JPH10257437A - Mpeg encoder/decoder - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an MPEG encoder/decoder for preventing noises from being generated at the time of reproduction start even without performing any complicated processing on the side of the decoder when starting reproduction from the head of GOP in the middle of the stream such as jump reproduction. SOLUTION: An audio frame detection part 105 retrieves an audio stream provided by MPEG encoding and detects the respective heads of plural audio frames consisting of that audio stream. Based on the detected result, an audio packeting part 106 packets the audio stream so as not to store the plural audio frames dividely in mutually different packets.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an MPEG encoding / decoding system.
More specifically, the present invention relates to an MPEG encoding / decoding device having a function of preventing noise generated when performing jump reproduction.

[0002]

2. Description of the Related Art In recent years, a large amount of video and audio information has been accumulated,
Research on the video server provided is underway. In a video server, video / audio information is digitized and encoded, and the amount of information is stored on a disk or the like as compressed data (stream) in which the information amount is largely compressed. Generally, the encoding method is MPEG (Moving Pictures), which is an international standard method for moving picture encoding.
Experts Group) is employed. It is provided in such a video server, etc.
A device that encodes / decodes audio information is an MPEG encoding / decoding device.

FIG. 7 is a diagram showing the structure of a stream used in MPEG. 7A shows a program stream, FIG. 7B shows a video elementary stream, and FIG.
Indicates an audio elementary stream. The program stream is composed of a PES packet storing a video elementary stream in a payload and a PES packet storing an audio elementary stream in a payload.
It is obtained by multiplexing with an ES packet.

The following describes how the streams shown in FIGS. 7 (1) to 7 (3) are constructed. A plurality of video frames are composed of a set of video data obtained by digitization and encoding by MPEG, and a picture header is added to the head of each video frame. In the picture header to be given,
The attribute (I / P / B frame) of the video frame is described.

[0005] Here, the I frame is an intra-coded frame, and can be decoded by itself. The P frame is a frame encoded with reference to a past I or P frame, and the B frame is a frame encoded with reference to two past and future frames. In the P and B frames, since the difference from the predicted frame is encoded, the data amount is reduced as compared with the I frame, but the frame cannot be decoded alone.

A plurality of video frames constitute one GOP (GOP).
group of Pictures) is composed and GO
A GOP header is added to the head of P. A GOP is the minimum unit of a set of video data that can be randomly accessed, and generally 15 frames are defined as one GOP. Further, one sequence is composed of a plurality of GOPs, and a sequence header is added to the beginning of the sequence. Then, a video elementary stream is composed of a plurality of sequences.

On the other hand, a plurality of audio frames are constituted by a set of audio data obtained by digitizing and encoding by MPEG, and a header is added to the head of each audio frame. An audio frame is the smallest unit of a set of audio data that can be decoded by itself.

[0008] The video elementary stream and the audio elementary stream configured as described above are packetized by dividing into a plurality of PES packets and storing them.
Since the length (payload length) of the payload portion of the PES packet is determined uniformly in advance, the video elementary stream and the audio elementary stream are each divided according to the payload length. Therefore, one video frame / audio frame may be divided into a plurality of PES packets and stored. Note that neither video nor audio data is stored in one PES packet.

In a PES packet including a head portion of a video frame or an audio frame, a time stamp (Presentation) indicating a time at which the frame is to be displayed is provided in each header portion (PES header).
Time Stamp (hereinafter, PTS) is included. Further, a pack is composed of a plurality of PES packets, and a pack header is added to the head of the pack. And
A program stream is composed of a plurality of packs.

[0010] By the way, video-on-demand (V), which supplies video and audio data as needed to one of the applications in the video server in response to a request from the client.
OD). In VOD, it is required to realize special playback functions such as high-speed playback, slow playback, and jump playback. Among these, the jump playback is a function of starting playback from a position arbitrarily designated by the user.
In a video server using G, since the GOP is a unit of random access, the reproduction of the program stream must be started from the beginning of the GOP including the specified reproduction start position.

