Method of setting a system time clock at the start of an MPEG sequence

Rzeczpospolitej Polskiej (21) Application number: 354840 (22) Application date: January 5, 2001 (86) Date and number of the international application:

01/05/2001, PCT / EP01 / 000110 (87) International application publication date and number:

07/19/2001, WO01 / 52554 (11) 222487 (13) B1 (51) Int.Cl.

G11B 27/10 (2006.01) H04N 7/56 (2006.01) H04N 5/92 (2006.01) G11B 27/00 (2006.01)

A method for determining the start value of a local system timer of a recording and / or reproducing apparatus, a playback timing method (54) from a real-time recording medium, a real-time recording apparatus on the recording medium, and a real-time playback apparatus from the recording medium

<td></td><td>(73) The right holder of the patent:</td>

<td>(30) Priority:</td><td>KONINKLIJKE PHILIPS NV, Eindhoven, NL</td>

<td>2000-01-10, EP, 00200038.8</td><td>(72) Inventor (s):</td>

<td>(43) Application announced:</td><td>DECLAN P. KELLY, Eindhoven, NL</td>

<td>23.02.2004 BUP 04/04</td><td>WILHELMUS J. VAN GESTEL, Eindhoven, NL PIETER B. IJDENS, Eindhoven, NL</td>

<td>(45) The grant of the patent was announced:</td><td></td>

<td>August 31, 2016 WUP 08/16</td><td>(74) Representative:</td>

<td></td><td>item. stalemate. Izabela Ludwicka</td>

PL 222 487 B1

Description of the invention

The present invention relates to a method for determining the start value of a local system timer of a recording and / or reproducing apparatus, a method for real-time playback from a recording medium, a real-time recording apparatus for a recording medium, and a real-time playback apparatus for a recording medium.

Digital information signals representing a real-time stream of audio video information, such as an MPEG-encoded Transport Stream, include time base information corresponding to the sending side. In the case of an MPEG-encoded Transport Stream, the time base information is determined by the Program Clock Reference (PCR) signals sent regularly in Transport Packets (TP). The time base information is used to synchronize the local clock on the receiving side with the clock on the sending side. However, the time base information is not sent in every transport (TP) packet. This causes the local clock to be initially out of sync with the time base information. This means that it is not known for the transport packets (TP) received before synchronization at which time these transport packets should be decoded (in the case of Access Units (AUs) using the Decode Time Stamp (DTS)) or presented (in the case of Access Units). (AU) using the Presentation Time Stamp (PTS)).

In addition, in the event of a discontinuity in the real-time stream resulting from the superposition of different streams of different programs with a mutually different time base, e.g. after editing, correct timing following such discontinuity should be restored after starting processing of the transport packets of the second sequence. However, the value of the count of the PAT (Packet Arrival Time) after the stream interruption will be discontinuous.

Thus, inter alia, it is an object of the invention to remedy the above-mentioned drawbacks.

The essence of the invention consists in the method of determining the start value of a local timer of a system recording and / or reproducing apparatus receiving real-time sequences of an information signal (TS) packet containing audio-video information, the information signal packets being MPEG2 transport stream packets, and the sequence includes at intervals of a multiple of the information signal packet software reference clock (PCR) information used to synchronize the local system time counter (STC) controlled by the generator VCO 2 with software reference clock (PCR) information:

- upon receipt of the first packet (TS) with a software reference clock (PCR) flag, the system time counter (STC) sets to the value of the software reference clock (PCR) flag,

- compares the received software reference clock (PCR) markers with the value of the system time counter (STC) by means of a phase detector for its synchronization, the phase difference being used by a low pass filter (LPF) to control a voltage controlled oscillator (VCO);

- received packets are buffered in the buffer,

- the time of receiving the APAT application packet is determined for each packet, using the APAT application packet reception time counter operating at the same frequency as the local system time counter (STC), and then

- multiplexing the corresponding application packet reception time stamp (APAT) with the received information signal packets to be recorded on a medium, the method characterized in that

- before receiving the first information signal packet (TS), the value of the application packet reception time (APAT) counter is set to an arbitrary amount, and then

