JPH10191331A - Method and device for encoding image data - Google Patents

Abstract

PROBLEM TO BE SOLVED: To switch data in a bit stream state without making a failure of a VBV buffer on a decoder side and to simplify a system structure, preventing an image quality from deteriorating. SOLUTION: An encoder 12 makes image data from a VTR 11 a GOP(group of pictures) in every plural pictures and performs compression encoding by utilizing the correlation property among pictures. A bit rate decoder 17 detects the bit rate of compression encoding data from the encoder 12, a stuffing bit generating means 19 sends a stuffing bit to an adder 14 based on the detected bit rate, and data quantity is made fixed length in two GOP units of the compression encoding data.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image data encoding method in which a plurality of pictures of input image data are grouped into a GOP (Group of Pictures) and compression-encoded using the correlation between pictures. Related to the device.

[0002]

2. Description of the Related Art In recent years, for example, an image signal or an audio signal is compressed by using a compression method of a predetermined standard to generate digital bit stream data, and these are multiplexed and transmitted as necessary. Is done.

For example, in a so-called video server for storing and transmitting image data, if a moving image signal is digitized as it is, the data amount becomes enormous. Therefore, the moving image signal is efficiently stored and transmitted with a limited storage capacity or transmission path. In order to do so, compression encoding of data is required.

A typical example of such a standard for moving picture compression is Moving Picture Coding (MPEG).
Experts Group) standards. This MP
EG stands for ISO / IEC JTC1 / SC29 (Intern
ational Organization for Standardization / Interna
nation Electrotechnical Commission, Joint Technica
l Committee 1 / Sub Committee 29: International Organization for Standardization /
International Electrotechnical Commission Joint Technical Committee 1 / Specialized Subcommittee 29)
Organization for Studying Video Compression Coding for Media Integration
Picture Experts Group), which is ISO11172 as the MPEG1 standard and ISO13818 as the MPEG2 standard. In these international standards, ISO11172-1 and ISO13818-1 have been adopted for multimedia multiplexing, and for video.
ISO11172-2 and ISO13818-2, and ISO111 for audio items
72-3 and ISO13818-3 are standardized respectively.

[0005] Here, ISO1 as an image compression coding standard is used.
According to 1172-2 or ISO13818-2, an image signal is compression-encoded in units of pictures (frames or fields) using correlation in the temporal and spatial directions of the image.

In the MPEG system, a virtual decoder model (STD: system target decoder) is defined, and a multiplexed system stream from the encoder is correctly written in the STD, that is, so as not to cause a buffer failure. Defined to be decrypted. The input / output model of the buffer on the virtual decoder side is a VBV (Video Buffering Verifier) model as ISO / IEC 11172-2 Annex C or ISO / IEC 13
818-2 The buffer of this VBV model is described in Annex C, and is referred to as a VBV buffer. Decoder V
The BV buffer size is indicated by “vbv_buffer_size” in the MPEG bit stream. The standard size is 1.75 Mbit in the case of MP @ ML (Main Profile at Main Level), for example.

FIG. 5 shows an example of the configuration of a conventional image data encoding system using the MPEG standard. In FIG. 5, image data as a baseband video signal from the VTR 101 is MPEG2
, And compression-encoded, and sent to the server 103 as a so-called MPEG bit stream. The server 103 accumulates the supplied compressed image data on a hard disk or the like, and supplies a bit stream of the image data to a plurality of decoders 104a to 104n as required. The baseband signals of a plurality of streams of image data decompressed by the decoders 104a to 104n are sent to the router 105 to be switched, sent to the encoder 106, subjected to MPEG2 compression coding, and processed as an MPEG bit stream. Retrieved and transmitted.

[0008]

By the way, as described above, MPEG1 (ISO11172-2) and MPEG2 (ISO13818)
The bit stream of image data compressed and coded based on the standard of -2) has a fixed rate from a few seconds to several minutes when viewed macroscopically, but 1 GOP (Group) which is a basic unit of the MPEG bit stream. Of Picture), it is variable length data.