However, as described above, the program stream is composed of a plurality of PES packets.
Input to the decoder is also performed in packet units. And
Each PES packet stores a video elementary stream / audio elementary stream divided according to a predetermined payload length. Therefore, at the start of playback, the beginning of the payload of the video packet to be decoded first includes the remaining portion of the video frame (hereinafter, a fragment of the video frame) of which a part is stored in the immediately preceding video packet. And the fragments are played back, which may result in video distortion.

Similarly, at the start of reproduction, at the head position of the payload of the audio packet to be decoded first, the rest of the audio frame (a fragment of the audio frame), a part of which is stored in the immediately preceding audio packet. May be included, and noise may be generated as a result of reproducing the fragment.

As a device for preventing the generation of this kind of noise, an information recording / reproducing apparatus disclosed in Japanese Patent Laid-Open No. 6-86224 is known. Although the noise generated is prevented, the noise generated when a still image is reproduced is equivalent to the noise generated during jump reproduction.) In this device, the playback device detects an audio synchronization signal added to the beginning of an audio frame. Until the audio synchronization signal is detected, the transmitted audio packet is not input to the decoder,
The operation of inputting to the decoder from the packet following the detection of the audio synchronization signal is started. This allows
At the start of the reproduction, no fragment of the audio frame is input to the decoder, so that generation of noise is prevented.

FIG. 8 is a diagram showing an example of a program stream near the reproduction start position during jump reproduction. In FIG. 8,..., VN, V0, V1,... Are video packets, respectively, and V0 stores the video data at the head of the GOP. …, AM, A0, A1,
Are audio packets.

When the stream shown in FIG. 8 is jumped to the playback start position shown in the figure to start playback, packets after V0 are sequentially input to the decoder. At the start of playback, the first input video packet (V0) contains a video frame fragment, which causes video disturbance, and the first input audio packet (A0) contains an audio frame fragment. It has already been explained that noise is generated due to noise.

In addition, the audio packet (A0) input first does not always correspond to the video packet after the reproduction start position, and may correspond to, for example, VM or VN. In this case, at the time of starting the reproduction, there is a disadvantage that the sound before the reproduction start position is reproduced.

In the apparatus disclosed in the above publication, in order to prevent this inconvenience, at the time of starting reproduction, the first audio packet is at least ΔT longer than the first video packet.
(= T2-T1; T1 is the transmission time of the video packet, T
(2 is the display time), the operation of inputting the audio packet to the decoder is started at the time when the time ΔT elapses after the first video packet is transmitted after the first video packet is transmitted. I have to. As a result, at the start of the reproduction, the sound before the reproduction start position is not reproduced.

[0018]

However, in the apparatus disclosed in the above publication, a reproducing apparatus (decoding apparatus) detects an audio synchronizing signal, and an operation of inputting an audio packet to a decoder based on the detection result. , The configuration of the reproducing apparatus is complicated. Therefore, when the same function is realized in an MPEG encoding / decoding device in which a decoding device is provided for each client, as the number of clients increases, the entire MPEG encoding / decoding device becomes larger and larger. It will be expensive.

In the apparatus disclosed in the above publication, it is assumed that the transmitted video packets are temporarily buffered, while the transmitted audio packets are not buffered and are reproduced in real time. ing. However, since the MPEG decoding apparatus normally buffers and reproduces the audio packet once, the operation of inputting the audio packet to the decoder is started when a time ΔT has elapsed since the first video packet was sent. However, the sound before the reproduction start position is not necessarily reproduced.

Therefore, an object of the present invention is to reduce noise generated at the start of reproduction, such as jump reproduction, when starting reproduction from the beginning of a GOP in the middle of a stream without performing complicated processing on the decoding device side. MP that can be prevented
An EG encoding device is provided.

Another object of the present invention is to provide an MPEG decoding system which can prevent reproduction of sound before a reproduction start position when reproduction is started from the beginning of a GOP in the middle of a stream, such as jump reproduction. It is to provide a chemical conversion device.

