CN101835055B

CN101835055B - Digital video signal record and playback device and record and playback method thereof

Info

Publication number: CN101835055B
Application number: CN2009101749172A
Authority: CN
Inventors: 三嵨英俊; 浅村吉范; 幡野喜子; 外田修司; 仓桥聪司; 中井隆洋; 加濑泽正; 长泽雅人; 大畑博行; 石田祯宣
Original assignee: Mitsubishi Electric Corp
Current assignee: Mitsubishi Electric Corp
Priority date: 1994-09-26
Filing date: 1995-09-25
Publication date: 2012-12-12
Anticipated expiration: 2015-09-25
Also published as: GB2293516A; CN1543209A; CN1158865C; CN1138273A; CN1487738A; GB9519553D0; CN101835055A; CN100348029C; GB2293516B; CN1487738B; DE19537508A1; DE19537508C2

Abstract

A digital video signal record and playback device for recording on a recording medium a digital video signal coded by using a motion compensation prediction and an orthogonal transform and playing back data from the recording medium. In a data arrangement of a digital video signal, an I picture which can be independently represented in a picture is divided into n areas in the vertical direction, and the data is arranged from the front of one GOP in the unit of area by giving a priority to an area located at the center of the screen. A playback picture is outputted by playing back the I picture read from the recording medium in the unit of area unit. In the case where the whole I picture area cannot be read within a definite time, the area which cannot be read are interpolated by the use of the data of the preceding screen.

Description

The method of the record of digital video signal and reproducing device and record thereof and playback

Dividing an application of the application's one Chinese patent application of the same name 200310120713.3 that to be applicant Mitsubishi Electric Corporation submit to September 25 nineteen ninety-five.

Technical field

The present invention relates to write down and the record and the reproducing device of the digital video signal of the digital video signal of resetting; Specifically, relate to and be used to write down and reset such as the record and the reproducing device of the digital video signal of the digital video signal of encoding according to moving compensation prediction and orthogonal transform on the medium of CD.

Background technology

Fig. 1 opens the record of the conventional CD shown in flat 4-114369 (1992) number and the circuit block diagram of reproducing device for japanese patent application laid one.Referring to Fig. 1, wherein label 201 expressions are used for vision signal, audio signal etc. is converted to A/D (mould/number) transducer of digital information.Label 202 expression data compression circuits; 203 expressions are used for packed data is converted to the frame sector change-over circuit of the sector data that equals a frame period integral multiple; 204 expressions are used for error correction signal is added to the error correcting encoder of sector data; 205 expressions are used for the interference signal between each coding of recording medium is modulated into the modulator of predetermined modulation code to reduce to disturb, and 206 represent to come according to modulation code the laser drive circuit of modulated laser, 207 expression laser output switches.In addition; The shaven head of label 208 expression emission laser, 209 expressions make from the laser beam of shaven head 208 actuator by track scanning, and 210 expressions are used to reciprocating motor that shaven head 208 is stretched out; 211 expressions make the coil motor of CD 212 rotations; 219 expression motor-drive circuits, 220 expression first control circuits, 221 then represent second control circuit.In addition; Label 213 expressions are used to amplify the playback amplifier from the replay signal of shaven head 208; 214 expressions are used for obtaining from the modulation signal that has write down the demodulator of data, 215 expression error correction decoders, 216 expression frame sector reverse conversion circuit; 217 expressions make packed data expanded data expanded circuit, and 218 expressions make expanded data convert for example D/A (D/A) transducer of analog video signal and audio signal to.

Fig. 2 is a block diagram, and the internal structure of data compression circuit 202 among Fig. 1 is shown.Among Fig. 2, the digital video signal of importing from A/D converter 201 is passed to a memory circuit 301.Deliver to from the vision signal 321 of memory circuit 301 output on second input of first input end and moving compensation prediction circuit 310 of subtracter 302.Output footpath one DCT (discrete cosine transform) circuit 303 of subtracter 302 is passed to a quantizer 304.The output of quantizer 304 is delivered on the input of transmission buffer 306 through variable length coder 305.The output of transmission buffer 306 is passed to frame sector change-over circuit 203.Meanwhile, the output of quantizer 304 is passed to reverse DCT circuit 308 through an inverse quantizer 307.The output of reverse DCT circuit 308 is delivered on the first input end of adder 309.The output 322 of adder 309 is delivered on the first input end of moving compensation prediction circuit 310.323 of the outputs of moving compensation prediction circuit 310 are delivered on second input of second input and subtracter 302 of adder 309.

Fig. 3 is a block diagram, and the internal structure of moving compensation prediction circuit 310 among Fig. 2 is shown.In Fig. 3, input terminal 401a is delivered in the output 322 of adder 309, and the output 321 of memory circuit 301 is delivered on the input 401b.Be input to frame memory 404a or the frame memory 404b through a switch 403 from the signal 322 of input terminal 401a input.Deliver on the first input end of moving vector testing circuit 405a from the reference picture of frame memory 404a output.Vision signal 321 by input terminal 401b input is input on second input of moving vector testing circuit 405a.Predictive mode selector 406 is delivered in the output of moving vector testing circuit 405a.Simultaneously, deliver on the first input end of moving vector testing circuit 405b from the reference picture of frame memory 404b output.Deliver on second input of moving vector testing circuit 405b from the vision signal 321 of input terminal 401b input.The output of moving vector testing circuit 405b is delivered on second input of predictive mode selector 406.Deliver on the 3rd input of predictive mode selector 406 from the vision signal 321 of input terminal 401b input.A zero-signal is delivered on second input of switch 407.Second output of predictive mode selector 406 is received on the 3rd input of switch 407.323 of the outputs of switch 407 are from lead-out terminal 402 outputs.

Fig. 4 is a block diagram, and the internal structure of data expanded circuit 217 among Fig. 1 is shown.In Fig. 4, the vision signal of reverse conversion circuit 216 inputs is passed to reception buffer 501 from the frame sector.The output of reception buffer 501 is passed to variable-length decoder 502, and the output of this decoder then is reversed quantification at inverse quantizer 503 places.Then, its output receives inverse discrete cosine transform at reverse DCT circuit 504 places.The first input end of adder 506 is delivered in output after the conversion.Meanwhile, prediction data decoding circuit 505 is delivered in the output of reception buffer 501, and the output of prediction data decoding circuit 505 is then delivered on second input of adder 506.The output of adder 506 is transported to D/A converter 218 through memory circuit 507.

The working condition of this equipment below is described.As a kind of high efficient coding mode in the video signal coding category, a kind of encryption algorithm by MPEG (active images are efficiently organized into groups) mode is arranged.This is a kind of hybrid coding mode, and it will utilize the inter prediction encoding of moving compensation prediction and frame internal conversion coding to combine use.Example commonly used of this respect is to utilize the data compression circuit 202 that structure shown in Figure 2 is arranged, and has adopted above-mentioned MPEG mode simultaneously.

Fig. 5 shows a kind of data layout structure (layer structure) of MPEG mode of simplification.In Fig. 5, label 621 expressions one sequential layer resembles (being referred to as " GOP " later on) comprising a picture group that contains a plurality of frame data clauses and subclauses.622 expressions contain the GOP layer of several images (screen), and 623 expressions are divided into several section with a screen, and 624 expressions have the lamella of several macro blocks, and 625 represent macroblock layer, the piece layer that 626 expressions are made up of 8 pixels * 8 pixels.

This macroblock layer 625 is pieces being made up of by for example MPEG mode the unit of 8 pixels * 8 pixels at least.This piece is in order to accomplish the unit of DCT.At this moment, whole 6 pieces promptly comprise four adjacent Y-signal pieces, corresponding to a Cb piece and a Cr piece of this Y-signal piece on the throne, are called macro block.A plurality of such macro blocks have been formed a section.In addition, several macro blocks has been formed a minimum unit of moving compensation prediction, and the moving vector that supplies moving compensation prediction to use is formed in some macroblock units.

Then the inter prediction encoding process will be described.Fig. 6 illustrates the overview of inter prediction encoding.Image is divided into three types, i.e. intraframe coding image (being called the I image later on), single directional prediction coding image (being called the P image later on) and bi-directional predictive coding image (being called the B image later on).

For example; Suppose that a images in the N images is the I image, the images in the M images is P image or I image, and then the (N * n+M) images constitutes width of cloth I image; The (N * n+M * m) images (m ≠ 1) constitutes width of cloth P image; ((N * n+M * m+M-1) each image of images has constituted the B image to the images to the of N * n+M * m+1), and wherein n and m are integer, and 1≤m≤N/M from the.At this moment, (images to the of N * n+1) (summarize and be called a GOP (resembling group) by each image of images of N * n+N) from the.

Fig. 6 illustrates a kind of situation of N=15 and M=3, and among Fig. 6, the I image does not stand inter prediction and only encodes through the frame internal conversion.P image is through near the image of the I before the P image or from the prediction of P image.For example, the 6th images among Fig. 6 is the P image.The 6th images is through the prediction from the 3rd width of cloth I image.Also have, the 9th width of cloth P image among Fig. 6 is through the prediction from the 6th width of cloth P image.B image process is from the prediction near I image before and after the B image or P image.For example, the 4th and the 5th width of cloth B image not only passes through from the 3rd width of cloth I image but also passes through the prediction from the 6th width of cloth P image among Fig. 6.Therefore, the 4th and the 5th images is through the coding after the 6th images coding.

Below, according to Fig. 2 the processing of the data that comprise circuit 202 is described.Memory circuit 301 output is rearranged behind the signal and the digital video signal of input by the coding order.In other words, for example by the above, the 3rd width of cloth I image is afterwards to the 1st width of cloth B picture coding in Fig. 6.Therefore, the order of image will be rearranged.Fig. 7 illustrates this operation of rearranging.Press the image of the input of order shown in Fig. 7 A and at this moment press the order output shown in Fig. 7 B.

