JP3123366B2 - Information storage management device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information storage and management apparatus for demultiplexing an information stream formed by multiplexing packets of different data types such as video and audio data.

[0002]

2. Description of the Related Art In general, in data transmission by a packet switching system, data is divided into blocks of a fixed length, and information is added to a header for each block to form a packet, which is multiplexed with other data. That is, they are multiplexed, arranged in series, and transmitted as a stream. On the receiving side, the packet stream is demultiplexed with reference to the header information to reconstruct the data. The information to be transmitted includes, for example, a video signal and an audio signal. In packet transmission, data of different data types can be transmitted by the same transmission system.

As a technique for demultiplexing a data stream formed by multiplexing packets having different data types as described above, for example, Japanese Patent Laid-Open Publication No.
What is shown in -224320 is known.
That is, when writing packets of different data types to the buffer memory, which is a common data memory, write the data length, data type, and other data information of the packet and the write start address of the packet in the buffer memory to the descriptor. When reading data, the data information and the like are read from the descriptor, and packets are read from the buffer memory according to the reading order.

[0004]

By the way, in the conventional apparatus as described above, when demultiplexing a data stream, packet data can be processed only in accordance with the reading order from the descriptor.
In order to process only a packet of a specific data type first, packets of another data type cannot be put on standby, and data processing for each data type cannot be performed.

[0005] The present invention focuses on the above problems,
The purpose of this information storage management was to devise an effective solution to this problem, in which packets of the same data type were taken out of a mixture of packets of different data types to enable processing for each data type. It is to provide a device.

[0006]

In order to solve the above problem, a first invention is to provide a common buffer which does not depend on the data type, and includes a packet group having data information including a data type and a data length as a header. An information storage management device that writes to a memory unit and reads a packet of a desired data type from the buffer memory unit as needed under the control of a central processing unit; a header detection unit that detects a header of the packet; A first-in, first-out memory unit that stores packet information and a first address of the packet in the buffer memory unit, and writes the packet information and the first address to the first-in, first-out memory unit based on information from the header detection unit. An interrupt is generated to the central processing unit, and the data in the central processing unit is A processing unit for writing the address information and the start address in a memory block allocated for each data type according to the data type, and when extracting a packet of a required data type, the memory corresponding to this data type The information is read from the position indicated by the read pointer on the block to obtain information on the packet in the buffer memory unit.

According to a second aspect of the present invention, a packet group having data information including a data type and a data length as a header is written in a common buffer memory unit irrespective of the data type. In an information storage management device that reads a packet of a desired data type as required by control, a header detection unit that detects a header of the packet, and is provided for each type of the packet. A first-in first-out memory unit that stores a first address in the buffer memory unit, and a processing unit that writes the packet information and the first address to the first-in first-out memory unit corresponding to the type of packet based on information from the header detection unit. When extracting packets of the required data type, Is read out from said corresponding first-in-first-out memory unit is obtained by configured to obtain information about the packets in the buffer memory unit.

[0008]

According to the first aspect, the packet stream is sequentially written to the buffer memory unit via the header detection unit, and the header information of the packet detected by the header detection unit and the packet information of the packet in the buffer memory unit are stored in the buffer memory unit. The start address is first-in first-out (hereinafter, FIFO
O: First-In First-Out) is sequentially written and stored in the memory unit. When the writing is completed, the processing unit generates an interrupt to the central processing unit, and stores the contents of the FIFO memory unit in a memory block attached to the central processing unit and allocated for each data type. Write in. When extracting a packet of a required data type, information on the packet is obtained from a memory block corresponding to the data type, and the packet is extracted from the buffer memory unit based on the information.

According to the second aspect, the packet stream is sequentially written to the buffer memory unit via the header detection unit, and information of the packet detected by the header detection unit and the start of the packet in the buffer memory unit The addresses are sequentially written by the processing unit to a plurality of FIFO memory units provided for each data type. When a packet of a required data type is extracted, information on the packet is obtained from the FIFO memory unit corresponding to the data type, and the packet is extracted from the buffer memory unit based on the information.

