systeme pour assembler les paquets de donnees en vue de leur transport
JP4497483B2 (ja) データストリームを処理する装置
JP4648489B2 (ja) データストリームを処理する方法
JP5119522B2 (ja) データストリームを処理する装置
JP5110721B2 (ja) データストリームを処理する装置
JP5264003B2 (ja) データストリームを処理する装置
JP4808812B2 (ja) データストリームを処理する装置
JP4317925B2 (ja) データストリームを処理する装置及び方法
JP4863326B2 (ja) データストリームを処理する方法
JP4754648B2 (ja) データストリームを処理する方法
JP2004357340A (ja) データストリームを処理する装置

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

17P Request for examination filed

Effective date: 19960528

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): DE ES FR GB IT PT

A4 Supplementary search report drawn up and despatched

Effective date: 19981118

AK Designated contracting states

Kind code of ref document: A4

Designated state(s): DE ES FR GB IT PT

17Q First examination report despatched

Effective date: 19990819

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE APPLICATION IS DEEMED TO BE WITHDRAWN

18D Application deemed to be withdrawn

Effective date: 19991230