And then; Vision signal 321 by memory circuit 301 outputs; The prediction of output image 323 of process and next automatic compensation prediction circuit 310 is done subtracting each other with after the redundancy on the minimizing time-axis direction between the image on subtracter 302, carries out the DCT on the spatial axes direction.Conversion coefficient is quantized, carry out the coding of variable-length, again through transmission buffer 306 outputs.Simultaneously, the conversion coefficient after the quantification is reversed and quantizes and carry out reverse DCT.Afterwards, this coefficient is added on the adder 309 and predictive image 323 additions, so obtain one decoded visual 322.Be input to moving compensation prediction circuit 310 with decoded visual 322, for later on to the usefulness of picture coding.

By Fig. 3 the work of moving compensation prediction circuit 310 is described below.The preparatory circuit 310 of moving compensation has utilized two width of cloth reference pictures that are stored among frame memory 404a and the 404b, to accomplish the moving compensation prediction from the vision signal 321 of memory circuit 301, so that export a predictive image 323.

Beginning when the image 322 of Code And Decode as above-mentioned or I image or P are visual, is stored in it among frame memory 404a or the frame memory 404b, so that later image is encoded.At this moment, diverter switch 403, thus in two frame memory 404a and 404b, select a frame memory, this memory has precedence over another in time and is updated.But, when decoded visual 322 are the B image, then in frame memory 404a and 404b, do not write.

For example, when the 1st among Fig. 7 and the 2nd images were encoded with this switching mode of switch 403, the 0th width of cloth P image just was stored among frame memory 404a and the 404b respectively with the 3rd width of cloth I image.In addition, when the 6th width of cloth P image was encoded and decodes, the decoded image of the 6th width of cloth P image just write among the frame memory 404a again.

Then, when the 4th and the 5th width of cloth B image was encoded, the 6th width of cloth P image and the 3rd width of cloth I image deposited in respectively among frame memory 404a and the 404b.And then when the 9th width of cloth P image was encoded and decodes, the decoded image of the 9th width of cloth P image just write among the frame memory 404b again.So when the 7th width of cloth B image and the 8th width of cloth B image were encoded, visual and the 9th width of cloth P image of the 6th width of cloth P was stored into respectively among frame memory 404a and the 404b.

When the vision signal 321 from memory circuit 301 was passed in the moving compensation prediction circuit 310, moving vector testing circuit 405a and 405b detected a moving vector according to the reference picture that is stored among frame memory 404a and the 404b, and exported a moving compensation prediction image.In other words, vision signal 321 is divided into a plurality of.Then, select a piece, make it minimum for the distortion of predicting in its reference picture of each piece.Then, the relative position of this piece is exported as motion block, and this piece itself is exported as moving compensation prediction image simultaneously.

Meanwhile, the predictive mode selector is always selected on average image of the minimum image of a width of cloth predicted distortion or the one of which width of cloth in the moving compensation prediction image of two width of cloth of vector detection circuit 405a and 405b automatically.Then this image of selecting is exported as the image of having predicted.At this moment, if vision signal 321 is not a B image, then always selects and export this moving compensation prediction image, this image is corresponding to the reference picture by time input before another images.And then predictive mode selector 406 makes the coding of selection that better code efficiency arranged according to selected predictive image, selects wherein to predict the coding in the still uncompleted image, or selects predictive coding.

At this moment, when vision signal 321 is the I image, always select the coding in the image.When the coding in the selection image, representative is exported as predictive mode with a signal of image pattern coding.Meanwhile, when the predictive coding between the selection image, represent a signal of selected predictive image to export as predictive mode.When the predictive mode of predictive mode selector 406 output is the coding mode the image, zero-signal of switch 407 outputs.When if predictive mode is not the coding mode in the image, then 406 predictions of output of predictive mode selector are visual.

Can be known that by above-mentioned situation when the vision signal 321 from memory circuit 301 output be the I image, moving compensation prediction circuit 310 was always exported zero-signal as predictive image 323, the I image is without inter prediction but receive the frame internal conversion and encode.Simultaneously,, move compensation prediction circuit 310 and carry out moving compensation prediction from the 3rd width of cloth I image among Fig. 6 when the vision signal from memory circuit 301 outputs is when the 6th width of cloth P is visual Fig. 6, and prediction of output image 323.And then when the vision signal 321 from memory circuit 301 output was the 4th width of cloth B image shown in Figure 6, then moving compensation prediction circuit 310 was carried out the moving compensation prediction from the 3rd width of cloth I image shown in Figure 6 and the 6th width of cloth P image, and the prediction of output visual 323.

The working condition of transmission buffer 306 below will be described.Transmission buffer 306 will be converted to the bit stream of MPEG vision signal by the video data that variable length coder 305 carries out variable length code.Among this paper, this mpeg stream has six layers of structure, and is as shown in Figure 5.For sequential layer 621, GOP layer 622, image layer 623, lamella 624 and piece layer 626 add header information as identification code, to form a layer structure.

Then, transmission buffer 306 is reduced into a plurality of packets of information respectively with the bit stream of vision signal and the bit stream of audio signal, makes these packets of information multiplexizations and comprises a synchronizing signal, thereby form a MPEG2-PS system flow (program flow).Among this paper, MPEG2-PS is made up of program covering and packet level, and is as shown in Figure 8.Then, header information is added in packet level and the program covering.In common instance, system flow should be included a GOP partial data of video data.

Here, the program covering has such structure, and wherein the upper strata of packet level " is tied " together.Each packet level of configuration program covering is called as the PES packets of information.In addition, the header information of program covering shown in Figure 8 also comprises a program package identification signal and a synchronizing signal that constitutes vision signal and audio signal basis.

Simultaneously, in the packets of information of configuration information covering, there are three kinds of PES packets of information as shown in Figure 9.Here; Second-level message bag shown in Fig. 9 is last is " video/audio/special use a 1 " packets of information, wherein be used for the anterior of identifying information bag and time tag information or similarly when decoding as each packets of information of header information a coding of required part (PTS and DTS) be added to before the packet data.But time tag information PTS is a time control information that reappears output, and is the information of the decoding order of a data flow of when reappearing, controlling each packets of information.And DTS is the time control information when decoding beginning, and is the information of control decoded data order of transmission.

Third level packets of information shown in Fig. 9 is last is " special-purpose a 2 " packets of information, has wherein write the special-purpose data of user.Bottom one-level packets of information is " filling " packets of information, and wherein all packet datas are all put on " 1 ".The header information of " special-purpose 2 " packets of information and " filling " packets of information is that startup sign indicating number and the packet length information by packets of information constitutes.

As stated, video data and voice data are the system flow that converts MPEG2-PS through transmission buffer 306 to, and this conversion is to each carries out in the frame vector.This information via correction process, simultaneously, this information is modulated onto in the interference between each yard on the CD minimum, and is recorded on the CD 212.At this moment, for example, set the data volume of each GOP unit for the equal amount that is roughly.So clearly, through data allocations just being accomplished the editing of each GOP unit to each sector that equals the frame period integral multiple.

The work of playback time below will be described.Playback time is recorded in the video information on the CD 212 with playback amplifier 213 amplifications.Then, with error correction decoder 215 places information reverting is become numerical data at demodulator 214, then reverse conversion circuit 216 places revert to pure former video data again in the frame sector, and do not contain the for example data of address and parity.Entering data into it subsequently forms in the data expanded circuit 217 like Fig. 4.The system flow of being made up of MPEG2-PS is input on the transmission buffer 501.

At transmission buffer 501 places, the system flow of input is reduced into a program package unit.Then, according to each PES packets of information of header information reduction, thus the video data that reduces in this PES packets of information unit of recombinating and the bit stream of voice data.And then, relating to video data, this bit stream is reduced into piece layer shown in Figure 5, thus reduction and IOB data and moving vector data.

Transport to variable-length decoder 502 from the blocks of data of transmission buffer 501 output, make the data of variable-length become the data of regular length, then the amount of being reversed and through reverse DCT outputs on the adder 506 again.Meanwhile, prediction data decoding circuit 505, is passed to it on adder 506 decoding from the predictive image of transmission buffer 501 according to moving vector then.

In this case, prediction data decoding circuit 505 is the same with moving compensation prediction circuit 310, and a frame memory is provided, and is used to store I image and the P pictorial data by adder 506 decodings.By the way,, no longer describe, because this method is identical with method to this digital coding about upgrading the method for reference picture data.

Adder 506 is added up the output of prediction data decoding circuit 505 with the output of reverse DCT circuit 504, transport to memory circuit 307 then.Here, when coded data, rearrange in the frame continuous video information in time according to digital coding order shown in Figure 7.Therefore in memory circuit 507, rearranged data, made pictorial data continuous in time, and then it is passed in the D/A converter 218 with the input of Fig. 7 B order.

Below, to the situation on the CD search of image and the process of high-speed replay thereof are described with the data record of this coding structure.When coding structure is as shown in Figure 6, can accomplish the high-speed replay of image, at this moment data are reset in the I image cell.In this case, the I image can carry out the jump of track after resetting immediately.Then, GOP subsequently or previous is visited, so the I image obtains resetting there.For situation shown in Figure 6, can realize high-speed forward playback and reverse playback through repeating this process.

But because this GOP speed is a variable bit rate, where the front portion that therefore can not distinguish GOP subsequently fully is positioned at.So, shaven head is suitably jumped on the anterior position of GOP.Like this, just can not confirm that which bar track should be visited.