[0022]

Means for Solving the Problems and Effects of the Invention The MPEG encoding apparatus according to the first invention converts audio data into MP data.
MPEG encoding means for encoding in the EG system, MPEG
An audio frame detecting unit that searches for an audio stream obtained by the encoding unit and detects the head of each of a plurality of audio frames forming the stream; And packetizing means for packetizing the audio stream, wherein the packetizing means performs packetization such that none of the plurality of audio frames is divided into different packets and stored. Features.

As described above, in the first invention, MPEG
An audio stream obtained by encoding according to the method is searched, and the head of each of a plurality of audio frames constituting the stream is detected. Then, based on the detection result, the audio stream is packetized so that any of the plurality of audio frames is not divided into different packets and stored.

GO in the middle of a stream such as jump playback
When playback is started from the beginning of P, at the beginning of the payload portion of the audio packet to be decoded first, the remaining portion of the audio frame whose part is stored in the immediately preceding audio packet (hereinafter, the audio frame) Fragment) is stored, and as a result of reproducing the fragment, noise may be generated. Therefore, when packetizing an audio stream, each audio frame is packetized so as not to be divided into different packets and stored. As a result, a fragment of an audio frame is not stored in an audio packet to be decoded first. As a result, noise is generated at the start of reproduction without performing complicated processing on the decoding device side. Disappears.

An MPEG decoding apparatus according to a second aspect of the present invention provides a buffer for temporarily storing a video stream and an audio stream, and an MPE for decoding the video stream and the audio stream stored in the buffer.
G decoding means, and control means for controlling the operation of the MPEG decoding means, wherein at the start of playback, the audio frame currently being decoded among a plurality of audio frames constituting the audio stream, The display time of the audio frame is compared with that of the video frame to be displayed first among the plurality of video frames constituting the video stream, and as a result of the comparison, the display time of the audio frame is more than that of the video frame. If it is early, it is characterized in that it instructs the MPEG decoding means to not output audio data obtained by decoding the audio frame.

As described above, according to the second invention, at the start of reproduction, the display time of the audio frame being read from the buffer and being decoded and the video frame stored in the buffer to be displayed first are displayed. Compare with that of If the comparison result shows that the display time of the audio frame being decoded is earlier than that of the video frame to be displayed first, the audio data obtained by decoding the audio frame is output. Absent.

GO in the middle of a stream such as jump playback
When the reproduction is started from the beginning of P, the audio frame to be decoded first corresponds to the video frame before the reproduction start position, and as a result, the sound before the reproduction start position may be reproduced. Therefore, at the start of playback, the display time of the audio frame currently being decoded is compared with the display time of the first video frame stored in the buffer to be displayed. If the result of the comparison shows that the display time of the audio frame being decoded is earlier than that of the video frame to be displayed first,
The audio data obtained by decoding the audio frame is not output. As a result, at the start of the reproduction, the sound before the reproduction start position is not reproduced.

[0028]

BEST MODE FOR CARRYING OUT THE INVENTION

(First Embodiment) Hereinafter, a first embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration of the MPEG encoding / decoding device according to the first embodiment of the present invention. The MPEG encoding / decoding device in FIG. 1 includes an encoding device 10 and a decoding device 11, and the encoding device 10 includes an MPEG encoder 101 and a G encoder.
OP detector 102, video frame detector 103, video packetizer 104, audio frame detector 10
5, an audio packetizing unit 106 and a multiplexing unit 107 are provided. The decoding device 11 includes an MPEG decoder 11
1. A buffer unit 112 is provided. Buffer unit 112
Are the channel buffer 112a and the header buffer 1
12b.

The MPEG encoder 101 encodes video data and audio data using MPEG. G
The OP detection unit 102 searches for a video stream obtained by encoding by the MPEG encoder 101, and detects the beginning of a GOP. The video frame detection unit 103
The EG encoder 101 searches for a video stream obtained by encoding, and detects the beginning of a video frame.
The video packetizer 104 includes the MPEG encoder 10
1 packetizes the video stream obtained by encoding.

