JPH1168881A - Data stream processor and its method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce power consumption by having a frequency calculating means which calculates an operation frequency that is supplied to an interfacing means based on transfer rate information from a transfer rate information extracting means and a signal generating means which generates a reference signal of an operation frequency that is calculated by the frequency calculating means and sends it to the interfacing means. SOLUTION: A demultiplexer 1 extracts bit rate information from an MPEG 2 stream, sends it to an optimum frequency calculation block 2 and also sends a data stream to DRAM I/F 5. The block 2 has a ROM table that preliminarily sets an optimum frequency which corresponds to a bit rate and calculates an optimum value for a clock that is supplied to the DRAM I/F 5 based on the bit rate information. A PLL 10 generates a clock of a frequency of the optimum value. Thus, it is possible to reduce power consumption by lowering the frequency of a clock that is supplied to the DRAM I/F 5 and SDRAM 7.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a data stream processing apparatus and method for processing a data stream that is variable-rate time-series information.

[0002]

2. Description of the Related Art MPEG2, which is a format of a color moving image compression system, is a common name of the ISO / IEC 13818 standard, which is a general-purpose encoding system that can be used for transmission media such as communication. This MPEG (Moving Pictures Expe
The rts Group 2) is expected to become popular because it has been adopted as a so-called digital video disc (DVD) for storing images on a disc of the same size as a compact disc having a diameter of 12 cm.

[0003] The moving picture data compression method of MPEG2 is as follows.
It is roughly divided into discrete cosine transform (DCT), which is compression using correlation within a screen, motion compensation, which is compression using correlation between screens, and Huffman coding, which is compression using correlation between code strings. It is a combination of the three.

[0004] DCT utilizes the spatial correlation of an image,
This is a method for removing redundancy in the spatial axis of data.
On the other hand, the motion compensation is a technique that uses the correlation in the time axis direction of an image to remove redundancy on the time axis of data. About 1/2 in compression by motion compensation processing, 1/10 to 1/20 in compression by DCT processing, 2/3 to 1/2 in compression by Huffman coding processing, and about 1 /
A compression ratio of 40 is obtained. If the data amount is reduced to about 1/40, for example, a 240 MBPS video signal becomes 6 MB
PS or 160MBPS video signal is 4MBPS
Can be transmitted.

In MPEG2, in order to perform predictive coding using motion compensation, as shown in FIG. 4, a GOP (Gr.Gr.) composed of three elements called an I picture, a P picture and a B picture is used.
oupOf Pictures) structure.

[0006] An I picture (Intra coded image) is created by intra-field coding and does not perform predictive coding from the previous image. If only images created using predictive coding are arranged, it is not possible to instantaneously display a screen in response to random access. Therefore, an access standard is created periodically so that random access can be supported. I
The picture exists because it has GOP independence.

The frequency of occurrence of I-pictures is determined by the performance of random access required for each application.
2 per frame), that is, one per 15 images. The data amount of one I picture is 2 times that of one P picture.
３3 times, that is, 5 to 6 times of one B picture. GOP
Is a group of pictures from one I picture to the next I picture. Therefore, image prediction is performed between pictures in this group.

A P picture (Predictive coded image) is
It is made by performing predictive coding from the previous image,
Created based on I-pictures. A P picture is defined as an "inter-frame quasi-prediction coded image" for an I picture that is an "intra-coded image".

B picture (Bidirectionally predictive
The “encoded image” is “bidirectional predictive encoded image”. B
A picture is created by performing prediction from two preceding and succeeding P pictures.

[0010] I picture, P picture, G of B picture
Looking at the relationship in the OP a little more in detail, as shown in FIG. 5, the first-stage prediction from the I-picture at the beginning of the GOP is performed in the forward or forward direction, and a P-picture is created.
At this time, the P pictures are arranged so as to skip a plurality of B pictures created later.

[0011] In the second stage of prediction, a plurality of B pictures are created between two frames of the first I picture and the P picture coded in the first stage, that is, by bidirectional prediction. Can be Furthermore, the first P picture and the second P picture
Similarly, a plurality of B pictures are created between pictures. In decoding a B picture, motion compensation is performed using two motion vectors and two (previous and subsequent) reference images. The bidirectional prediction, which is a feature of MPEG, has a feature that high prediction efficiency can be obtained because two temporally past images and future images are used in prediction.