In addition, the I image has lot of data, therefore, when having only the I image with the continuation mode playback time, a kind of playback time of special use for example, image can not be reset with general activity figure 30Hz frequency that way, because be limited to the reading speed of dish.Even shaven head jumps later in the playback of accomplishing the I image, because it is longer to jump to the gap of next I image again, its work also lacks stationarity.

The record and the reproducing device of conventional digital video signal constitute in a manner described.Carrying out under the situation of high-speed replay with I image and several times of speed of P image, I pictorial data and P pictorial data be in the front portion that from the bit stream of recording medium etc. (for example CD etc.) record, detects GOP afterwards quilt read.Therefore; When data volume visual as I and the P image is very big; Perhaps when the removal search GOP that takes long to anterior, it is just not enough to be used for from the recording medium the time of reading of data, so thereby the problem that I image and the visual total data of P can not realize high-speed replay can occur reading.

In the record and reproducing device of the digital video signal of routine,,, also to spend a lot of times to go to import the I pictorial data if it has lot of data even only when carrying out high-speed replay with the I image.Therefore, the special use that surpasses tens of times is reset and can not be realized.One width of cloth I image that at this moment, can supply several GOP to use through only resetting is realized the special use playback of fair speed.Problem is that the gap of upgrading the playback image will increase, thereby the content of image just thickens.

Because the record and the reproducing device of conventional vision signal are encoded in a manner described, so it is decoded when jumping over search (through fast speed playback observed data) only to have an I image of mass data.Therefore, shaven head is jumped and do not reset and be enough to the data of supply decoding.Otherwise when the data of playback q.s, the playback duration of data will elongate, and the position that GOP will jump to just must be set in place quite far away, thereby produce the problem that the number of pictures that outputs on the screen reduces.

In addition, because the sevtor address of subsequently GOP can not be identified because of variable bit rate, whether be positioned on the track that " jump " reached so be difficult to differentiate the front portion of GOP.Therefore bring such problem, promptly, need the front portion of GOP be navigated on the target trajectory for many times CD rotation, then when special use is reset, output on the screen the picture data order very little.Have again; Also bring such problem,, also do not have what way to judge that the data of much scopes when shaven head jumps can be reset even the sector can be identified; Thereby before not through Video Decoder, can not make judgement, so the effect that shaven head jumps has just reduced.

Record and reproducing device as the digital video signal of some other routine proposed some equipment, were proposed in for example japanese patent application laid-Kai Ping 6-98314 (1994) number, 6-78289 (1994) number etc.Figure 10 illustrates an example.Wherein label 775 is flying-spot video generator, for example gamma camera, VTR etc.; 776 is audio signal generator, for example microphone, VTR etc.; 762 is video coder; 763 is audio signal encoder; 777 is the system layer bit stream generator; 778 is error correcting encoder; 779 is digital modulator; 780 is CD; 756 is playback amplifier; 786 is detector; 781 is digital demodulator; 758 is error-corrector; 759 is the system flow processor; 782 is video signal decoder; 783 is audio signal decoder; 784 is monitor; 785 is loud speaker.

At present, used usually its diameter of CD is for being 120mm, and this CD generally can write down 600M byte or more data.Recently, this type CD can reach 74 minutes with data transfer rate 1.2Mbps left and right sides recorded video signal and audio signal.During data record, vision signal is input to video coder 762 with to this encoding video signal from flying-spot video generator 775.From audio signal generator 776 audio signal is input to audio signal encoder 763 with to this coding audio signal.Multipath conversion such as header are become the processing procedure of these two code signals undertaken by system layer bit stream generator 777.Error correcting encoder 778 with the error correcting code affix after, carry out digital modulation by 779 pairs of these error correction signals of digital modulator, produce to supply recording bit stream.Make the stamper of tape recording device (not shown) by this bit stream, again with the content replication of stamper to CD 780, so just obtained the floppy disk sold on the market.

In the reproducing device that supplies the user, amplify with 756 pairs of information of playback amplifier from the video floppy disk of CD, again replay signal is transported to demodulator 786.This attention signal after detecting, carries out digital demodulation by digital demodulator 781 to it in detector 786, correct a mistake with error-corrector 758 again.Afterwards, from being taken out video signal portions the signal of error correction, the data of taking-up are decoded in video signal decoder 782, then with the audio signal with audio signal decoder 783 decodings, transport to respectively in monitor 784 and the loud speaker 785.

Typical method to this video signal coding is MPEG1 and the MPEG2 that is called MPEG (the special-purpose marshalling of active images) method, and this is a kind of compiling method of international standard.The instantiation of coding method can use the example of MPEG2 to explain.

Figure 11 illustrates the block diagram of video signal coding part in record and the reproducing device of conventional digital signal, in order to the coding method of explanation MPEG2.Figure 12 is the block diagram of video signal decoding unit in record and the reproducing device of conventional digital signal, in order to the explanation coding/decoding method.Figure 13 has represented the notion that the used active images of video signal coding are handled in record and the reproducing device of conventional digital signal, in order to explanation according to the coding method of MPEG2 situation to the active images marshalling.Referring to Figure 13, in the sign such as IBBPBBP..., I represents width of cloth I image among the figure, and B represents width of cloth B image, and P represents width of cloth P image.For example among Figure 13 A, in a limited number of frame, to organizing into groups to each active images that another width of cloth I occurs near a preceding width of cloth from I.Constitute frame number 15 frames normally in most of the cases of the image of this group.But this number is not limited to any specific number.

GOP promptly constitutes the group that resembles of this group, comprise at least one frame can be fully in a frame decoded I image.This GOP also comprises width of cloth P image and width of cloth B image, and this P image is to encode through the moving compensation prediction of being undertaken by the single directional prediction of time system according to the I image, and this B image is to be encoded by the bi-directional predicted of time system according to I image and P image.By the way, the arrow among Figure 13 A and Figure 13 B is expressed projected relationship.

In other words, the B image can only just can be encoded and decode after I image and P image are ready to.The P image that begins among the GOP can be encoded and decode all set by the I image before the P image.Second width of cloth P image and thereafter P image can be encoded and decode all set the time near the P image before this P image.Therefore, when not having the I image, no matter the visual still B image of P all can not be encoded and decode.

Referring to Figure 11, wherein label 787 is that image is reset device, and 788 is scan converter; 789 is encoder buffer, and 790 is the mode decision device, and 702 is the moving vector detector; 706 is subtracter, and 708 be a DCT circuit, comprising field store device, frame memory and DCT calculator.Label 710 is a quantizer, and 714 is inverse quantizer, and 761 is reverse DCT circuit, and 718 is adder, and 720 is shadow memory, and 722 is rate controller, and 726 is variable length coder.

Referring to Figure 12, wherein label 733 is a variable-length decoder, and 736 is reverse DCT circuit, and 737 is shadow memory, and 788 is adder, and 739 are the reverse scan transducer.By the way, moving vector detector 702 is combined with mode decision device 790 and has been represented a moving vector detecting unit.

Below, according to the record of Figure 11 to 13 explanation digital video signal and the working condition of reproducing device.Referring to Figure 11, wherein 787 pairs of images of image rearrangement device are rearranged, so that encode by order shown in Figure 13.Then, scan converter 788 conversion scanning processes make it to convert block scan into from raster scan.Thisly rearrange image and be commonly referred to preprocessing process from the conversion process that raster scan transfers block scan to.And image is reset device 787 and scan converter 788 is generically and collectively referred to as preprocessor.The pictorial data of input stands block scan by the coding order.When image was the I image, this image passed subtracter 706.When image was P image or B image, this image subtracted each other with reference picture in subtracter 706.

At this moment; Moving vector detector 702 confirms that (input of delivering to the former image of this moving vector detector 702 can be the images after image is rearranged a later images or the scanning of former picture block, and circuit size is less under then a kind of situation for the direction of motion and amount of exercise.In addition, reference picture must be imported from shadow memory 720, and has omitted the arrow among the figure), so, can from shadow memory 720, read the signal in this zone, wherein considered direction and value part.At this moment, by the mode decision device confirm 790 confirm with bi-directional predicted still be to use single directional prediction.

On subtracter 706, consider moving vector and subtraction with reference to screen.Even the image that constitutes has only less electrical power, its code efficiency is also very high.The output of subtracter 706 both can focus in the FU and also can focus in the frame unit of DCT circuit 708, handled thereby carry out DCT, and converted the data in the frequency component to.These data are input in the quantizer 710, wherein different weights are arranged for each frequency.On two yardsticks of low frequency component and high fdrequency component, these data are carried out the scanning of zigzag mode, make it do coding and Huffman coding on the length.

Control through these data of length coding and Huffman coding carrying out the coding of variable-length, thereby draw quantization table in proportion with rate controller 722 so that the value of data fit object code.The data of variable length code are usually through encoder buffer 789 outputs.Data after the quantification are got back to inverse quantizer 714; Revert to the data of former pattern space again by reverse DCT circuit 716; Utilize adder 718 with the data of former pattern space and the data addition of subtracter 706 references at last, thereby obtain and same data of decoded data.

Figure 12 is the block diagram of decoder architecture.733 pairs of variable-length decoders comprise that the pictorial data of header information (like moving vector, coded system, image pattern etc.) decodes.After these decoded data are quantized, carry out reverse DCT by reverse DCT circuit 736 and calculate (by the way: omitted the inverse quantizer that is arranged in reverse DCT circuit 736 primes among Figure 12).With reference to from the pictorial data of shadow memory 737 and consider moving vector, the data addition behind this pictorial data that will contrast through adder 738 and the reverse DCT, thus moving compensation prediction is decoded.With reverse scan transducer 739 this data transaction is become raster scan, so obtain and export an interleaved image.