That is, FIG. 6 shows a specific example of an MPEG bit stream. One GOP includes a so-called I picture (intra-frame coded image), a P picture (forward predicted coded image), and a B picture (bidirectional). (Predicted coded images) are arranged in a predetermined order, but the amount of data of each of I, P, and B pictures differs depending on whether the image is complicated or simple. The data length of each GOP is different.

Even if a plurality of such bit streams are used and, for example, the bit streams are switched to be unified, the VBV buffer in the decoder may eventually overflow or underflow and break down.

For this purpose, as shown in FIG. 5, decoders 104a to 104n are provided on the input side of the router 105 to decode the MPEG bit stream from the server 103 into a baseband video signal, and to output the It is necessary to provide an encoder 106 to encode the baseband video signal from the router 105 into an MPEG bit stream, which complicates the configuration and raises the cost. It is also the cause.

[0012] The present invention has been made in view of the above-mentioned circumstances, and does not cause a VBV buffer failure on the decoder side even if switching is performed with a compression-encoded bit stream, and a video server or router. It is an object of the present invention to provide a method and apparatus for encoding image data that can prevent image quality deterioration and cost increase when applied to.

[0013]

According to the present invention, in order to solve the above-mentioned problems, a plurality of pictures of input image data are used as a GOP (group of pictures) as a unit of encoding and the correlation between pictures is utilized. Compression-encoding, detecting the bit rate of the compression-encoded data, and based on the detected bit rate, the 2GOP of the compression-encoded data
It is characterized in that the length is fixed in units.

[0014] The fixed length means that a shortage of a fixed data amount in 2 GOP units is filled with stuffing bits.

In the compression encoding, a so-called MPEG is used.
(MPEG1 and MPEG2).

The bit stream 2 of the compression encoded data
By making the length fixed in GOP units, even if the bit stream is switched in units of 2 GOPs with this fixed length,
The VBV buffer on the decoder side does not break down.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of an image data encoding apparatus to which the image data encoding method according to the present invention is applied will be described with reference to the drawings.

FIG. 1 is a block diagram showing a system using an image data encoding apparatus according to an embodiment of the present invention.

In FIG. 1, a VTR 11 for reproducing image data of a program or CM as a baseband video signal
Is sent to the encoder 12 which performs compression processing by the above-described MPEG2, and is subjected to compression processing.
The data is sent to an adder 14 via an FO (First In First Out) memory 13, to which a stuffing bit described later is added, and sent to a video server 15. The video server 15 stores the MPEG bit stream, which is compressed image data, on a hard disk or the like, and supplies the MPEG bit stream of the compressed image data to a plurality of systems in response to a transmission schedule request. A plurality of MPEG bit streams from the server 15 are sent to a router 16 for switching and extraction.

The bit rate of the MPEG bit stream from the encoder 12 is detected by a bit rate decoder 17. The encoder 12 encodes the video signal so as to have the target bit rate. However, since the video signal does not always have a fixed length, an error from the target is detected by the bit rate decoder 17 and the error amount is determined by a bit rate control circuit (bit). The data is fed back to the encoder 12 via a rate controller 18. In the bit rate decoder 17, the next GOP (Group Of Pictur)
The data amount of e) is also detected and sent to the stuffing bit generation circuit 19.
At 9, a stuffing bit is generated to correct the error and sent to the adder 14. At this time, 2 GOP
From the memory 13 so as to have a fixed length in units of
A stuffing bit is added to the PEG bit stream. The stuffing bits are dummy data, such as 0, filled to make the data amount constant in units of 2 GOPs, and the decoder ignores or removes the stuffing bits and performs decoding processing.

That is, in the embodiment of the present invention, the encoder 12 uses a plurality of pictures (frames or fields) of input image data as GOPs (groups of pictures) as basic units of encoding to determine the correlation between pictures. MPE using compression encoding using compression encoding
The bit rate of the data of the G bit stream is detected by the bit rate decoder 17, and based on the detected bit rate, the stuffing bit generation circuit 19 and the adder 14 execute the 2GOP of the compression encoded MPEG bit stream data. The unit is fixed length.