The audio frame detection unit 105 outputs the MP
The EG encoder 101 searches the audio stream obtained by encoding, and detects the head of the audio frame. The audio packetizing unit 106 packetizes the audio stream obtained by encoding by the MPEG encoder 101. The multiplexing unit 107 includes a PES packet (video packet) obtained by packetizing the video packetizing unit 104 and the audio packetizing unit 10.
6 multiplexes PES packets (audio packets) obtained by packetization.

The MPEG decoder 111 includes a multiplexing unit 10
7 temporarily stores the program stream obtained by multiplexing in the buffer unit 112 and then decodes the program stream. In the buffer unit 112, the channel buffer 112a stores a video stream and an audio stream in packet units, and the header buffer 112b stores a pack header and a PES header.

The operation of the MPEG encoding / decoding apparatus shown in FIG. 1 will be described below. In the apparatus of FIG. 1, for example, an encoding device 10 is provided on a video server side, and a decoding device 11 is provided on a client side. Also, the output of the encoding device 10 (a program stream as shown in FIG. 7A) is temporarily stored on a disk or the like (not shown), read out from the disk in response to a request from the client side, and decoded. It is sent to the device 11 side.

In the encoding device 10, first, video data and audio data are transmitted to the MPEG encoder 101.
Is input to The MPEG encoder 101 converts the input video data and audio data into MP
The video elementary stream and the audio elementary stream as shown in FIGS. 7 (2) and (3) are output by encoding with EG. GOP detector 102
Is G from the output video elementary stream.
Group_start_cod indicating the start position of the OP
e (0x000001B8) is detected. Then, it notifies the video frame detecting unit 103 and the video packetizing unit 104 of the detected byte position of the group_start_code.

Note that the GOP detection unit 102
_Start_code, instead of detecting sequence_head indicating the start position of the sequence header
er_code (0x000001B3) and gro
Any of the up_start_code may be detected. In this case, the sequence header and G
Whichever byte position of the OP header is detected first is notified to the video frame detecting unit 103 and the video packetizing unit 104.

Next, the video frame detecting unit 103
From the byte position notified by the OP detection unit 102, that is, from the top of the GOP, the picture header
A search for _start_code (0x0000100) is started. The first pict found after starting the search
Based on the byte position of ure_start_code,
The position of the first byte of the I frame is known. Also, based on the byte position of the second detected picture_start_code, B or P immediately after the I frame is used.
The first byte position of a frame (hereinafter, a non-I frame) is known. Therefore, the video frame detection unit 103
Is the non-
The first byte position of the I frame is notified to the video packetizer 104.

The video packetizer 104 calculates the first byte position of the GOP notified by the GOP detector 102 and the first byte position of the I frame notified by the video frame detector 103 and the immediately following non-I frame. And packetizes the video elementary stream. The PES obtained by packetization
It is assumed that the payload length of the packet is a predetermined length for both the video packet and the audio packet.

Here, the operation of the video packetizer 104 packetizing the video elementary stream will be described with reference to FIG. That is, FIG. 2 is a diagram for explaining an operation in which the video packetizing unit 104 in FIG. 1 packetizes a video elementary stream. As shown in FIG. 2, in the video elementary stream, the data from the head of the sequence header to the data at the end of the I frame is divided into a plurality of PES packets and stored.

P at which the data at the end of the I frame is stored
In the case where the stored data length is less than the payload length in the ES packet, padding processing is performed on the remaining payload portion where data is not stored (portion indicated by (a) in the figure). By performing this process,
The head data (non-I frame data) of the next frame is not stored in the PES packet storing the data at the end of the I frame.

In the padding process, a padding packet equal to the length of the remaining payload portion is generated and inserted.

Next, data of a series of non-I frames (B frame, P frame,..., P frame in the figure) from the above I frame to the next I frame is added to the payload portion of a plurality of PES packets. Stored separately. At this time, unlike the above I frame, each non-frame
In a PES packet in which data at the end of an I frame is stored, padding is not performed even if the stored data length is less than the payload length.