In addition; The variable bit rate VBR disc system of introducing in " variable bit rate VBR disc system with sign indicating number an amount control method " literary composition of in TV circle annual meeting in 1994, delivering according to people such as Sugiyama has proposed the suggestion of relevant a kind of higher-quality encoding digital video signals method.This method fixes encoding rate with a program (the for example first cover program), so each GOP is set on a certain speed, this degree of difficulty with design and coding is relevant.Figure 14 is the block diagram of the record of conventional digital signal and the video signal coding unit in the reproducing device.Wherein, label 791 is moving compensation prediction device, and 792 is sign indicating number amount memory, and 793 is GOP rate setting unit, and 794 is sign indicating number amount assignment unit, and 795 is subtracter, and 796 is the sign indicating number batching counter, and 797 is switch.GOP setup unit 793 among Figure 14 is set, to change the setting of quantized value according to the degree of difficulty of design.In other words, when switch 797 is connected on void coding side, then yard batching counter 796 is transported in the output of variable length coder 726, calculated the sign indicating number amount that will be stored in yard amount memory 792 by it.

The empty sign indicating number amount in the whole program is confirmed according to the sign indicating number amount that is stored in yard amount memory 792 in GOP rate setting unit 793, to set and to calculate the optimal encoding code check among each GOP.The assignment of this time-code is calculated by sign indicating number amount assignment unit 794, uses for the actual coding prediction.When switch 797 is connected on the actual coding side, the assignment amount of sign indicating number amount and the value of sign indicating number batching counter 796 are compared, so can control quantizer 710 according to a real sign indicating number amount console switch 797.By this way, give the little sign indicating number amount that is of easy design, and give the big sign indicating number amount that is of difficulty, so the coding difficulty that gradually changes in the program just is overcome.Therefore according to reports, in this way with the approaching image quality of the image quality of 3Mbps speed record with the 6Mbps rate coding.

Consider the possibility of in the record of the digital video signal of CD and reproducing device, jumping over search; Even when playback I image with the P image so that when also can high speed in the front portion of GOP carrying out fast rewind under by the visit situation; The P image is positioned at the appropriate location of GOP; So during the data on search bit stream, need handle shaven head.But this control can not be accomplished on time, and this is owing to servomechanism (like actuator etc.) has the regular hour constant to cause.A GOP generally includes 15 frame images, and in the NTSC scan mode, available 0.5 second time was sought the front portion of GOP.But, in order to detect the front portion of a certain GOP, bit stream; Even when attempting when jumping over search, only to read I image or P image; Also need with a certain frame rate read its 1/2 or more so that read 1/3 image, result; When bare headed traveling time was set in 200 milliseconds, reading speed will be set in faster than 2.5 times of common speed or higher.This has exceeded the reaction limit of actuator.With common playback method, in fact can not accomplish and jump over search.

For the record and the reproducing device of the digital video signal of routine, its signal is programmed in such a way.So, when attempt to carry out as in the video tape recorder jumping over search the time, can not obtain the width of cloth image of perfectly resetting, at this moment replay data can not obtain all original images from the pictorial data clauses and subclauses of width of cloth B image and so on.Particularly, when jumping over search, can undergo mutation (improper displacement), this has influenced the processing of the output in the frame unit.In the time of will carrying out reading of variable bit rate with the high-quality image of resetting; The problem that the anterior difficulty of visit GOP itself increases can appear; This is the cause that changes because of the anterior address location of GOP, and the result has formed a space because of GOP has uneven unit in one section disk area.

Summary of the invention

The objective of the invention is to provide a kind of record and reproducing device of digital video signal, it can carry out special-purpose playback with the I image with big data quantity, and can obtain high-quality playback image, provides the method for its record and playback simultaneously.

Another object of the present invention is record and the reproducing device that a kind of digital video signal will be provided, and it can carry out high-speed replay with I image and P image with big data quantity, and can obtain high-quality playback image, provides the method for its record and playback simultaneously.

A purpose more of the present invention is record and the reproducing device that a kind of digital video signal will be provided; It can improve at the access performance that adopts GOP under the variable bit rate situation that can expect; Thereby obtain the good search of jumping over, provide the method for its record and playback simultaneously.

Another purpose of the present invention is record and the reproducing device that a kind of digital video signal will be provided; It can improve at the access performance that adopts GOP under the variable bit rate coding situation that can expect; And effectively utilize the area of space on the storage medium; Thereby accomplish and jump over search, provide the method for its record and playback simultaneously.

Record and reproducing device according to digital video signal of the present invention; When video signal recording is on the GOP unit; The one of which frame is divided into n the zone that relates to I image, makes each zone according to from being positioned at order encoding that the screen center subregion begins and the front portion that is recorded in a GOP.Simultaneously, each regional address information of I image is also noted as header information.At special-purpose playback time, only read the data that are arranged in the subregion I of screen center image, and the special-purpose playback image of output, wherein hide the data that relate in the sense data zone not fixed number value really.Therefore, have the situation of whole I images of mass data to compare with playback, special-purpose playback can be carried out at faster speed.

In the record and reproducing device of above-mentioned digital video signal, make it on whole screen, to be read through the expansion center, to export each dedicated video image.Therefore, comprehensively because screen center's partial data is expanded be the image of resetting, thereby zone that can not sense data is just not obvious, so the image of resetting seems very comfortable.

Record and reproducing device according to digital video signal of the present invention; Its video signal recording is on the GOP unit; And a frame is divided into n the zone that relates to I image, makes each zone according to from being positioned at order encoding that screen center partly begins and the front portion that is recorded in a GOP.When reading from recording mediums such as CDs (wherein each regional address information is used as header information simultaneously and writes down the I image) with the playback of video signal; Reset if at this moment carry out special use, the I pictorial data that then only is arranged in the screen center subregion is read.And the zone of sense data not wherein then lies in wherein with the data mode of hiding certain numerical value when image output is reset in special use.Therefore, compare with the situation of whole I images of resetting, special-purpose playback can realize at faster speed.

In the record and reproducing device of above-mentioned digital video signal, arrive whole screen through the core zone that expansion is read, the exportable special-purpose image of resetting.Therefore, can not sense data regional not obvious, it is very comfortable that the image of playback seems.

Record and reproducing device according to another digital video signal of the present invention; When video signal recording is on the GOP unit; The one of which frame is divided into n zone, makes each zone according to from being positioned at order encoding that the screen center subregion begins and the front portion that is recorded in a GOP.Simultaneously, each regional address information of I image is also noted as header information.At special-purpose playback time, only read the data of the I image in this territory element, and these data are exported as the image of resetting.Within a certain period of time, when the visual Zone Full of I all can not be read, can carry out interior slotting special-purpose visual output of resetting that make through making the data on this image and the last screen.Therefore, the zone of screen center part just obtains the priority of resetting, and the playback image after inserting in the result seems very comfortable.

Record and reproducing device according to another digital video signal of the present invention; When video data recording is on the GOP unit; The one of which frame is divided into n zone, makes each zone according to order encoding that partly begins from screen center and the front portion that is recorded in a GOP.Simultaneously, each the regional address information with the I image also writes down not come as header information.At special-purpose playback time; Only read the data of the I image in this regional unit, and from continuous n images one width of cloth read that each is regional 1,2 with connecing a width of cloth ... the scope among the n, result; Each image that can a comprehensive screen portions, and as the image output of resetting.When the Zone Full of I image all can not be read within a certain period of time, insert in then carrying out and make the special-purpose image output of resetting through the data that make this image and last screen.Therefore, the zone that is positioned at screen center part just obtains the priority of resetting.Owing to be to become a screen with n width of cloth I image synthesis, the therefore interior playback image of inserting is not obvious.

According to the record and the reproducing device of another digital video signal of the present invention, when video image was recorded on the GOP unit, it is regional and to each regional code that the one of which frame is divided into n of relating to I image.When the I image is recorded in a GOP anterior for each regional complex, with in the GOP unit during beginning this regional position " scrolling " of record so that record.Simultaneously, each regional address information of I image is also noted as header information.At special-purpose playback time, only read the I pictorial data in this territory element, and it is exported as the image of resetting.When all the I image all can not be read within a certain period of time, insert in then carrying out and make the special-purpose image output of resetting through the data that make this image and last screen.So, position " scrolling " that should the zone in the GOP unit, thus can smooth fashion reset a screen image.

Record and reproducing device according to another digital video signal of the present invention; When video data recording is on the GOP unit; It is regional and to each regional code that the one of which frame is divided into n of relating to I image; And be divided into an error correction block unit, and be recorded in the front portion of a GOP by the order that the zone that is positioned at screen center's part begins.Simultaneously, each regional address information of I image is also noted as header information.At special-purpose playback time, only read the I pictorial data in the error correction block unit, and as the image output of resetting.When the Zone Full of I image all can not be read within a certain period of time, then the special-purpose image of resetting is exported through making the image and the data of last screen carry out interior inserting.So, owing to be positioned at the priority that the zone of screen center's part obtains resetting, thereby the image of resetting looks very comfortable.

Record and reproducing device according to another digital video signal of the present invention; When on the unit of imagery data recording at GOP; It is regional and to each regional code that the one of which frame is divided into n of relating to I image and P image, and by being positioned at the front portion that order that screen center zone partly begins is recorded in a GOP.Simultaneously, each regional address information of I image and P image is also noted as header information.At special-purpose playback time, read I image and P image in this territory element, and as the image output of resetting.When the Zone Full of I image or P image can not be read within a certain period of time, insert in then carrying out and make the special-purpose image output of resetting through the data that make this image and last screen.So, owing to be positioned at the priority that the zone of screen core obtains resetting, thereby interior slotting playback image looks very comfortable.