FIG. 2 shows a specific example of an MPEG bit stream to which stuffing bits SB are added. As shown in FIG. 2, the stuffing bits SB are fixed in length by adding stuffing bits SB in units of 2 GOPs. The MPEG bit stream is sent from the adder 14 of FIG.

The server 15 stores the bit stream fixed in length in units of 2 GOPs. The bit stream reproduced by the server 15 is sent to the router 16 in accordance with the transmission schedule, and the router 16 switches the bit stream in units of the fixed length 2 GOP. The switching of the bit stream in units of 2 GOPs may be performed in the server 15 without being performed in the router 16.

As a specific example of the encoder 12, an encoder that performs compression encoding in accordance with the MPEG1 or MPEG2 standard as described above can be used.
This MPEG encoder uses a coding unit as a macroblock, and performs motion compensation, DCT (discrete cosine transform),
Compression coding is performed by combining quantization and variable length coding. A GOP is formed by grouping a plurality of, for example, 15 screens (pictures), and an I picture (intra-frame code) is set as a picture type. Image), P
A picture (forward coded image) and a B picture (bidirectional coded image) are used.

An MPE for decoding an MPEG bit stream obtained from such an MPEG encoder
An example of the G decoder will be described with reference to FIG.

In FIG. 3, the MPEG bit stream is stored in the VBV buffer 41,
The data is sent to a motion compensation circuit 45 via an LC (inverse variable length coding: variable length decoding) circuit 42, an inverse quantization circuit 43, and an inverse DCT circuit 44, and is extracted as a baseband video signal.

In the VBV buffer 41, the GOP
The data of each picture of I picture, P picture, and B picture is stored, and at a fixed time interval in synchronism with the video output timing, MPEG data is read out at a different readout amount for each picture.
Read the data of the bit stream. In the inverse VLC circuit 42 and the inverse quantization circuit 43, the inverse processing of quantization and variable length encoding on the encoder side is performed, and the inverse DCT circuit 44 returns the frequency axis data to the time axis data, and the motion compensation circuit 45 Sent to. The motion compensation circuit 45 uses the frame memories 46 and 47 to correct the motion generated between frames, and performs motion compensation inter-frame prediction.

FIG. 4 shows such an MPEG decoder.
2 shows a specific example of an MPEG bit stream stored in a BV buffer, and shows an example of a change in the data amount between the start point of a GOP and a point after 2 GOPs.

In FIG. 4, the slope of the bold line is the MPEG
Indicates the input rate of the bit stream. In the case of a fixed bit rate, the slope is constant. Since the data of each picture is extracted from the VBV buffer when decoding each picture of I, B, and P, the remaining amount of accumulated data is reduced. This extraction of data is ideally performed instantaneously. In general, the data amount of an I picture is the largest,
It decreases in the order of pictures. The size of the VBV buffer (VBV size) is, for example, MP @ M of MPEG2.
1.75 M for L (Main Profile at Main Level)
It is specified as bit. FIG. 4A shows a case where the difference ΔB of the remaining data amount between the start point of the GOP and the point after the lapse of 2 GOPs is negative (ΔB <0), and FIG.
B> 0), and FIG. 4C shows that the difference ΔB is 0.
(ΔB = 0) is shown. As described above, since the difference ΔB in the remaining data amount between two GOPs changes, when such an MPEG bit stream is switched, the VBV buffer may overflow or underflow. Buffer remaining is VBV
A state where the size has been reached is called an overflow state, and the input data is discarded, so that an image is skipped or noise is generated. Conversely, a state in which the buffer remaining amount becomes 0 is called an underflow state, and the image is frozen because the previous picture is continuously output until valid data for one picture is input.

Therefore, in the embodiment of the present invention,
2 GOP is fixed length as one unit.
By switching the MPEG bit stream in GOP units, it is possible to prevent a failure such as overflow or underflow of the VBV buffer from occurring. That is, the stuffing bit is added as the data amount of 2 GOPs so that the remaining data amount at the start point and the remaining data amount at the end point of the 2 GOP in the VBV buffer are equal. The data amount of 2 GOPs with the stuffing bits added is equal to the 2 GOP period (3
(During 0 frames).