However, in the data at the end of the last non-I frame, that is, in the PES packet storing the data at the end of the GOP, if the stored data length is less than the payload length, the data is stored. For the remaining parts not shown (parts shown in (b) in the figure),
Perform padding processing. By performing this processing, GO
The head data (I frame data) of the next GOP is not stored in the PES packet storing the data at the end of P.

As described above, the PES packet storing the data at the end of the I frame contains the next frame (non
The first data of the non-I frame (fragment of the video frame) is decoded when only the I frame is to be selectively reproduced, such as during high-speed reproduction, by not storing the first data of the I-frame. As a result, it is possible to prevent video disturbance during high-speed reproduction.

Also, by preventing the head data (I frame data) of the next GOP from being stored in the PES packet storing the data at the end of the GOP,
When playback is started from an arbitrary GOP in the middle of a stream such as during jump playback, data at the end of the immediately preceding GOP (fragment of a video frame) is not decoded, and as a result, video is disturbed at the start of playback. You can not.

On the other hand, the audio frame detecting section 105
Searches the audio elementary stream to detect a synchronization bit (0xFFF) included in the header.
Then, it notifies the audio packetizer 106 of the detected byte position of the synchronization bit. The audio packetizer 106 packetizes the audio elementary stream with reference to the notified byte position.

The operation of the audio packetizing section 106 for packetizing an audio elementary stream will be described with reference to FIG. That is, FIG.
FIG. 6 is a diagram for explaining an operation of the audio packetizing unit 106 packetizing an audio elementary stream. As shown in FIG. 3, the audio packetizer 106 first stores the data of the first audio frame in the payload of the first PES packet. Then, the remaining payload length in which no data is stored is compared with the data length of the second audio frame.

As a result of the comparison, since the remaining payload length is longer than the data length of the second audio frame, the audio frame is stored in the remaining payload portion of the first PES packet. Then, the remaining payload length in which no data is stored is compared with the data length of the third audio frame.

As a result of the comparison, since the remaining payload length is shorter than the data length of the third audio frame, the audio frame is not stored in the remaining payload portion of the first PES packet and the second audio frame is not stored. PES
Store it in the payload part of the packet. Then, the remaining payload portion of the first PES packet ((c) in the figure)
), The same padding processing as described above is performed.

Thereafter, similarly, each time an audio frame is stored in a PES packet, the remaining payload length of the packet is compared with the data length of the next audio frame. If the result of the comparison indicates that the remaining payload length is shorter than the data length of the next audio frame and that all the data of the next frame cannot be stored in the remaining payload portion, the remaining The operation of performing the padding process on the payload portion of is repeated. As a result, the data of one audio frame is not divided and stored in a plurality of PES packets.

Next, the multiplexing unit 107 divides the PES packet (video packet) obtained by packetizing the video packetizing unit 104 into an audio packetizing unit 106.
Multiplexes a PES packet (audio packet) obtained by packetization, adds a pack header to the head of the multiplexed program stream, and outputs the result. The output program stream is temporarily stored on a disk or the like (not shown).

Next, the operation of the decoding device 11 during jump reproduction will be described. The program stream stored in the disk or the like as described above is read from the beginning of the GOP in the middle of the stream according to the instruction of the client, and sent to the decoding device 11. Decryption device 1
In step 1, the MPEG decoder 111 extracts a pack header and a PES header from the sent program stream, and stores them in the header buffer 112b. In addition, a video elementary stream and an audio elementary stream are extracted and
2a.

When the amount of data stored in the channel buffer 112a exceeds a preset threshold, the MPEG
The decoder 111 starts an operation of reading and decoding the video stream and the audio stream from the channel buffer 112a in packet units. The video data and audio data obtained by decoding are respectively converted into analog video signals by a D / A conversion unit (not shown),
It is converted to an analog audio signal. Then, display of an image is started on a screen (not shown), and output of sound from a speaker (not shown) is started.

At this time, on the encoding device 10 side, GO
The next G is added to the PES packet storing the data at the end of P.
Since the video stream is packetized so that the data at the beginning of the OP is not stored, the video packet to be decoded first does not include a fragment of the video frame before the reproduction start position. At the start,
No image distortion occurs.