Record and reproducing device according to digital video signal of the present invention; When on the unit of video signal recording at GOP; It is regional and to each regional code that the one of which frame is divided into n of relating to I image and P image, and by being positioned at the front portion that order that screen center zone partly begins is recorded in a GOP.Simultaneously, each regional address information of I image is also noted as header information.At special-purpose playback time, only read I image and the visual data of P in this territory element, and the scope of from continuous n width of cloth I image and P image, reading each

zone

1,2...n, so that comprehensively go out the image of a screen portions, and as the visual output of resetting.When the Zone Full of I image or P image can not be read within a certain period of time, insert in then carrying out and make the special-purpose image output of resetting through the data that make this image and last screen.So, owing to be positioned at the priority that the zone of screen core obtains resetting, thereby interior slotting playback image is not obvious.

According to the record and the reproducing device of another digital video signal of the present invention, when on the unit of video signal recording at GOP, it is regional and to each regional code in this frame unit that the one of which frame is divided into n of relating to I image and P image.When the frame of being told is fixedly to anterior each zone located of a GOP, then make the position " scrolling " in the zone of opening entry in this frame unit.Simultaneously, each the regional address information in the I image is also noted as header information.At special-purpose playback time, only read the I pictorial data in this territory element, and export as resetting image.When whole I images can not be read within a certain period of time, insert in then carrying out and make the special-purpose image output of resetting through the data that make this image and last screen.So, the record order of I image and P image region in the GOP unit by " scrolling ", thereby can be visual with reset one playback of shielding part of mode stably.

According to the record and the reproducing device of another digital video signal of the present invention, when on the unit of video signal recording at GOP, it is regional and to each regional code that the one of which frame is divided into n of relating to I image and P image, and tell the error correction block unit.Then by being positioned at order that the screen center subregion begins with the front portion of the frame recording of being told at a GOP.Simultaneously, each regional address information of I image is also noted as header information.At special-purpose playback time, only read the data of the I image in the error correction unit, and as the image output of resetting.When whole I images can not be read within a certain period of time, insert in then carrying out and make the special-purpose image output of resetting through the data that make this image and last screen.So, owing to be positioned at the priority that the zone of screen center's part obtains resetting, thereby the playback image after interior the inserting seems very comfortable.

Playback method (equipment) according to digital video signal; The I image intra-encoded to the major general separates according to frequency domain, quantization level or spatial resolution; To constitute the bit stream of a video data, the prior data of image in the data of wherein telling that relate to the I image at least and so on are arranged at the front portion.And the address information in the data of telling is arranged in the front portion of the bit stream of video data as header information, to constitute a packets of information.Be recorded in the data on the recording medium,, must its data sequence be rearranged according to the header information in the packets of information to be exported before data are separated at common playback time.Carry out special-purpose playback time, decode, as special-purpose output of resetting arranging fore data.Like this,, reduced, reset visual so can obtain level and smooth special use with regard to the data that special-purpose playback time is visited through data being demarcated by frequency domain, quantization level or spatial discrimination.In addition,, therefore can know the byte number that playback time should immediately be reset, like this, just can jump effectively in that special-purpose playback time is bare headed owing to write down address as the data of having divided of the header information of system flow.Have, at common playback time, data based address is rearranged again, so the shortcoming that can avoid playback time to bring because of dividing data.

Record and playback method (equipment) according to digital video signal; The I image intra-encoded to the major general separates according to frequency domain, quantization level and spatial resolution; To constitute the bit stream of a video data, the prior data of image in the data of wherein telling that relate to the I image at least and so on are arranged at the front portion.And the address information in the data of telling is arranged in the front portion of the bit stream of video data as header information, to constitute a packets of information.Be recorded in the data on the recording medium,, must its data sequence be rearranged according to the header information in the packets of information to be exported before data are separated at common playback time.Carry out special-purpose playback time, decode, as special-purpose output of resetting arranging fore data.Like this,, reduced, reset visual so can obtain level and smooth special use with regard to the data that special-purpose playback time is visited through data being demarcated by frequency domain, quantization level or spatial discrimination.In addition,, therefore can know the byte number that playback time should immediately be reset, like this, just can jump effectively in that special-purpose playback time is bare headed owing to write down address as the data of having divided of the header information of system flow.Have, at common playback time, data based address is rearranged again, so the shortcoming that can avoid playback time to bring because of dividing data.

Record and playback method (equipment) according to another digital video signal; To the major general when record intra-encoded I image be divided into n regional (n＞1); Thereby in the unit in zone, rearrange the I image that is divided into n zone; To form the bit stream of a video data, the regional arrangement that wherein is positioned at the screen center forwardly.Then, the front portion that the address information of the I image that is divided into n zone is arranged in data bit flow to be constituting a packets of information, and is recorded on the recording medium.At common playback time, rearrange the I pictorial data in this territory element according to being arranged in the anterior header information of packets of information, and with its output.At special-purpose playback time, can accomplish and special use to reset through the data of only exporting the I image that to read from the packets of information front portion within a certain period of time.So the data of dividing during through record in the screen zone have reduced the data that special-purpose playback time should be visited.Owing to the address of the data of having divided is that header information as system flow writes down; Therefore know that playback time answers the byte number of instant replay; So shaven head can jump at special-purpose playback time effectively, thereby in regular hour unit, can accomplish the jump of address.In addition, at common playback time, owing to rearrange data according to the address, the shortcoming that outcome record and playback time factor cause according to division can be avoided.

Record and playback method (equipment) according to another digital video signal; To the intra-encoded I pictorial data of major general be divided into n the zone (n＞1); Thereby in the unit in zone, rearrange the I image that is divided into n zone; To form the bit stream of a video data, the regional arrangement that wherein is positioned at the screen center forwardly.Then, the front portion that the address information of the I image that is divided into n zone is arranged in video data bit stream is as header information, to constitute a packets of information.At common playback time, rearranged in this regional unit according to the I pictorial data that the header information that is arranged in the packets of information front portion is rearranged, and exported from the recording medium that has write down these data.At special-purpose playback time, accomplish and special use to reset through only exporting the data that to read within a certain period of time.So, at special-purpose playback time,, can in the certain hour unit, accomplish the jump of address through dividing the data in the screen zone, the result has reduced the data that special-purpose playback time is wanted addressing.In addition, owing to the address of the data of having divided is that header information as system flow writes down, so the byte number that playback time should be reset can be detected at once, so shaven head can jump at special-purpose playback time effectively.In addition, because playback time is rearranged data according to the address usually, the shortcoming of bringing because of dividing data in the time of therefore can not producing data playback.

Record and playback method (equipment) according to another digital video signal; To the major general when record intra-encoded I image divide with low frequency range, high frequency region, quantization level and spatial resolution; The master data of the I image that making controls oneself divides is rearranged on screen in each regional unit; To constitute the bit stream of a video data, wherein be positioned at I image the screen core regional arrangement forwardly.The address information of the zone of having divided, the data of division and image, the front portion of bit stream that is arranged at video data be as header information, thereby constitute a packets of information, and be recorded on the recording medium.At common playback time, according to the header information that is arranged in the packets of information previous section data in the territory element are rearranged, then dateout.The data of having divided are rearranged by the order of original data.At special-purpose playback time, only export the I pictorial data that can read within a certain period of time from the packets of information front portion, reset to carry out special use.Therefore, when record, data are divided according to frequency, quantification and spatial resolution, and it is assigned in the unit in screen district.So, at special-purpose playback time, reduced, thereby can obtain a level and smooth special use and reset through reducing data volume that special-purpose playback time will visit gradually by the data of being visited.In addition,, and can detect the byte number that reset at once, therefore can accomplish the jump of shaven head effectively at special-purpose playback time at playback time owing to write down as header information the address of the data of having told.Have again, relate to multiple dividing mode and come dividing data, the data volume that can will read according to special-purpose playback speed adjustment, thus can adapt to very wide special-purpose playback speed scope.In addition, rearrange data according to the address, therefore when replay data, can not produce the shortcoming of bringing because of dividing data at common playback time.

Playback method (equipment) according to another digital video signal; I image to the major general is intra-encoded is divided according to low frequency range, high frequency region, quantization level or spatial resolution; The master data of the I image that making controls oneself divides is rearranged in each zone on screen; To constitute a video data stream, wherein be positioned at I image the screen core regional arrangement forwardly.The address information of the zone of having divided, the data of division and image, the front portion of bit stream that is arranged at video data be as header information, thereby constitute a packets of information, and be recorded on the recording medium.At common playback time, according to the header information that is arranged in the packets of information previous section arranging data of resetting, thereby by dateout on the medium.The data of having divided are rearranged by the order of original data.At special-purpose playback time, only export the I pictorial data that can read within a certain period of time from the packets of information front portion, reset to carry out special use.Therefore, data are divided according to frequency, quantification and spatial resolution, and it is assigned in the unit in screen district.So the data that special-purpose playback time is visited have reduced owing to data are assigned in the Ping Qu unit.In addition,, and can detect the byte number that reset at once, therefore can accomplish the jump of shaven head effectively at special-purpose playback time at playback time owing to write down as header information the address of the data of having told.Have again, relate to multiple dividing mode and come dividing data, the data volume that can will read according to special-purpose playback speed adjustment, thus can adapt to very wide special-purpose playback speed scope.In addition, rearrange data according to the address, therefore when replay data, can not produce the shortcoming of bringing because of dividing data at common playback time.

According to the record and the playback method (equipment) (or playback method of digital video signal (equipment)) of another digital video signal, only read the zone of the screen core that is positioned at the I image.For the data of not reading in the zone, be that a limit value comes the comprehensive image of resetting then to hide data.Therefore, the situation that has whole I images of mass data with playback is compared, and this special use is reset and can be realized with higher speed.