In the first and second invention devices, the buffer memory section is usually of a ring type, and when writing reaches the upper limit of the buffer memory section, writing is continued from the beginning. If a packet longer than the upper limit of the buffer memory is transmitted during writing to the memory, the packet is written from the beginning of the buffer memory between the header detector and the buffer memory. By providing the second processing unit, the central processing unit does not consider the boundary of the buffer memory unit, and the load can be reduced accordingly.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the information record management device according to the present invention will be described below in detail with reference to the accompanying drawings. FIG. 1 is a packet structure diagram showing the structure of a packet used in the present invention, FIG. 2 is a diagram showing a stream when the packet shown in FIG. 1 is multiplexed, and FIG. 3 shows an information storage management device of the first invention. FIG. 4 is a block diagram showing the configuration of the FIF used in the apparatus shown in FIG.
FIG. 5 is a diagram showing a storage state of an O memory unit, and FIG. 5 is a diagram showing a storage state of a memory block in a central processing unit used in combination with the device shown in FIG.

As shown in FIG. 1, a packet 1 used in the present invention comprises a header area 2 and a data area 3. The header area 2 has a start code 4, a data type code 5, a data length code 6, The auxiliary information code 7 is sequentially arranged. The start code 4 is a code of several bytes indicating the head of the packet, and the data type code 5 is divided into video and audio, and each is classified into several channels. The header length code 6 is the header area 2
Indicates the length of subsequent valid data. The auxiliary information code 7 includes information such as a time stamp. The packet 1 configured as described above is multiplexed by a multiplexer as shown in FIG. 2 to be streamed and transmitted. Further, invalid data 8 can be included between packets.

In this embodiment, the required packet information is a data type, a data length, a time stamp as auxiliary information, and the data type is a video and audio MPEG (Motion Picture) of channel 1.
(Expert Group) format.

The information storage management device 9 according to the present invention shown in FIG. 3 is used for receiving and appropriately distributing the packet stream configured as described above. The header detection unit 10 detects the header of each packet 1 in the packet stream, and all the packets that have passed through the detection unit 10 are stored in the common buffer memory unit 11 regardless of the data type. .

The processing unit 12 sequentially receives the header information detected by the header detection unit 10 and also receives the head address information of the stored packets in the buffer memory unit 11 and stores these information in a FIFO memory unit. 13 sequentially. The processing unit 12
When the writing to the FIFO memory unit 13 is completed,
Immediately, an interrupt signal S1 is sent to the central processing unit 14 to interrupt it, and the FIFO memory unit 13
Is readable.

As is well known, the central processing unit 14 includes a CPU body 15, a RAM 16 for temporarily storing data being processed, and an R for storing processing programs and the like in advance.
It has an OM 17 and the like, and is connected to the buffer memory unit 11 and the FIFO memory unit 13 via a BUS line 18.

The RAM 16 in the central processing unit 14 is allocated and partitioned for each data type by a program, and has memory blocks 16A, 16B... For each data type as shown in FIG. . FIG.
5A shows the data type of video of channel 1 and FIG. 5B shows the data type of audio of channel 1.

The number of memory blocks is not limited to two, but increases according to the number of data types.
The counter unit 19 counts the number of packets input to the buffer memory unit 11 and the processing unit 12 based on information from the header detection unit 10.

Next, the operation of the device of the present invention will be described based on the device configured as described above. First, the RAM 16 in the central processing unit 14 is allocated in advance for each data type as shown in FIG. 5 to form memory blocks 16A and 16B for each data type, and a read pointer and a read pointer for accessing the memory block are respectively provided. A write pointer is prepared in advance.