- temporarily buffering in the buffer the APAT [START] application packet receipt time stamp of the first information signal (TS) packet for a given sequence of the first information signal packet, and the application packet receipt (APAT) [PCR] time stamp of the packet, which includes software clock information reference (PCR), where the time stamp of the application packet (APAT) of a given packet is a value representing the time of receipt of the given packet, then

- the number of application packet receipt (APAT) count counts occurring between the APAT [START] and (APAT) [PCR] application packet receiving time stamps, and then

PL 222 487 B1

- subtract the count counts of the APAT application packet receipt timer from the value of the software reference clock (PCR) having the same time unit as the VCO clock period to obtain the initial value of the system time counter (STC) [START], then

the initial value of the system timer (STC-Start) as an attribute of the stored sequence is captured on a recording medium.

Moreover, the essence of the invention consists in that in a method for synchronizing the reproduction from a real-time recording medium of a sequence of information signal (TS) packets, including video audio information, being MPEG2 transport stream packets, the information signal packets further comprising recorded packet receipt time stamps applications (APAT), which indicate the time of receipt of the corresponding packet by the recording device:

- demultiplexing the sequence of reconstructed information signal (TS) packets so as to obtain stored application packet receipt time stamps (APAT),

- the received packets (TS) are buffered in the buffer together with the saved application packet receiving time stamps (APAT), in parallel, for each packet (TS) the APAT application packet receiving time stamp is determined, using the APAT application packet receipt time counter operating at the same frequency every local system timer (STC),

- for the same packet (TS), the buffered application packet receipt time stamp (APAT) is compared in the comparator (38) with the determined application packet receipt time stamp (APAT),

- an information signal (TS) packet from the buffer is output to the output circuit if the generated application packet receipt time stamp (APAT) coincides with the buffered application packet receipt time stamp (APAT), characterized by that

- the sequence of reconstructed information signal (TS) packets is demultiplexed so as to obtain further a sequence attribute which is the start value (STC-Start) of the system timer, the start value (STC-Start) of the system timer corresponds to a software reference clock (PCR) value reduced by the difference between the application packet receipt time (APAT) [PCR] stamp of the first signal packet, which included the value of the software reference clock (PCR) counter and the time stamp of the application packet (APAT) [START] of the first packet in the sequence, then

- before receiving the first packet of information signals (TS), the system time counter (STC) is set to the demultiplexed start value of the system timer (STC-Start) and the value of the application packet receipt timer (APAT) to the value of the application packet receipt time stamp ( APAT) [START].

According to a further aspect of the invention, the essence thereof consists in that a device for real-time recording on the recording medium of a sequence of information signal packets containing audio video information, the information signal packets being MPEG2 transport stream packets, and the sequence including information signal packet programming information of the reference clock (PCR), for synchronizing the local system time counter (STC) with software reference clock (PCR) information, containing:

- a receiving unit for receiving information signal (TS) packets,

- a buffer for temporarily storing the received information signal packets,

- a time stamp generation unit for generating an application packet receipt time stamp (APAT) corresponding to the reception time of information signal packets (TS), which comprises a system time counter (STC) for generating counts and a phase detector connected to its output and and a generator connected to its input VCO,

- a multiplexer for multiplexing the generated application packet receipt time stamps (APAT) and information signal (TS) packets,

- a recording unit for storing the generated application packet receipt time stamps (APAT) and information signal packets (TS) on the recording medium, characterized in that

- a timestamp generating unit, further comprising an application packet receipt (APAT) timer for generating an application packet receipt (APAT) time stamp at the time of receiving an application packet, based on the local system timer (STC) indications, furthermore the time stamp generating unit is adapted for determining the number of cycles of the local counter

System time (STC) between the application packet receipt (APAT) [PCR] time stamp of the first information signal packet, which contained a software reference clock (PCR) counter value, and the application packet receipt time (APAT) [START] time stamp of the first packet in sequence, and the marker generating unit is adapted to subtract the number of counts thus obtained from the software reference clock (PCR) value to obtain the start value of the system time counter (STC-start), and the multiplexer is adapted to multiplex the generated application packet receipt time stamps (APAT), signal packets information (TS) and the obtained system timer start value (STC-start) as an attribute of the sequence.