According to the record and the playback method (equipment) (or playback method of digital video signal (equipment)) of another digital video signal, only read the zone of the screen core that is positioned at the I image.Through with the area extension of reading comprehensive image of resetting to the whole screen.Therefore, the situation that has the entire I image of mass data with playback is compared, and this special use is reset and can be realized with higher speed, so the zone that data can not be read and not obvious.

The recording equipment of digital video signal of the present invention comprises: first code device; Be used for containing the encoding video signal of coding image, this coding image comprises the intra-encoded images with the digital video signal of moving compensation prediction and orthogonal transform coding that comes from least; Second code device is used for remaining component being encoded through coding with the vision signal output of first code device; The data placement device is used for that each is resembled the group data and resembles each dateout of arranging on the precalculated position of organizing data from first and second code devices at each.Compare with the situation that first code device is encoded to all videos signal, can reduce the zone that to visit through the essential part of active images is encoded at least.Second code device is encoded by the video information of first code device coding to the end, so just all videos information has been made coding with two code devices.In addition, the data placement device is rearranged the data that obtain with two code devices, makes data be convenient to bare headed access.Therefore, the coding of being done can make the sign indicating number amount that should be visited at special-purpose playback time at least reduce.So, the arrangement of the data that can carry out effectively being visited at special-purpose playback time at least.

In the recording equipment of above-mentioned digital video signal, comprise at least one width of cloth frame video image of the coding image of coding image with respect to containing, in a predetermined interval, video information coding is become to dredge (thin).So first code device is dredged the video image changed to this and encoded, the zone that therefore will be visited at least just reduced.When only visiting the data of first code device, can be with after the video image code Cheng Zaiqi decoding, this picture is enough to understand for the people.

In the recording equipment of above-mentioned digital video signal, the low frequency region of first code device pair of orthogonal conversion is encoded.First code device is encoded to the pictorial data that belongs to a part of frequency, has so just reduced the zone that will be visited at least.When the data of having only first code device are visited, can be with after the video image code Cheng Zaiqi decoding, this picture is enough to understand for the people.

In the recording equipment of above-mentioned digital video signal, first code device slightly quantizes on the quantization level that will encode.First code device is encoded to influencing the bigger last Bit data of image through thick quantification, then reduced the zone that will be visited at least, and can not reduce resolution.When only visiting the data of first code device, can be with after the video image code Cheng Zaiqi decoding, this picture is enough to understand for the people.

The recording equipment of another digital video signal of the present invention; From data row, extract the data in the low frequency component; Vision signal among this data row is with predetermined bit segments, this vision signal comprise contain at least one intra-encoded, come from the image of having encoded with the image of the vision signal of moving compensation prediction and orthogonal transform coding.To each block of information with data with predetermined bit segments, thereby to the low frequency region segmentation of vision signal.Therefore, be easy to the sign indicating number amount is limited in the fixing length.In addition, in carrying out low frequency region, during the decoding of data, can encode the data to the content that to understand image roughly.

The reproducing device of digital video signal of the present invention; Data in the low frequency region and the data in the high-frequency region are rearranged; Becoming predetermined order, make mode that the data of rearranging are decoded or the mode that data are carried out the selectivity decoding is in low frequency region.At common playback time,, can obtain a complete decoded picture through connecting the coded data of two segmentations.At special-purpose playback time, only the data of low frequency region are decoded.Therefore, can decode to data, thereby make the image that is obtained reach the degree that the big activation of its content is understood according to the operating state of equipment.

In the reproducing device of above-mentioned digital video signal, when data being decoded, have only and decoded data to obtain decoding with the mode of only data of low frequency range being decoded.Near can not decoded data the be dropped edge of predetermined number of bits is so the data in the high-frequency region are just replaced by the fixed numbers by reverse orthogonal transform.When special-purpose playback time is decoded to the low frequency region that comes from two segment encoding data, have only and can decoded data obtain decoding, the bit that can not decode then is dropped.So can avoid the data of abnormality are decoded.For remaining high-frequency region, then replace and decode with fixing value, then can obtain decoded image, and avoided distortion.

In the recording equipment of another digital video signal; One end of information block code is added on the coded data of each block of information of vision signal, this vision signal comprise contain at least one intra-encoded, come from when the bit that obtains as the predetermined number of low frequency range data a width of cloth coding image with the image of the digital signal of having encoded of moving compensation prediction and orthogonal transform.Exceed the above-mentioned coded data of predetermined number of bits purpose as the data of high-frequency range and be encoded.The low frequency region or the high-frequency region that no matter are block of information are all encoded by this way, promptly seem to terminate in the end of block of information (EOB) sign indicating number on the block of information surface.Therefore, when having only low-frequency data decoded, can obtain can be decoded coded data, and do not need the unnecessary circuit that for example is used for abandoning data.

The recording equipment of another digital video signal of the present invention comes recombination data according to the data in the low frequency region, data and EOB sign indicating number in the high-frequency region.Then, the mode that selection is decoded to the data of reorganization, or only the data of low frequency region are carried out the mode of selectivity decoding, thus the coded data of recombinating according to the result who selects is decoded.For high-frequency region, then with its data of fixed value replacement, to carry out reverse orthogonal transform.At common playback time, from obtaining the image that a width of cloth is decoded completely the coded data of segmentation respectively with EOB.At special-purpose playback time, have only the data in the low frequency region from coded data, to decode, thereby no matter can both operate according to the operating state of equipment, so can obtain the thick image that we can understand its picture for the still special-purpose playback of common playback.In addition, from coded data, decode when coming out when low frequency range, remaining high-frequency information piece zone then replaces with fixed value, and decodes, then should the zone can be decoded, avoided the distortion of data.The high frequency region or the low frequency range that no matter are block of information can both be decoded, seem that it is to terminate in the EOB place that this block of information looks.

The recording equipment of another digital video signal of the present invention comprises: the low code device of differentiating; Be used for the data of low resolution component are encoded; Pixel in this component is to become to have dredged with respect to vision signal; Said vision signal comprises the coding image that contains image at least one width of cloth frame, and this image is to come from the digital image of having encoded that has utilized moving compensation prediction and orthogonal transform; The difference component code device, the dateout that is used for the code device through interior slotting low resolution is encoded to the difference component that has the image before thinization of pixel; And the information adding device, be used for output with the code device of low resolution and difference component code device and be divided into predetermined zone adding error correcting code, thus composition data.The pictorial data changed is encoded and when only visiting this coded data, this pictorial data can be encoded when dredge in the space, so after should image decoded, its picture be enough to understood by the people.Inserted by interior from the decoded data that the code device of low resolution is next; Thereby through should image and the low resolution conversion before image make comparisons and obtained a difference component, so the pictorial data of the high-resolution part that can not be obtained by the low resolution code device can be encoded.Like this, the pictorial information of just low resolution degree just can not obtain coding.

The reproducing device of another digital video signal of the present invention combines the data of decoded low resolution component and the data of difference component.At common playback time; The coded data of low resolution component and the coded data of high-resolution component (it is the low resolution part and is dredged the difference component that turns between the low resolution data before) are integrated, can be decoded so have the image of whole resolution components.

In the reproducing device of above-mentioned digital video signal of the present invention, the data that will be through comprehensive low resolution and the data of difference component and the mode of image decoding is switched to a mode of decode low-resolution component.At common playback time; The coded data that is segmented into coded data and the high-resolution component (it is dredged the data and the difference component between the low resolution partial data that turn to before the low resolution) of two low resolution component integrates, and feasible image with whole resolution can be decoded.At special-purpose playback time, come the switching encoding/decoding mode according to the operating state of equipment, make and only just can decode thick image through the coded data of low resolution is decoded.

In the reproducing device of above-mentioned digital video signal, when to the image decoding of low resolution, only produce the image of inserting in the decoding back.At special-purpose playback time, when only the coded data of low resolution being decoded, the then interior video data of inserting the low resolution component is so that change back to its original size with the size of image.

The recording equipment of another digital video signal of the present invention comprises: judgment means, the degree of distortion when judging that image carries out Code And Decode according to moving compensation prediction and orthogonal transform; The adaptive coding device in order to the judgement output according to judgment means, through adaptively modifying speed, is encoded to data transfer rate; The information adding device is used for additional informations such as audio signal, header and error correcting code are added together; The data transfer rate setting device is in order to set centrifugal pump, with the speed of adaptively modifying data.In the code device of variable bit rate, speed is restricted to only can reach a limited value.Therefore, the data rate information of GOP (the sign indicating number amount that is equivalent to GOP) can be with a spot of bit representative.

The recording equipment of another digital video signal of the present invention comprises: judgment means, the degree of distortion when judging that image carries out Code And Decode according to moving compensation prediction and orthogonal transform; The adaptive coding device in order to the judgement output according to judgment means, through adaptively modifying speed, is encoded to data transfer rate; The information adding device is used for additional informations such as audio signal, header and error correcting code are added together, and wherein said equipment is consisted of can carry out multipath conversion to data rate information on shaven head etc., perhaps on the presumptive area of recording medium, it is write.Except that video data; When pictorial data is when encoding with variable bit rate, also to the data transfer rate set information be recorded on the recording medium, so; Data rate information can be read together, so when recorded information, also want and can the predetermined position of GOP on CD be noted at once.

The recording equipment of another digital video signal comprises: judgment means, the degree of distortion when judging that image carries out Code And Decode according to moving compensation prediction and orthogonal transform; The information adding device is used for additional informations such as audio signal, header and error correcting code are added together; First code device is used for being dredged the encoding video signal of having changed at predetermined interval with respect to a vision signal, and an aforementioned vision signal comprises that a width of cloth contains the coding image of intra-encoded image; Second code device; Be used for by first code device coding and the component of all the other vision signals after the output is encoded, wherein said equipment is consisted of can make first or second code device data transfer rate one of at least from change and be encoded rightly according to the judgement output of judgment means.Utilize variable speed, can realize high-quality coding.In the GOP that speed greatly increases, the space is dredged the video data changed and is encoded, and the coding of its completion can make the zone of being visited reduce at least.