The audio stream is packetized so that the data of one audio frame is not divided into a plurality of PES packets and stored on the encoding device 10 side. The audio packet does not include a fragment of an audio frame before the reproduction start position, and therefore, no noise is generated at the start of reproduction.

As described above, according to the present embodiment, when starting playback from the beginning of a GOP in the middle of a stream, such as jump playback, the decoding device 11 does not need to perform complicated processing, so that playback can be started at the beginning of playback. It is possible to prevent image distortion and noise from occurring.

(Second Embodiment) Hereinafter, a second embodiment of the present invention will be described with reference to the drawings. FIG.
FIG. 3 is a block diagram showing a configuration of an MPEG encoding / decoding device according to a second embodiment of the present invention. 4 includes an encoding device 10 and a decoding device 11a.
Consists of The encoding device 10 of FIG. 4 has a configuration similar to that of the encoding device 10 of FIG. Decoding device 1 of FIG.
1a is an output control unit 1 in the decoding device 11 of FIG.
13 is further provided. The output control unit 113
The output operation of the G decoder 111 is controlled.

The operation of the MPEG encoding / decoding apparatus shown in FIG. 4 will be described below. However, the encoding device 10
Is similar to that of FIG. 1, and the description is omitted. FIG. 5 is a diagram for explaining an operation in which the MPEG decoder 111 in FIG. 4 stores a header in the header buffer 112b and a video elementary stream and an audio elementary stream in the channel buffer 112a. At the time of jump reproduction, the decoding device 1
In 1a, first, the MPEG decoder 111 extracts a header from a transmitted program stream and stores it in the header buffer 112b, and extracts a video elementary stream and an audio elementary stream and stores them in the channel buffer 112a. I do. The operation will be described below with reference to FIG.

The client has a GO in the middle of the stream
When the reproduction is instructed from P, a program stream as shown in FIG. 7A is read from a disc (not shown) and input to the decoding device 11a. In the decoding device 11a, the MPEG decoder 111 extracts a pack header (not shown) from the input stream and stores it in the channel buffer 112a.

Next, the MPEG decoder 111 converts the PES packets (packet (1),
The PES headers (header (1), header (2),...) Are extracted from the packets (2,...) And stored in the header buffer 112b. Also, video data / audio data (data (1), data (2),...) Are extracted and stored in the channel buffer 112a.

At this time, the MPEG decoder 111
The value of the write pointer is written following the last byte of the ES header. The values of the written write pointer (write pointer (1), write pointer (2),...) Correspond to the corresponding video data / audio data (data (1),
The address (of the channel buffer 112a) at which the first byte of the data (2),... Is written.

Next, the MPEG decoder 111 sequentially reads the PES header from the header buffer 112b and analyzes the contents. As a result of the analysis, PT
If the STS exists (as described above, the PTS packet including the head data of the video frame / audio frame includes the PTS in the PES header), the PTS
(PTS (1), PTS (2),...) And the values of the write pointers (Write Pointer (1), Write Pointer (2),...) Are combined into a PTS table as shown in FIG. Accumulate.

FIG. 6 is a diagram showing an example of the PTS table created by the MPEG decoder 111 of FIG. MPE
The G decoder 111 creates one PTS table as shown in FIG. 6 for each of a video frame and an audio frame. By referring to these tables, the relationship between the storage position of each video frame / audio frame and the display time can be understood.

When the amount of data stored in the channel buffer 112a exceeds a preset threshold value, the MPEG
The decoder 111 starts the operation of reading and decoding the video stream and the audio stream from the channel buffer 112a. However, at the start of playback, the video data obtained by decoding is output, but the audio data obtained by decoding is not output, but invalid data is output (mute processing) instead. .

On the other hand, the output control unit 113 monitors the decoding operation of the MPEG decoder 111. And MP
When the EG decoder 111 is reading and decoding audio data from the channel buffer 112a, the output control unit 113 obtains the value of the write pointer of the audio data from the MPEG decoder 111 and sets the value of the PTS corresponding to the value. From the audio frame PTS table managed by the MPEG decoder 111. The value of the obtained PTS is now M
This is the display time of the audio frame that the PEG decoder 111 is trying to decode.