[0012] In MPEG2, five types of "profiles" combining functions suitable for various applications and four types of "levels" indicating processing capability are defined in order to ensure compatibility between media. .

As the profile / level, FIG.
As shown in the above, 11 types are currently proposed. Of these profiles / levels, Main Profile / Main Level
l; MP @ ML) records DVD-Video
Has been adopted. This MP @ ML is expected to be used in many applications until the digital HDTV system becomes widespread in the future.

MP @ ML is a standard division corresponding to television video in the MPEG2 standard, and performs in-field chroma line thinning for a video signal. The high profile does not perform chroma line thinning, and is an excessive specification for general broadcasting.

The numbers in the row of levels in FIG. 1 represent the number of horizontal pixels × the number of vertical pixels × the standard value of the frame frequency. The “SNR” of the “SNR scalable” in the row of the profile is a signal noise ratio (Signal Noise Ratio).
Ratio).

[0016] The DVD-Video employs a variable transfer rate for high-efficiency image recording. Here, the transfer rate indicates the amount of data that can be transmitted within a unit time, and is expressed in bits per second (bp).
s). As the numerical value of the transfer rate is higher (the transfer rate is higher), a larger amount of video and audio data can be sent, so that image quality and sound quality can be improved.

When image compression is performed by MPEG2, NTS
A transfer rate of about 6 Mbps is required to obtain an image quality equivalent to that of a C-type television. However, in practice, a larger amount of data is required in a portion where the movement of video is intense, and a higher transfer rate must be used to maintain high image quality. On the other hand, a portion close to a still image with little video motion has a small data amount, and a sufficiently high image quality can be obtained even at a low transfer rate.

The problem here is the relationship between the transfer rate and the recording time. In order to maintain high image quality regardless of the state of the video, it is necessary to use a high transfer rate according to the part where the amount of video data is maximum, but transfer rate × recording time Therefore, the higher the transfer rate, the greater the waste generated when the recording time is short.

For this reason, DVD-Video employs a technology of a variable transfer rate of up to 9.8 Mbps. In this method, when a large amount of data is required, the transfer rate is reduced and the image is recorded / reproduced according to the state of the image, and the recording capacity of the disc can be used more efficiently and without waste. .

In the case of DVD-Video, by adopting this variable transfer rate, an image quality that requires a transfer rate of about 3.5 Mbps on average and a rate twice or more that of a fixed rate is realized. By the way, MPEG1
And compress the image with VideoCD, 1.15Mbps
, Which is one of the technical differences from DVD-Video.

[0021]

SUMMARY OF THE INVENTION MPEG @ MP @ M
L, the maximum video bit rate is typically 25M
Hz, and the bandwidth of access to the external memory is defined based on this value.

However, the video broadcasting and the D
In VD, it is a data stream with a variable rate of about 4 to 10 Mbps. For this reason, actually, there is a considerable amount of useless portion in the above bandwidth, and during that time, a clock is supplied, so that power consumption is large.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and has as its object to provide a data stream processing apparatus and method which reduce power consumption.

[0024]

In order to solve the above-mentioned problems, a data stream processing apparatus according to the present invention comprises a transfer rate information extracting means for extracting transfer rate information from a data stream which is variable-rate time-series information. An interface means for transmitting and receiving data to and from an external storage means using data provided from the transfer rate information extracting means to which the data stream is input, and a transfer rate information provided from the transfer rate information extracting means. Frequency calculating means for calculating an operating frequency to be supplied to the interface means, and a signal generating means for generating a reference signal of the operating frequency calculated by the frequency calculating means and supplying the reference signal to the interface means. .

Further, in the data stream processing method according to the present invention, a transfer rate information extracting step of extracting transfer rate information from a data stream which is variable-rate time-series information; An interface step of transmitting and receiving data to and from an external storage means using data provided from the step, and a frequency for calculating an operating frequency to be supplied to the interface step based on the transfer rate information provided from the transfer rate information extracting step A calculating step; and a signal generating step of generating a reference signal of the operating frequency calculated in the frequency calculating step and supplying the reference signal to the interface step.