In yet another aspect, the essence of the invention consists in that the device for real-time reproducing from the recording medium of a sequence of information signal (TS) packets including video audio information, the information signal packets being MPEG2 transport stream packets, the information signal packets further comprising recorded time stamps of receiving an application packet (APAT), indicating the time of receipt of the corresponding packet by the recording device, the device comprises:

- a reading unit for reading the information recorded on the recording medium,

- a demultiplexer adapted to demultipleset the sequence of reconstructed information signal (TS) packets, so as to obtain stored application packet receipt time stamps (APAT),

a buffer for temporarily storing information signal packets read from the recording medium and for demultiplexed time stamps of application packet receipt (APAT),

- a time stamp generation unit for generating an application packet receipt time stamp (APAT) corresponding to the time of receiving an information signal (TS) packets, the unit (37) including a system timer (STC) for generating the counts and a phase detector connected to its output and connected with its input VCO generator (16),

a comparator comparing the application packet receipt time stamp (APAT) temporarily stored in the buffer, with the application packet receipt time stamp (APAT) generated by the time stamp generation unit,

- the output device outputting information signal packets from the buffer when the generated time stamp value is equal to the value of the corresponding application packet reception time stamp (APAT) stored in the information signal packet, characterized in that

- the demultiplexer is adapted to demultipleset the sequence of reconstructed information signal packets (TS), so as to further obtain a system time counter start value (STC-start) written in the MPEG transport stream as a sequence attribute, and the marker generation unit comprises an application packet receipt timer (APAT) ), which counts from the value of the time stamp of receiving the APAT [start] application packet of the first received packet, and a system timer (STC) that counts from the demultiplexed start value of the local system timer (STC-start).

Computing the system time clock value of the first information signal packet improves the reproducibility and simplifies the processing during playback.

The subject of the invention has been presented in a preferred embodiment in the drawing, in which: Fig. 1 shows schematically the sequence of transport packets of an MPEG-encoded stream at the time of starting transmission, Fig. 2 shows a discontinuity between two sequences of transport packets of a stream encoded with the use of the standard MPEG, fig. 3 4 shows an example of recording transport packets at the beginning of a sequence according to the invention, FIG. 5 shows a time stamp generator means of the recording / reproducing apparatus in the recording / reproducing apparatus in the recording / reproducing apparatus. 6 shows an example of recording a transport packet during a discontinuity time according to the invention, Fig. 6. 7 shows a recording apparatus using the time stamp generator means of Fig. 3, Fig. 8 shows a reproducing apparatus using the time stamp generator means of Fig. 5, Fig. 9 illustrates a preferred embodiment of the invention for reading data from an encoded data stream to implement effects 10 shows an example of random access to an MPEG2 transport stream data stream after a PID change, and Fig. 11 shows the special effects for the MPEG2 transport stream data stream after the PAT / PMT has changed.

PL 222 487 B1

Figure 1 illustrates the sequence of MPEG transport packets (TS packets). The sequence begins with the first TS packet 1. The TS 2 packets constitute an encoded access unit AU 3 to be presented as a decoded Presentation Unit at a time interval determined by the corresponding Presentation Time Stamp (PTS). The access unit was received before the local system clock at the receiving side, such as the 27MHz PLL phase locked loop, was synchronized to the time base in the data stream. A software reference clock (PCR) flag is received with TP 5. Thus, it is not known when the access unit should be presented because the presentation time stamp (PTP) points to the time interval before receiving the first software reference clock (PCR).

Figure 2 shows a break in the stream of MPEG transport packets (TS packets). The first sequence of TS 6 packets is followed by a second sequence of TS 7 packets. Each sequence has its own time base information or software reference clock (PCR) flag. This can happen after editing a stream. Thus, the counter that measures the time of packet receipt is discontinuous. The last access unit (AU) formed by the first sequence TS packets 8 is displayed as the presentation unit 9 uninterrupted with the successive presentation units 10, 11 and 12 of the second sequence. However, the first TS packet with a software reference clock (PCR) flag of the second sequence 7 arrives with the TS packet when the TS packet 14 to be presented has been received earlier. Thus, the local system clock has not yet been synchronized with the PCR of the second sequence.