Next, the output control unit 113 controls the video frame PTS managed by the MPEG decoder 111.
Referring to the table, the acquired PTS value (the display time of the audio frame currently being decoded) and the PTS stored in the video frame PTS table with the smallest value (first decoded PTS) Video frame display time).

As a result of the comparison, the obtained PTS value is
PT stored in the video frame PTS table
If the value of S is smaller than the smallest value, it is understood that the audio frame currently being decoded is before the reproduction start position. On the other hand, when the value of the acquired PTS is larger, it is understood that the audio frame currently being decoded is after the reproduction start position.

Therefore, the output control unit 113 sets the acquired P
At the moment when the value of TS becomes larger than the smallest value of the PTS stored in the PTS table for video frame, it instructs the MPEG decoder 111 to stop outputting invalid data and decode the data. The output of the audio data obtained is started.

The video data and audio data obtained by decoding are converted into an analog video signal and an analog audio signal by a D / A converter (not shown). Then, display of an image is started on a screen (not shown), and output of sound from a speaker (not shown) is started. At this time, since the output control unit 113 performs the output control as described above, the sound before the reproduction start position is not reproduced at the time of the reproduction start.

As described above, according to the present embodiment, when playback is started from the beginning of a GOP in the middle of a stream, such as jump playback, the decoding device 11a, as in the first embodiment.
In addition to being able to prevent video distortion and noise at the start of playback without performing complicated processing on the side,
At the start of playback, audio before the playback start position can be prevented from being played back.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an MPEG encoding / decoding device according to a first embodiment of the present invention.

FIG. 2 is a diagram for explaining an operation in which the video packetizer 104 of FIG. 1 packetizes a video elementary stream.

FIG. 3 is a diagram for explaining an operation in which the audio packetizing unit 106 in FIG. 1 packetizes an audio elementary stream.

FIG. 4 is a block diagram illustrating a configuration of an MPEG encoding / decoding device according to a second embodiment of the present invention.

FIG. 5 is a diagram for explaining an operation in which the MPEG decoder of FIG. 4 stores a header in a header buffer 112b and a video elementary stream and an audio elementary stream in a channel buffer 112a, respectively.

FIG. 6 shows a PT created by the MPEG decoder 111 of FIG.
It is a figure showing an example of an S table.

FIG. 7 is a diagram showing the structure of a stream used in MPEG.

FIG. 8 is a diagram illustrating an example of a program stream near a reproduction start position during jump reproduction.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 ... Encoding device 11, 11a ... Decoding device 101 ... MPEG encoder 102 ... GOP detection part 103 ... Video frame detection part 104 ... Video packet formation part 105 ... Audio frame detection part 106 ... Audio packet formation part 107 ... Multiplexing part 111: MPEG decoder 112: Buffer unit 112a: Channel buffer 112b: Header buffer 113: Output control unit

Claims (2)

[Claims] An MPEG encoding means for encoding audio data in an MPEG system, an audio stream obtained by encoding by the MPEG encoding means is searched, and a plurality of audio frames constituting the stream are searched. Audio frame detection means for detecting the head of each of the plurality of audio streams, and packetization means for packetizing the audio stream in association with the detection result of the audio frame detection means, wherein the packetization means comprises: An MPEG encoding apparatus, characterized in that a frame is packetized so that none of the frames is divided into different packets and stored. 2. A buffer for temporarily storing a video stream and an audio stream; an MPEG decoding unit for decoding the video stream and the audio stream stored in the buffer; and an operation of the MPEG decoding unit. Control means, the control means, at the start of playback, of the plurality of audio frames constituting the audio stream, the display time of the audio frame that is currently being decoded,
If the display time of the audio frame is earlier than that of the video frame as a result of the comparison, the display time of the audio frame is compared with that of the video frame to be displayed first among the plurality of video frames constituting the video stream. An MPEG decoding apparatus, characterized in that the MPEG decoding means is not instructed to output audio data obtained by decoding the audio frame.