[0026]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A data stream processing apparatus and method according to the present invention will be described below with reference to the drawings.

As shown in FIG. 1, the data stream processing apparatus according to the present invention comprises a demultiplexer 1 serving as transfer rate information extracting means for extracting a bit rate which is a transfer rate from an MPEG2 data stream which is variable rate time-series information. And an optimum frequency calculation block 2 as frequency calculation means for calculating a frequency based on the bit rate extracted by the demultiplexer 1.

The demultiplexer 1 is a header parser for analyzing a slice header in additional information indicating an attribute of MPEG2 which is a variable rate data stream and extracting bit rate information. This demultiplexer 1
Thereafter, the bit rate information is sent to the optimum frequency calculation block 2 and the data stream is sent to the DRAM interface unit.

The optimum frequency calculation block 2 is a block for calculating the optimum value of the clock supplied to the DRAM interface based on the bit rate information supplied from the demultiplexer. The optimum frequency calculation block 2 can be configured as a ROM table in which the optimum frequency corresponding to the bit rate is set in advance in a non-volatile memory such as a ROM.

As an example of the values set in this table, as shown in FIG.
bps is 30 for the original frequency of 25 MHz.
The original frequency is set to 20% for 1.5 to 4 Mbps corresponding to digital broadcasting, to 10% for 4 to 10 Mbps corresponding to DVD, and to normal for 10 to 25 Mbps. be able to.

Further, the data stream processing device includes:
A frequency counter 3 which is a counter for counting a frequency,
Voltage-controlled oscillator that generates a frequency corresponding to the voltage (Voltag
e Controlled Oscillator (VCO) 4.

The frequency counter counts the frequency of the signal supplied from the VCO 4, and receives, for example, a reference timing supplied from the outside and an optimum frequency from the optimum frequency calculation block 2. A voltage for controlling the VCO 4 is generated to control the VCO 4.

The VCO 4 is a voltage-controlled oscillator that generates a signal having a frequency corresponding to the voltage supplied from the frequency counter 3.

The frequency counter 3 and the VCO 4
Is a so-called PLL (Phase Locked Loop)
10). The PLL 10 generates a reference signal of the optimum frequency given from the optimum frequency calculation block 2 synchronized with the reference timing. The PLL 10 supplies the reference frequency to the DRAM interface unit 5 and the SDRAM.

Further, the data stream processing device includes a DRAM for mediating signal transmission to the SDRAM.
It has an interface unit 5 and an audio / visual decoding block 6 for decoding audio / visual data provided as a data stream.

The DRAM interface unit 5 receives the data stream from the demultiplexer 1 and sends and receives information to and from the SDRAM as an external storage unit in accordance with the timing of the reference signal supplied from the PLL 10. Specifically, data and an address are sent to the SDRAM in synchronization with the reference signal, and data is received from the SDRAM. Also, the DRAM interface unit 5 stores the compressed audio / visual data in the audio / visual decode block 6.
It provides visual data and receives decoded audio / visual information from the audio / visual decoding block 6.

The data stream processing device is
The demultiplexer 1, the optimal frequency calculation block 2,
The frequency counter 3, VCO 4, DRAM interface unit 5, and A / V decode block 6 can be configured as an MPEG2 decoder IC formed as an integrated circuit on a one-chip semiconductor substrate.

In the data stream processing apparatus, as an external storage means, a synchronous dynamic random access memory (SDRAM) having a high access speed and a wide bandwidth is used.
emory) 5 is connected. The SDRAM 5 stores audio / visual information provided from the data stream processing device via the DRAM interface unit 5 and reads out the audio / visual information by the DRAM interface unit 5.

More specifically, the SDRAM 5 has the D
The data and address are received from the RAM interface unit 5 and the data is provided to the DRAM interface unit 5. The SDRAM 5 has the PL
A reference signal is supplied from L10.

As described above, when decoding an MPEG2 data stream of usually about 10 MBPS,
When the DRAM interface unit 5 operating at the fastest frequency and the SDRAM 7 are used as external storage means, power consumption can be reduced by lowering the frequency of the clock supplied to the memory itself.