Next, the sequence of operation of the above device will be described. (A) First, the read pointer and the write pointer are initialized so as to be at the head position of each memory block. (B) Next, when the stream of the multiplexed packet is input to the header detection unit 10, this detection unit 1
0 detects the start code 4 from the header area 2 of each packet and determines the data type of the packet. (C) If the packet input here is a packet required by the header detection unit 10 (here, video and audio of channel 1), the header information and data of the data length are written into the buffer memory unit 11. . It should be noted that only the data of the data length may be written into the buffer memory unit 11 without the header information.

Further, in parallel with the write processing to the buffer memory section 11, header information and the like of each packet are input to the processing section 12, and the processing section 12 determines the type of the packet, the start address in the buffer memory, and the like. , A data length, a time stamp, and the like are sequentially written in the FIFO memory 13. Memory 1 where information is written
The state of No. 3 is shown in FIG.

(D) FIFO memory unit 13 as described above
When the writing to the buffer memory unit 11 is completed, the processing unit 12 sends an interrupt signal S1 to the central processing unit 14 that manages the buffer memory unit 11 to cause an interrupt. (E) When an interrupt occurs, the central processing unit 14
The storage contents are read out to the FIFO memory unit 13 to obtain a series of information on the packet, and the address, the data length, and the like on the buffer memory are obtained from the position indicated by the write pointer on the memory blocks 16A and 16B corresponding to the data type.
The time stamp is stored as shown in FIG. 5, and the write pointer is moved to the head position to be stored next.

The above steps (b) to (e) are repeated, and the header information is fetched into the central processing unit 14 while being distinguished for each data type. However, since the input of the packet stream to the header detector 10 in the step (b) is performed as needed, the sequence operates simultaneously and in parallel.

Apart from the above processing, when extracting a packet of a required data type, first, if there is a packet to be processed based on the positional relationship between the read pointer and the write pointer of the memory block corresponding to this data type, the read is performed. By reading the contents of the memory block from the position indicated by the pointer, the write start address, data length,
Get a timestamp. Then, the read pointer is moved to the head position to be read next.

Next, the central processing unit 14 reads out a corresponding packet from the buffer memory unit 11 based on a series of information of the obtained packet. For example, if the data type is channel 1 video, the video (DMA) is transmitted to the video decoder, and if the data type is audio, the video is transmitted to the audio decoder.
By performing ct memory access (ct memory access) transfer, the bit stream can be transferred at high speed.

Further, the time difference between video and audio is obtained from the obtained time stamp information, whereby the start of the video decoder and the start of the audio decoder are made different, and lip sync can be obtained. In this case, even if one of the decoders cannot accept the data due to the difference in the start of the decoder, the processing is delayed and the other types of packets are processed first. The management can be easily performed by managing the read pointer and the write pointer for each.

As described above, the RAM of the central processing unit 14
Since the packet information and the like for each data type are read from the memory block for each data type allocated to 16 and the packets are extracted from the buffer memory unit 11 based on this, it is possible to individually perform processing for each same type. . Further, since the FIFO memory unit 13 is provided, the central processing unit 14 can easily obtain the memory information. Therefore, it is easy to create a program for the central processing unit 14 and to increase the processing speed. Can be achieved.

In the first aspect of the present invention, one FIFO memory unit is provided to allocate the RAM 16 of the central processing unit 14 for each data type to form a memory block. However, the present invention is not limited to this. The second described below
Multiple FIs corresponding to data types as in the invention device of
The allocation of the RAM may be eliminated by providing the FO memory unit.

FIG. 6 is a block diagram showing an information storage management device according to the second invention, and FIG. 7 is an explanatory diagram for explaining a storage state of a FIFO memory unit used in the device shown in FIG. The same parts as those shown in FIG. 3 are denoted by the same reference numerals, and description thereof will be omitted.

In the device of the present invention, a plurality of FIFO memory units 20 and 21 are provided corresponding to each data type, unlike the first embodiment. One FIF in the figure
The O memory unit 20 is provided, for example, for channel 1 video, and the other FIFO memory unit 21 is provided for channel 1 audio. The FIFO memory units 20 and 21 are provided with register units 22 and 23 for holding the number of packets managed in each memory unit.
Note that the number of the FIFO memory units 20 and 21 is not limited to two, and can be further increased in accordance with the number of data types.