Figure 3 shows a time stamp generator means in a recorder / reproducer according to a first embodiment of the invention, a 27 MHz voltage controlled oscillator controls a system time counter (STC) that is set to a value arbitrarily selected at the start time and counts according to the standard. MPEG (PCR, PTS, DTS). Upon receipt of the first software reference clock (PCR) flag, the system time counter (STC) takes the value of the software reference clock (PCR) flag. Further synchronization is performed with a phase detector which compares the received software reference clock (PCR) marker information with a system clock counter (STC) value. The phase difference is used by a low pass filter (LPF) 10 to drive a voltage controlled oscillator (VCO) forming a phase locked loop (PLL). The system clock is used to control the Application Packet Receive Time (APAT) binary counter 2 to generate the appropriate APAT time stamps.

At the start, the APAT counter starts counting from an arbitrary value. APAT timestamps are appended to each TS packet received. Timestamps represent the time when TS packets were received. The initial APAT [START] timestamp of the sequence TS packet and the APAT tag [_PCr] of a TS packet that contains a software reference clock (PCR) flag is temporarily stored in memory. By subtracting APAT [_start] from APAT [_pCr], the number of 27 MHz cycles between the two time stamps is calculated. Knowing this difference, the start time of the system clock counter (STC) is calculated by subtracting this difference from the earlier PCR value. The start of the STC is the value the STC counter would take if it were synchronized right from the start. The start value STC is preferably stored as a segment attribute when recording the MPEG stream to a recording medium such as a disk.

Figure 4 shows an example of reconstructing transport packets from the initial sequence according to the invention. TS 21 packets received at irregular intervals are shown, the time of receiving TS 21 packets is determined by the APAT time stamp. The digital interface during playback should keep the time between consecutive TS 21 packets constant. The start segment 22 need not start with a programmable reference clock (PCR) marker, this information is received later with the TS packet 23. The programmable reference clock (PCR) information repetition rate should be 100 ms, with 40 ms recommended. The received TS 21 packets are temporarily stored in the smoothing buffer 24. Causes a start delay until the presentation unit (PU) 25 determined by the access unit (AU) 26 containing the respective TS 21 packets is presented. Note that this delay is necessary if stream APAT synchronization is to be maintained.

During playback from the contents of the smoothing buffer, it is possible to restore the original sequence and timing, as shown with reference to Fig. 5. Fig. 5 shows the means of the time stamp generator for generating proper timing of the stored stru6

The disclosed embodiment is largely similar to that disclosed in Fig. 3, so the reference numbers are identical. The difference is the ability to set the system clock 17 (STC) counter 17 and the application packet receipt time (APAT) counter 20. Immediately after the start, the system clock counter (STC) 17 receives the start value STC that was stored as e.g. a segment attribute as discussed earlier. The system clock (STC) counter 17 is now synchronized with a programmable reference clock (PCR). The Application Packet Receiving (APAT) time counter 20 obtains the Application Packet Receiving (APAT) Time Stamp value from the first TS packet. TS packets are received from the smoothing buffer 24 at the time which is indicated by the Application Packet Receive Timestamp (APAT). It should be noted that the internal decoder does not require such a solution, however the stream interface should start by assigning the contents of the Reference Clock (PCR) software packet as the starting value of STC.

Figure 6 shows an example of recording transport packets during the discontinuity of the stream. The packet receipt time stamps (APAT) of the first sequence 27 and the second sequence 28 are discontinuous at the junction point. The offset between both counters should be calculated. Thereafter, proper timing can be restored in the smoothing buffer. An access unit (AU) 32 is shown forming at least one segment shown as a presentation unit from the first sequence 27. Another presentation unit (PU) 30 corresponds to an access unit (AU) 33 of the second sequence 28. The presentation unit (PU) 29 includes the presentation time stamp PTS- 1e which relates to the first local STC-1 system clock counter. The presentation unit (PU) 30 includes the presentation timestamp PTS-2b corresponding to a local system clock counter STC-2. It is assumed that the junction point is a type C point, which implies by definition that there are no buffer problems during the stream discontinuity, there is no APAT time stamp overlap from the first and second sequences, and presentation units 29 and 30 are presented in a gapless manner. .