Here, it is conceivable to add stuffing bits in units of one GOP to make the length fixed, but in terms of data use efficiency, the unit of two GOPs is more preferable. This is preferable in that the data amount can be more efficiently controlled by feedback. Also, 2G
It is conceivable that the length is fixed in units of several GOPs larger than the OP, but the unit of switching the bit stream in the router 16 is long, and the editing efficiency and the switching waiting time are long. On the other hand, the 2GOP is about 1 second in time, and is a practically appropriate length in consideration of the fact that the switching unit of general video is often about 1 second.

It should be noted that the present invention is not limited to only the above-described embodiment, and for example, the router 16 shown in FIG.
In addition to transmitting the MPEG bit stream directly from the decoder, the MPEG bit stream may be decoded by a decoder, converted into a baseband video signal, and then transmitted.

[0033]

According to the method and apparatus for encoding image data according to the present invention, a plurality of pictures of input image data are used as a GOP (group of pictures) as a unit of encoding by utilizing the correlation between pictures. Compression encoding, detecting a bit rate of the compression encoded data, and performing 2GOP of the compression encoded data based on the detected bit rate.
Since the length is fixed in units, even if the bit stream is switched in units of 2 GOPs, the VBV buffer on the decoder side does not fail, and the image quality deteriorates due to the failure of the VBV buffer, that is, noise appears on the decoded output screen. And the disadvantage that the time axis is skipped or the image freezes can be prevented beforehand. In addition, since switching can be performed without changing the bit stream, it is not necessary to decode and convert to a baseband video signal in advance when sending to a server or a router to perform switching, and it is necessary to encode after switching. And the structure is simplified, and the cost can be reduced.

Further, by setting the unit of fixed length to 2 GOPs, the unit of switching becomes an appropriate edit unit time of about 1 second, the data amount can be controlled by feedback to the encoding operation, and the data use efficiency can be reduced. In terms of point, waste can be reduced.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a schematic configuration of a system using an image data encoding device according to an embodiment of the present invention.

FIG. 2 is a diagram illustrating an MPEG bit stream obtained by an image data encoding device according to an embodiment.

FIG. 3 is a block diagram illustrating a basic configuration example of an MPEG decoder.

FIG. 4 is a diagram illustrating a buffer remaining amount when 2 GOP data of an MPEG bit stream is supplied to a VBV buffer.

FIG. 5 is a block diagram showing a configuration of a system using a conventional image data encoding device.

FIG. 6 is a diagram showing a conventional MPEG bit stream.

[Explanation of symbols]

12 encoders, 13 memories, 14 adders,
15 server, 16 router, 17 bit rate decoder, 18 bit rate control circuit, 19
Stuffing bit generator

Claims (6)

[Claims] An input image data is stored in a GO for each of a plurality of pictures.
A step of compressing and encoding the data into a group of pictures (P) using correlation between pictures; a step of detecting a bit rate of the compressed and encoded data; and a step of performing the compression based on the detected bit rate. Fixedly lengthening the data amount in units of 2 GOPs of encoded data. 2. The image data encoding method according to claim 1, wherein said fixed length is performed by adding a stuffing bit. 3. The image data encoding method according to claim 1, wherein an MPEG standard encoding is used for said compression encoding. 4. The method according to claim 1, wherein the input image data is GO
An encoder for compressing and encoding the data as a P (group of pictures) using correlation between pictures; a bit rate detecting means for detecting a bit rate of the compressed and encoded data; An image data encoding device, comprising: fixed length means for fixedly lengthening a data amount in units of 2 GOPs of compressed encoded data. 5. The image data encoding apparatus according to claim 4, wherein said fixed length lengthening means adds a stuffing bit to said compressed and coded data so as to have a constant data amount in units of 2 GOPs. 6. The image data encoding apparatus according to claim 4, wherein an MPEG encoder is used as said encoder.