The recording equipment of another digital video signal of the present invention comprises: judgment means, the degree of distortion when judging that image carries out Code And Decode according to moving compensation prediction and orthogonal transform; The information adding device is used for additional informations such as audio signal, header and error correcting code are added together; First code device is used for only the low frequency region of having done orthogonal transform with respect to vision signal being encoded, and said vision signal comprises that a width of cloth contains the coding image of intra-encoded image; Second code device; Be used for the component of all the other vision signals after by first code device signal being encoded is encoded; Wherein said equipment is consisted of can make first or second code device data transfer rate adaptively modifying one of at least according to the judgement of the judgment means that is designed output, with this to encoding video signal.Utilize variable speed, can realize high-quality coding.For the GOP that speed greatly increases, to the coding video data in the local district frequently of each block of information, the coding of its completion can make the zone of being visited reduce at least.

The recording equipment of another digital video signal comprises: judgment means, the degree of distortion when judging that image carries out Code And Decode according to moving compensation prediction and orthogonal transform; The information adding device is used for adding together such as additional informations such as audio signal, header and an error correcting code; First code device is used for respect to a vision signal through will slightly quantizing to be arranged on the quantization level and to encoding video signal; Said vision signal comprises that a width of cloth contains the coding image of intra-encoded image; Second code device; Be used for the component of all the other vision signals after by first code device signal being encoded is encoded, wherein said vision signal is according to the judgement of the judgment means that is designed output first or second code device data transfer rate adaptively modifying one of at least to be encoded.Utilize variable speed, can realize high-quality coding.In the GOP that speed is greatly increased, the data that influence in the very big last bit of image to be encoded, the coding of its completion can make the zone of being visited reduce at least.

The reproducing device of another digital video signal of the present invention is switching replay mode usually between replay mode and the special-purpose replay mode, to extract data rate information.With special-purpose replay mode the time, recording medium is in the position at special-purpose replay data place, is that the data rate information during according to special-purpose replay mode is calculated.When carrying out playback time through the data rate information of extracting each GOP, under playback situation usually, coded data dichotomous is integrated and decodes with different data rates.At special-purpose playback time, then calculate the position of wanting the GOP on the accessed record medium.Then, the data of resetting and will visit at least are so that conduct interviews to next target GOP again.At this moment, calculate the positional information of the GOP that will visit on the recording medium, to help special-purpose playback and to obtain high-quality variable bit rate.

In the reproducing device of above-mentioned digital video signal, according to the result and the special-purpose replay rate of position calculation, the control shaven head position to the recording medium.Calculate the positional information of GOP on CD according to special-purpose replay rate as access destination.Can be according to special-purpose replay rate with the Position Control of shaven head position to target GOP, so available various speed are done the playback of high-quality variable bit rate in special-purpose replay mode.

In the recording equipment of another digital video signal of the present invention; The sign indicating number amount is controlled to be corresponding to being assigned to a zone that resembles group; This resembles group is to be formed by the digital video signal that moving compensation prediction of basis and orthogonal transform are encoded; Said this equipment of the present invention comprises: code device is used for coding; Sign indicating number amount comparison means is used for the output of comparison code device and the data of scheduled volume; And feed, be used for unnecessary data are buried the white space that resembles group with clear area.Under the situation of with variable bit rate data being encoded and writing down, can shorten the access time through GOP is positioned on the easy position of visiting of shaven head, thereby increase the data volume of being read in digital coding and the recording process at special-purpose playback time.In addition, the unnecessary part of blank parts for example on the identification CD, thus utilize these parts to improve image quality, perhaps help to prolong writing time by these parts.

Another kind of digital video signal reproducing device of the present invention comprises in order to rebuild the data reconstructing device and the data deciphering device in order to the data of being rebuild by the data reconstructing device are decoded of the video signal coding data that embed original that picture group elephant.Blank parts is embedded with the coded data of other GOP data and can rebuilds, and data is decoded with activation undistortedly.And at specific playback time, data volume still increases, and still can obtain high-quality playback image.

Which of three decoding devices another digital video signal reproducing device of the present invention switch with regard to according to specific playback speed.First decoding device is reset visual in order to decode first and second coded datas and acquisition in these three decoding devices; Second decoding device is in order to the decoding first coding data and obtain such playback image; This image low frequency region of resetting corresponding to the image of process intraframe coding; The number of pixel is thinning to be dredged or thick the quantification, and the 3rd decoding device is in order to the decoding first coding data and obtain such playback image, and this image is corresponding to inter prediction image and the visual low frequency region of intraframe coding; The number of pixel is thinning to be dredged, and perhaps slightly quantizes.Since decoding only and show the pattern of I image and show the I image and the pattern of P image between carry out mode switch; So can realize the special use of I image and P image with slower special use playback (for example double speed playback) resets; Consequently reset and compare, can realize not having accurate special-purpose playback of frame-skipping with the special use of having only the I image.And, can handle the various playback speeds of for example special-purpose playback I image at the high-speed dedicated playback time.

Another digital video signal reproducing device of the present invention comprises in order to extract corresponding to from the video data draw-out device of the data of the vision signal of the sign indicating number of resetting, in order to decoding with reset from pictorial data decoding and the replay device and the mode-changeover device of the video data of video data draw-out device output.Mode-changeover device switches the pattern of normal playback mode, playback and demonstration odd field or even field and through putting upside down the pattern that field structure shows odd field or even field.At special-purpose playback time, make the field structure optimization according to pattern.When reverse reproduction, equipment provides demonstration with reverse manner work till the field shows.At the for example quick frame-skipping playback time of winding etc., through all exporting same video image, be set in certain value to number of fields and the special use that can obtain to watch is easily reset visual at even field and odd field.

Description of drawings

Will seem clearer from following detailed description together with accompanying drawing above-mentioned and other purpose of the present invention and characteristic.

Fig. 1 is the traditional video disc recording and the block diagram of reproducing device.

Fig. 2 is the block diagram of the video signal coding unit among the traditional M PEG.

Fig. 3 is the block diagram of traditional moving compensation prediction circuit.

Fig. 4 is the block diagram of the video signal decoding unit of traditional M PEG.

Fig. 5 is the sketch map of data arrangement structure of the video coding algorithm of expression traditional M PEG.

Fig. 6 is the sketch map of instance of gop structure of the video coding algorithm of expression traditional M PEG.

Fig. 7 A and 7B are the sketch mapes of an instance of the video bit stream of expression traditional M PEG.

Fig. 8 is the sketch map of the instance of system flow among the PS of expression traditional M PEG.

Fig. 9 is the sketch map of instance of the PES data packet stream of expression traditional M PEG.

Figure 10 is the traditional digital signal record and the block diagram of reproducing device.

Figure 11 is the block diagram of video signal coding unit in traditional digital video signal recording and the reproducing device.

Figure 12 is the block diagram of video signal decoding unit in traditional digital video signal recording and the reproducing device.

Figure 13 A and 13B are the sketch mapes of the principle of moving image treatment in traditional digital signal record of explanation and the reproducing device.

Figure 14 is the block diagram of traditional digital signal record and the video signal coding unit in the reproducing device.

Figure 15 is according to the digital signal record of embodiment 1 and the block diagram of the register system in the reproducing device.

Figure 16 is according to the digital signal record of embodiment 1 and the block diagram of the playback system in the reproducing device.

Figure 17 is the schematic diagram of illustration macro block.

Figure 18 is the schematic diagram that the illustration picture is divided.

Figure 19 is the schematic diagram of illustration data arrangement.

Figure 20 A to 20D is the schematic diagram of the special-purpose playback method of resetting of illustration.

Figure 21 A to 21B is the schematic diagram that under the situation of data extending, carries out special-purpose method of resetting.

Figure 22 is the schematic diagram that illustration is divided according to the picture of embodiment 3.

Figure 23 is the schematic diagram of illustration according to the data arrangement of embodiment 3.

Figure 24 A to 24E is illustration is carried out special-purpose method of resetting according to embodiment 3 a schematic diagram.

Figure 25 A to 25B is the schematic diagram that illustration is arranged according to the error correction block of embodiment 3.

Figure 26 A to 26D is illustration is carried out special-purpose method of resetting according to embodiment 4 a schematic diagram.

Figure 27 A to 27F is illustrated in the schematic diagram that carries out the method for special playback under the situation of carrying out data interpolating among the embodiment 4.

Figure 28 is the schematic diagram of the data arrangement in the illustrative embodiments 5.

Figure 29 A to 29E is the schematic diagram of illustration according to the playback method of the special use playback of embodiment 5.

Figure 30 is the schematic diagram of data arrangement in the illustrative embodiments 6.

Figure 31 A to 31F is illustrated in the schematic diagram that carries out special-purpose method of resetting among the embodiment 6.

Figure 32 A to 32B is illustrated in the schematic diagram that error correction block is arranged among the embodiment 6.

Figure 33 A to 33G is the schematic diagram that is illustrated in the playback method among the embodiment 7.

Figure 34 A to 34F is illustrated in the schematic diagram that carries out the method for special playback under the situation of carrying out data interpolating among the embodiment 7.

Figure 35 is the schematic diagram of data arrangement in the illustrative embodiments 8.

Figure 36 A to 36F is illustrated in the schematic diagram that carries out special-purpose method of resetting among the embodiment 8.