That is, the frequency of the clock supplied to the DRAM interface unit 5 by the built-in VCO 4 is
It is varied according to the bit rate of the data stream. This is different from the conventional clock in which the bit rate is fixed to the upper limit that can be taken. Therefore, in this data stream processing device, the power consumption is reduced by the amount of the delay of the clock.

Next, a series of steps according to the data stream processing method will be described with reference to a flowchart shown in FIG.

In the first step S1, a bit rate, which is a transfer rate, is extracted from an MPEG2 data stream, which is variable-rate time-series information. That is, MP
Since EG2 includes a slice header in the additional information indicating the attribute, the header parser analyzes the slice header to extract bit rate information. Then, the process proceeds to the next step S2.

In step S2, an optimum clock value is calculated based on the bit rate information. Here, in step S2, the optimum frequency corresponding to the bit rate is set in the ROM which is a non-volatile memory and set in advance in the ROM.
An OM table may be prepared, and an optimum frequency corresponding to a given bit rate may be read from the ROM table. Then, the process proceeds to step S3 following this.

In step S3, a reference signal having the optimum frequency is generated. For generating the reference signal, a PLL having a frequency counter and a VCO can be used.
Then, the process proceeds to step S4.

Step S4 is an interface process for transmitting and receiving the input variable-rate MPEG2 data stream to and from external storage means based on the reference signal. In step S4, audio / visual information is transmitted / received to / from an external storage unit based on a reference signal having a frequency optimally set according to the variable rate. Then, this series of steps ends.

As described above, in the data stream processing method, the reference signal of the interface is set to an optimum frequency according to the bit rate of the data stream. Therefore, unlike the related art, it is not necessary to always fix the frequency of the reference signal to the upper limit of the bit rate, and the frequency can be reduced, so that power consumption can be reduced.

In the above description, the frequency of the reference signal is set to an optimum frequency according to the bit rate of the data stream. However, the frequency of the reference signal is determined by the number of audio channels, the GOP structure as an image group, and the attribute of a picture. Fine control can be achieved by taking into account such factors. For example, during the decoding of an I-picture, a motion vector is not used, so that the fact that access is small can be used.

Further, if information is prefetched at the time of header parsing and feedforward control is performed, decoding does not break down.

Further, in the above description, SDRAM is exemplified as the external storage means, but the storage means is SDRA
It is not limited to M. As an external storage means, for example, D
RAM can be used.

[0051]

As described above, when decoding the MPEG2 data stream, the power consumption can be reduced by lowering the clock frequency of the DRAM interface according to the transfer rate of the data stream.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a schematic configuration of a data stream processing device.

FIG. 2 is a diagram showing an example of a table stored in a ROM table of an optimum frequency calculation block of the data stream processing device.

FIG. 3 is a flowchart showing a series of steps of a data stream processing method.

FIG. 4 is a diagram illustrating an example of an arrangement of each picture in a GOP of MPEG2.

FIG. 5 is a diagram illustrating MPEG2 forward and backward predictive encoding.

FIG. 6 is a diagram showing a list of profiles / levels.

[Explanation of symbols]

1 demultiplexer, 2 optimal frequency calculation block,
4VCO, 5 DRAM interface

Claims (3)

[Claims] 1. A transfer rate information extracting means for extracting transfer rate information from a data stream which is variable-rate time-series information, and externally using data provided from the transfer rate information extracting means to which the data stream is input. A storage means and an interface means for transmitting and receiving data; a frequency calculation means for calculating an operating frequency to be supplied to the interface means based on the transfer rate information given from the transfer rate information extraction means; A signal generating means for generating a reference signal having the calculated operating frequency and supplying the reference signal to the interface means. 2. The data stream processing device according to claim 1, wherein the data stream is based on the MPEG2 standard. 3. A transfer rate information extracting step for extracting transfer rate information from a data stream which is variable-rate time-series information, and externally using data given from the transfer rate information extracting step to which the data stream is input. An interface step of transmitting and receiving data to and from the storage means; a frequency calculating step of calculating an operating frequency to be supplied to the interface step based on the transfer rate information given from the transfer rate information extracting step; Generating a reference signal having the calculated operating frequency and supplying the reference signal to the interface step.