From the fact that the presentation is continuous, it follows that according to the local time base STC-1, the first presentation unit of the second segment should be presented at PTC-1e + T. As can be seen from the example of the first presentation unit 30 of the second segment, it is known when this presentation unit should be presented relative to the local time base STC-2: PTS-2b. The number of clock cycles between the first TS packet receipt time and the presentation time is known: PTS-2b - STC-start (2). Thus, it is possible to calculate at what point in relation to the local STC-1 time base, the STC-2 local time base should assume the value STC-start (2). It should be noted that the decoder requires a stall for STC-1 and STC-2 (approximately 1 second).

Figure 7 shows a recording apparatus having an input terminal 34 and receiving means 35 for recording received information signals representing transport packets. Packet detector 36 detects that it has received the received transport packets and the embedded software reference clock (PCR) signals. A time stamp generator 15 as disclosed with reference to Fig. 3, is synchronized with its local system timer with software reference clock (PCR) signals. At startup, the time stamp generator 15 takes an arbitrary selected value used to set the local system clock as disclosed with reference to Fig. 3. The generated time stamps, together with the start value (STC-start) of the system clock counter, are combined with the received transport packets at the combining unit 38. The combined signals are channel-coded using channel coding means 39 and recorded on recording medium 40 using recording means 41. The recording medium may be a disc-type medium in which case it is rotated by the drive means 42 while the read / write beam moves in a radial direction by the translation means 43. The recording medium 40 may be an optical type medium, such as a recordable CD, DVD . In this case, the recording means 41 generates a laser beam for recording and furthermore comprises suitable focusing means. In another embodiment, the recording medium 40 may be a magnetic medium, such as a magnetic disk.

Figure 8 shows a reproduction apparatus adapted to scan a recording medium 40 recorded according to the method of the invention using a radiation beam and corresponding reading means 44. In the case where the recording medium 40 is of the type

In optical such as CD, DVD, the reading means include a laser beam and suitable focusing means allowing the scanning of the recording medium 40. The detected signal is channel-decoded by channel decoding means 45. The decoded transport packets along with the time stamps are delivered to the demultiplexer means 46 for separating the time stamps from the transport packets. The time stamps are provided to the means of comparator 38. The values ​​of the time stamps generated by the digital stamp generating means 37, such as disclosed with reference to Fig. 5, are also provided to the means of comparator 8. The stored start value of the system clock counter (STC-start) is provided to the timestamp generating means 37 to synchronize with this value of the timestamp counter each time it is necessary according to the method of the invention. The generated timestamp value is compared with the saved and read timestamp values. When the two values ​​are equal, the corresponding transport packet stored in the memory buffer 47 is sent to the exit means 48 to generate a real-time stream of the transport packets at the exit terminal 49.

As mentioned, the transport packets may contain real-time audio-video information. A combined recording and reproducing apparatus as described with reference to Fig. 7, respectively Fig. 8 can be used as a disk VCR. For the user's convenience, it is possible to provide the possibility of specifying markers at key places of the recorded audio-video program in order to mark key scenes, end of broadcasting of advertisements and the like. The key points are typically video signal entry points such as I-frames in MPEG2 standard. However, in order to allow the playback apparatus to decode the video signal at these points, additional information is required.

Note that a complete description of the MPEG2 format can be found in the relevant publications of the international standards ISO / IEC 13818. I-frames are internally encoded frames that can be decoded independently of the others, unlike P-frames, which are predictive-encoded and require knowledge to be decoded. previous P frame or I frame. Moreover, B-frames or bidirectional frames, which require knowledge of the previous and next I or P frames to be decoded, may also be distinguished.

The advantageous embodiment is realized by storing additional information with the mark point, which allows decoding to be performed at the mark point. If this is not followed, it may take some time (1-2 seconds) to start the decoding process and this portion of the video will not be displayed correctly.