The operation sequence of the above device will be described below. (A) First, the registers 22 and 23 arranged in the two FIFO memory units 20 and 21 are initialized to zero. (B) Next, when the stream of the multiplexed packet is input to the header detection unit 10, this detection unit 1
0 detects the start code 4 from the header area 2 of each packet and determines the data type of the packet. (C) If the packet input here is a packet required by the header detection unit 10, the header information and data corresponding to the data length are written into the buffer memory unit 11.
It should be noted that only the data of the data length may be written into the buffer memory unit 11 without the header information.

Further, in parallel with the write processing to the buffer memory section 11, header information and the like of each packet are input to the processing section 12, and the processing section 12 transmits the head address and data of the packet on the buffer memory. As shown in FIG. 7, a series of information such as length and time stamp is written to the FIFO memory units 20 and 21 corresponding to the packet type, and the number of packets in the corresponding register units 22 and 23 is increased by one at the same time. Each sequence from the above steps (b) to (c) is repeatedly performed. However, step (b)
Since the input of the packet stream to the header detection unit 10 in the above is performed as needed, the sequence operates simultaneously and in parallel.

In order to extract a packet of a required data type separately from the above processing, the central processing unit 14 first reads the contents of the register unit corresponding to this data type, and the number of stored packets is If the value is larger than zero, it means that a necessary packet is stored. Therefore, by reading the contents of the FIFO memory unit corresponding to the data type, the information of the corresponding packet, that is, the write start address in the buffer memory unit 11, Data length and time stamp can be obtained. Here, when a series of information of the packet is read from the FIFO memory unit,
The number of packets in the register section corresponding to this is reduced.

Based on a series of information of these packets,
The central processing unit 14 reads a corresponding packet from the buffer memory unit 11. For example, a bit stream can be transferred at a high speed by DMA-transferring a packet to a video decoder if the data type is channel 1 video and to an audio decoder if the data type is audio.

Also, a time difference between video and audio is obtained from the time stamp information, and the start of the video decoder and the audio decoder is changed from this difference, whereby lip sync can be obtained. At this time, if one of the decoders cannot accept the data due to the change of the start of both coder,
By making this process wait, it is easy to process packets of other types first by managing the register unit for each data type.

In the above two invention devices, the buffer memory unit 11 is of a ring type, and the data of the packet is not always well contained at the upper limit of the buffer memory unit 11. When an address used by another device is mapped adjacent to the upper limit of the located address, the second processing unit 24 is connected to the header detection unit 10 and the buffer memory unit 11 as shown in FIG. It can be interposed between them. The second processing unit 24 stores the packet in the buffer memory unit 11
If the upper limit is exceeded, an operation is performed to write the head of the packet from the head of the buffer memory unit 11.

That is, the second processing unit 24 obtains the current address of the buffer memory unit 11 from the counter unit 19, obtains the header length from the header of the packet from the header detection unit 10, and obtains the above address and data length. If the value obtained by adding the header length exceeds the size of the buffer memory unit 11, the counter unit 19 is initialized to the start address of the buffer memory unit 11, and the packet to be transferred is written from the start address of the buffer memory unit 11. I do.

By adding the second processing unit 24 in this way, the central processing unit 14
The data portion of each packet can be transferred to a video decoder or an audio decoder at a time by block transfer such as DMA transfer without worrying about the upper limit.

In each of the above embodiments, video and audio have been described as examples of data types. However, the present invention is not limited to this, and can be applied to any data type.