For the MPEG2 transport stream, the marker point should store the following information: a programmable reference clock (PCR) marker at the entry point, presentation time stamp (PTS) for the I-frame, decoding time stamp (DTS) for the I-frame, and packet identification (ID) mapping onto stream. This information enables the decoder to start correct decoding from the mark point.

For the digital video stream of the MPEG2 type described above, to realize the special effects, i.e. to play back a video sequence at a speed other than the normal playback speed, only a part of the video stream is downloaded and decoded and the rest is skipped. In many cases, such as in the case of a DVD, pointers have been introduced to indicate both the start of the required data and the end of the required data without the need to analyze the stream.

The preferred method and embodiment will be described with reference to the case where the end of the required data is not stored, forcing the reproduction apparatus to analyze the stream to determine which parts of it should be ignored.

If the reproduction apparatus does not identify where the end of the information for which the special effects are performed is located in the stream, then a simple solution is to read all the data of the stream from the first start point to the next start point. Such a solution increases the amount of memory required that must be installed in the device and that is required to implement special effects, and the performance requirements of the recording medium are also increased. The preferred method and embodiment disclosed introduce a method for reducing the amount of data that must be read from the recording medium and then stored in the device memory.

PL 222 487 B1

Two types of special effects were considered. The first, in which only I-frames are read from the stream, and the second, in which I and P-frames are read. It is assumed that the start locations of I-frames are recorded, however, their end locations are unknown, and neither of the P-frame points is known.

The basic observation underlying the preferred embodiment and method is that instead of reading a complete group of pictures (GOP) to obtain an I-frame, only a fragment of the GOP can be read based on the estimated size of the I-frame. The Picture Group (GOP) is defined in MPEG2 format (ISO / IEC 13818), and includes at least one I-frame and one or more P and B-frames. For example, in a DVD section, the average I-frame size may be 28 sectors and the average GOP size may be 199 sectors. This leads to a fairly obvious choice that you only need to read a quarter (50 sectors) of the GOP to get the I-frame. This is twice the average, so it should be sufficient for even the worst case scenario. The estimates used should be based on measurements of the broadcast streams and may differ for HDTV streams and SD streams.

The same approach works for the implementation of special effects using P-frames as well as I-frames. In this case, the process of the read GOP will be larger.

Based on the focal point information used in the implementation of special effects, such as information, for example, disclosed in international patent application EP 99/08 285 (PHN 17 161), the presentation time stamp (PTS) for the I-frame and the subsequent I-frame is known. This makes it possible to calculate the number of frames in a GOP. This information can advantageously be used to modify the overall estimate for each specific GOP structure.

With this approach, it is possible for a situation where the entire I-frame is not to be read. If this occurs, it is not a problem. It only means that the refresh rate for that special effect will be reduced.

Reading a stream and frames that are larger than expected will cause the image to appear incorrectly when playing the special effect. In order to avoid this phenomenon, the algorithm used is an adaptive algorithm. For example, if two I-frames that are larger than estimated are observed in a given time period, the reading percentage of the GOP is increased. If the situation repeats itself, the GOP reading percentage is incremented again. This algorithm should stabilize very quickly on a value that is large enough. It is also possible to adaptively reduce the amount of data read. This solution may be particularly useful if P-frames are used to implement a special effect on a stream devoid of B-frames.

Certain encoders, and hence specific streams, are very regular with respect to the actual size of the images they use. In addition, encoders typically use a fixed GOP size. Thus, the adaptive approach should be very effective in practice. By using the Presentation Time Stamp (PTS) time in the Characteristic Point Information (CPI) to calculate the number of images in a GOP, this method is also applicable to irregular GOP structures.

Alternatively, the stream can be analyzed during the write execution for the end of the I-frame and the percentage of GOP to be read and retained during the execution of the special effect in order to obtain a full I-frame 95% or 99% of the time.

This method will be equally effective for multiple video streams broadcast on a single channel. In this case, the percentage of GOP readings will be the same, but the actual amount of data will be greater.

Figure 9 shows an example of an MPEG2 encoded data stream, arrows indicate entry points stored in characteristic point information (CPI). Without knowing the position of the end points of the I-frames, the reproducing apparatus according to the invention does not need to read the data stream until the next entry point when performing the special effect. It should be noted that the amount of data read depends on the amount of data in the GOP.

Another preferred embodiment will be described with reference to handling packet identification (PID) changes in a writer when receiving an information signal packet stream such as MPEG2 transport streams. Such a situation may, for example, occur when broadcasting a digital television program based on the use of MPEG2 transport streams. Packet identifiers (PIDs) are used to identify different streams

PL 222 487 B1 in multiplexed streams. For example, a PID for a video signal, a PID for an audio signal, a PID for a timing information, and a PID for teletext information may be entered. In the case of a TV broadcast with multiple video streams or audio streams in one program, the PID will be entered for each video stream and each audio stream. When broadcasting digitally, the PID may change whether the new PID replaces the old PID or whether there is a change in the association between the PID and the streams. The PID mapping change is signaled by program association tables (PAT) and program map tables (PMT) in the MPEG transport stream. Thus, if a digital TV signal is processed as a stream, the decoder will know when the PID changes and will also know the new mapping for the PID.

Note that according to the MPEG2 standard, the program association table (PAT) maps program units to their corresponding program transport streams. The PAT indicates the PID of the bitstream containing the program map table (PMT) for the program.

The problem is that a digital TV signal is recorded, it is not always completely reproduced from start to finish. The reading device may jump within the stream (random access), or it may select only fragments of the stream (special effects) for reproduction. Thus, the reproduction device may not recognize that the PID mapping has changed before starting to decode the stream. For example, when performing a special effect, the audio signal is normally filtered out of the stream. If the correct PID mapping is unknown, then it is impossible to filter the audio signal, in some cases the video signal may be filtered instead of the audio signal (if the PID of the audio and video signal is changed). In addition, the writer may introduce additional PID changes due to stream editing.

The method and apparatus of the invention include storing metadata relating to the record to record points where PID change occurs. New PID mapping can also be saved. For each PID change, at least the following information should be recorded: 1) the time in the stream where the PID changes, 2) the location in the stream where the PID changes, e.g. by specifying a transport stream packet (TS) where the new PID, 3) program number, 4) PID of programmable reference clock (PCR), 5) PID of video signals, 6) PID of audio signals.

In the case of multiple video streams or multiple audio streams, the relationship between the streams should be maintained. For example, it can be done directly. The sequence of flows in the structure would define such a relationship.

Figure 10 illustrates the case of random access to an audio video stream such as an MPEG2 transport stream, after the PID change, when the playback device hops to the entry point in the stream, it needs to know the PID mapping to start decoding and presenting the data. The PAT / PMT tables that define the PID mapping are repeated in the stream but are generally not present just before the entry point. By searching for meta-data that stores PID changes, the reproduction apparatus can know the correct PID corresponding to that part of the program, and thus perform proper stream multiplexing and decoding. In the case of multiple video streams or multiple audio streams, the reproducing device in this way can be sure to present a video stream which exactly matches the previous one if this is the case. If the stream is transmitted using a digital interface, then the reproducing apparatus may use the meta-data to embed a new PAT and PMT table to indicate a new PID mapping.

Figure 11 shows the special effects related to the audio stream or video stream after the PAT / PMT table change. The data subject to the realized special effects and to be reproduced are indicated by pointers 59. The meta-data defining the PID mapping enables the filtering of non-video streams and ensures that special effects will be applied to the correct video stream in the case of multiple video streams. If the special effect stream is sent via the interface, then the PID of the video can be changed during playback, then standard playback is also resumed, the reproduction apparatus can read new PAT and PMT tables reflecting the new PID mapping.

Although the invention has been described with reference to preferred embodiments thereof, it should be noted that these are not limiting examples. Thus, it is possible for those appropriately educated to make a number of modifications without departing from the scope of the invention as defined by the claims. The same device element may represent a plurality of "measures. Moreover, the invention10

Any and all novel features or any and all combinations of features are all. It should be noted that the word "comprises does not exclude the presence of other elements or steps than those mentioned in the claim. Any reference signs do not limit the scope of the invention.