[0040]

As described above, according to the information storage management device of the present invention, the following excellent operational effects can be exhibited. According to the first aspect, the packet information and the like for each data type are read from the memory block for each data type allocated to the RAM of the central processing unit, and the packet is extracted from the buffer memory unit based on this. Processing can be performed for each data type. In addition, since the central processing unit can easily obtain the packet information, it is easy to create a program for the central processing unit and the processing can be speeded up. According to the second invention, the F storing the packet information and the like for each data type is stored.
By providing the IFO memory unit, packet information for each data type can be obtained without generating an interrupt to the central processing unit, and therefore, processing for each data type can be performed. Further, since it is not necessary to generate an interrupt as described above, the processing can be further speeded up. In the first and second inventions, by providing the second processing unit, when the packet exceeds the upper limit of the buffer memory unit, the head of the packet can be written from the head of the buffer memory unit. The unit can easily perform data block transfer and the like to a subsequent device without worrying about the boundary of the buffer memory unit, and can reduce the load and further increase the processing speed.

[Brief description of the drawings]

FIG. 1 is a packet configuration diagram showing the structure of a packet used in the present invention.

FIG. 2 is a diagram showing a stream when the packets shown in FIG. 1 are multiplexed.

FIG. 3 is a block diagram showing an information storage management device of the first invention.

4 is a diagram showing a storage state of a FIFO memory unit used in the device shown in FIG.

FIG. 5 is a diagram showing a storage state of a memory block in a central processing unit used in combination with the device shown in FIG. 3;

FIG. 6 is a block diagram showing an information storage management device according to a second invention.

7 is an explanatory diagram for explaining a storage state of a FIFO memory unit used in the device shown in FIG.

FIG. 8 is a diagram illustrating a connection state when the second processing unit is used.

[Explanation of symbols]

1 ... packet, 2 ... header area, 3 ... data area, 4 ...
Start code, 5: data type code, 6: data length code, 7: auxiliary information code, 9: information storage management device,
10: Header detection unit, 11: Buffer memory unit, 12:
Processing unit, 13: FIFO memory unit, 14: Central processing unit, 15: CPU body, 16: RAM, 16A, 16B
... Memory block, 17 ROM, 18 BUS line, 19 counter, 20, 21 FIFO memory (first-in first-out memory), 22, 23 register
24: second processing unit, S1: interrupt signal.

Claims (3)

(57) [Claims] 1. A packet group having data information including a data type and a data length as a header is written in a common buffer memory unit irrespective of the data type, and the packet group is controlled by a central processing unit from the buffer memory unit. An information storage management device that reads out a packet of a desired data type in response to the packet; a header detection unit that detects a header of the packet; a first-in first-out memory unit that stores packet information of the packet and a head address of the packet in the buffer memory unit And writing the packet information and the head address to the first-in first-out memory unit based on the information from the header detection unit, generating an interrupt to the central processing unit after the writing is completed, and generating a data type in the central processing unit. Depending on the data type in the memory block allocated for each And a processing unit for writing the address information and the head address. When a packet of a required data type is taken out, the packet is read out from a position indicated by a read pointer on the memory block corresponding to the data type and the buffer is read out. An information storage management device configured to obtain information on a packet in a memory unit. 2. A packet group having data information including a data type and a data length as a header is written in a common buffer memory unit irrespective of the data type, and is required by the central processing unit from the buffer memory unit. An information storage management device that reads a packet of a desired data type in response to the packet information; a header detection unit that detects a header of the packet; and a buffer memory unit that is provided for each packet type and that stores the packet information of the packet and the packet. A first-in first-out memory unit that stores the first address in the first and second memory units, and a processing unit that writes the packet information and the first address to the first-in first-out memory unit corresponding to the type of packet based on information from the header detection unit. When extracting a packet of the data type, the An information storage management device configured to read information from a first-in first-out memory unit and obtain information on a packet in the buffer memory unit. 3. A second process for writing a head of the packet from the head of the buffer memory unit between the header detection unit and the buffer memory unit when the packet exceeds the upper limit of the buffer memory unit. The information storage management device according to claim 1, further comprising a unit.