Figure 37 is the block diagram of the encoding digital video signals unit among the embodiment 9.

Figure 38 is the sketch map of

expression embodiment

9 and 11 medium frequency division principle.

Figure 39 is the flow chart that encoding digital video signals is handled among the embodiment 9.

Figure 40 is the sketch map of the header of bit stream in the illustrative embodiments 9.

Figure 41 A to 41D is the sketch map that bit stream is arranged again among the expression embodiment 9.

Figure 42 is the sketch map of the address information instance of system flow among the expression embodiment 9.

Figure 43 is the block diagram of the decoded digital video signal unit among the embodiment 9.

Figure 44 is the sketch map of the decoding processing principle among the expression embodiment 9.

Figure 45 is the flow chart of decoding processing among the embodiment 9.

Figure 46 is the block diagram of encoding digital video signals unit among the embodiment 10.

Figure 47 is the block diagram of decoded digital video signal unit among the embodiment 10.

Figure 48 is the sketch map of the embodiment of expression embodiment 10 picture areas.

Figure 49 is that to be illustrated among the embodiment 10 when video data be the sketch map of the bit stream instance of unit when arranging again with the picture area.

Figure 50 is the flow chart that encoding digital video signals is handled among the embodiment 10.

Figure 51 is the sketch map of the address information instance of system flow among the expression embodiment 10.

Figure 52 is the sketch map of the instance of system flow among the expression embodiment 10.

Figure 53 A to 53E is illustrated in can the reset sketch map of the instance of playback picture in the image of playback time.

Figure 54 A to 54E is illustrated among the embodiment 10 sketch map of instance of playback picture that has only the core output of picture at playback time.

Figure 55 A and 55b are that the core that is illustrated in picture among the embodiment 10 is exaggerated and the sketch map of the playback picture instance that shows.

Figure 56 is the flow chart that decoding digital video is handled among the embodiment 10.

Figure 57 is the block diagram of digital signal encoding unit among the embodiment 11.

Figure 58 is the flow chart that encoding digital video signals is handled among the embodiment 11.

Figure 59 A to 59D is the sketch map of system flow instance among the expression embodiment 11.

Figure 60 is the sketch map of the address information instance of system flow among the expression embodiment 11.

Figure 61 is the block diagram of decoded digital video signal unit among the embodiment 11.

Figure 62 is the flow chart that decoded digital video signal is handled among the embodiment 11.

Figure 63 is the block diagram of encoding digital video signals unit among the embodiment 11.

Figure 64 is the sketch map that the resolution on the picture becomes the principle of changeing among the expression embodiment 12.

Figure 65 is the sketch map that the data in

illustrative embodiments

12,13 and 14 are formed result's instance.

Figure 66 is the block diagram of encoding digital video signals unit among the embodiment 13.

Figure 67 is the sketch map of an instance of the data arrangement of DCT coefficient in the illustration DCT piece.

Figure 68 is a block diagram of implementing encoding digital video signals unit in 14.

Figure 69 is the sketch map of the instance of coded data statistic in

illustrative embodiments

12,13 and 14.

Figure 70 is the sketch map of an instance of processing sequence among the

expression embodiment

12,13 and 14.

Figure 71 A to 71D is the sketch map of illustrative embodiments 15 distribution shape of the frequency component of the bit stream of DCT piece in a piece.

Figure 72 A is the block diagram of decoded digital video signal unit among the embodiment 15.

Figure 72 B is the sketch map of the operation principle that decoded digital video signal is handled in the illustrative embodiments 15.

Figure 73 A is the block diagram of encoding digital video signals unit among the expression embodiment 16.

Figure 73 B is the sketch map of the operation principle that encoding digital video signals is handled in the illustrative embodiments 16.

Figure 74 is the block diagram of decoded digital video signal unit among the embodiment 16.

Figure 75 is the block diagram of decoded digital video signal unit among the embodiment 17.

Figure 76 is the block diagram of GOP address generator and optical disc controller among the embodiment 18.

Figure 77 comprises reset the GOP address generator of processing and the block diagram of optical disc controller among the embodiment 18.

Figure 78 be when in embodiment 19, carry out by frequency partition and when quantize dividing the block diagram of decoded digital video signal unit.

Figure 79 is the block diagram of decoded digital video signal unit when in embodiment 19, carrying out dividing by bit length.

Figure 80 is the block diagram of decoded digital video signal unit when in embodiment 19, carrying out dividing by resolution.

Figure 81 is the block diagram of encoding digital video signals unit among the embodiment 20.

Figure 82 is the block diagram of decoded digital video signal unit among the embodiment 20.

The sketch map of Figure 83 A and the 83B principle that to be illustration handle with digital video signal recording and reproducing device among the embodiment 20.

Figure 84 be in embodiment 21, carry out by frequency partition or when quantize dividing the block diagram of decoded digital video signal unit.

Figure 85 carries out dividing or the block diagram of decoded digital video signal unit when quantizing to divide by bit length in embodiment 21.

Figure 86 carries out dividing and the block diagram of decoded digital video signal unit when quantizing to divide by resolution in embodiment 21.

Figure 87 be in embodiment 22, carry out by frequency partition and when quantize dividing the block diagram of decoded digital video signal unit.

Figure 88 carries out dividing and the block diagram of decoded digital video signal unit when quantizing to divide by bit length in embodiment 22.

Figure 89 A and 89B are the sketch mapes that is illustrated in the principle of handling when jumping search among the embodiment 22.

Figure 90 A and 90B are the sketch mapes of the principle handled when being illustrated in reverse reproduction among the embodiment 22.

Embodiment

Accompanying drawing according to each embodiment of expression carries out illustrated in detail to the present invention now.

At first explain embodiments of the invention 1.Figure 15 is the block diagram of the register system of digital video signal recording and reproducing device among the expression embodiment 1.Consult Figure 15, from the digital video signal pattern of the input circuit 3 of input terminal 1 output.Import the first input end of subtracter 4 and second input of moving compensation prediction circuit 11 from the video electrical signal of formating circuit 3 outputs.The output signal of subtracter 4 is through DCT circuit 5 input quantizers 6.The output signal of quantizer 6 is through the first input end of variable length coder 7 input buffer storages 12.Simultaneously, the output signal of quantizer 6 is also through the reverse DCT circuit 9 of inverse quantizer 8 inputs.The output signal of reverse DCT circuit 9 is then delivered to the first input end of adder 10.

The output signal of adder 10 is delivered to the first input end of moving compensation prediction circuit 11.Second input of adder 10 and second input of subtracter 4 are delivered in first output of moving compensation prediction circuit 11.In addition, second input of buffer storage 12 is delivered in second of moving compensation prediction circuit 11 output.The output of buffer storage 12 is then through format encoded device 13 input modulators 14.The output of modulator 14 is recorded on the recording medium such as CD etc. through lead-out terminal 2.

Figure 16 is the circuit block diagram of expression according to playback system in the digital video signal recording of embodiment 1 and the reproducing device.With reference to Figure 16, the video information of reading from recording medium is from input terminal 20 input modulators 21.The output of demodulator 21 is then through buffer storage 22 pattern of the input decoders 23.Variable-length decoder 24 is then imported in first output of format decoder 23, and carries out inverse quantization in inverse quantizer 25.Export signal then and stand reverse DCT, deliver to the first input end of adder 28 again at reverse DCT circuit 26.Simultaneously, prediction data decoding circuit 27 is then imported in second of format decoder 23 output.Deliver to the output of prediction data decoding circuit 27 second input of adder 28 then.The output of adder 28 is then exported from lead-out terminal 30 through unformatted circuit 29.

Follow the working condition of explanation facilities.Digital video signal is imported from input terminal with behavior unit, and is transported to formating circuit 3.Here when moving compensation prediction, as shown in figure 6 as conventional example, a GOP is configured to 15 frames, carries out predictive coding, 1 frame I image, 4 frame P images (P1 to P4), and 10 frame B images (B1 to B10).In this case, in formating circuit 3, arrange again, and be that unit exports with the order shown in Fig. 7 with the frame with the video data of continuation mode input.

In addition, be that unit is arranged again with the data of behavior unit input with the piece of 8 * 8 pixels, thereby constitute macro block (6 pieces altogether, on for example adjacent four brightness signal Y pieces and the position with corresponding two color difference signal Cr of Y piece and Cb piece.Data are that unit exports with the macro block.Here, macro block is determined with the least unit of moving compensation prediction, and the moving vector of moving compensation prediction is to be that unit is determined with the macro block.

And, for I image, be divided into three zones to one-frame video data with formating circuit 3, in this zone, be that unit carries out piecemeal therefore with 8 * 8 pixels, constitute and export macro block.Here, three zone of being divided that kind as shown in Figure 18 are set at

zone

1,2 and 3 from picture jack-up.In Figure 18, the zone 2 that is positioned at the picture core is of a size of 720 pixels * 288 row, and picture two end regions are of a size of 720 pixels * 96 row.Simultaneously, in P image and B image, carry out piecemeal regionally, and be unit ground output with the macro block without separating into each.

The output of formating circuit 3 is input to subtracter 4 and moving compensation prediction circuit 11.The working condition of subtracter 4, DCT circuit 5, quantizer 6, variable length coder 7, reverse digital quantizer 8, reverse DCT circuit 9, adder 10 and moving compensation prediction circuit 11 is identical with traditional embodiment, and the explanation of relevant their working conditions is just in advance to omit.

Input to buffer storage 12 from the video data of variable length coder 7 outputs and the moving vector of driven compensation prediction circuit 11 outputs.The moving vector and the video data of a GOP part of record in buffer storage 12, data order then output to format encoded device 13.The output of format encoded device 13 is transfused to modulator 14, and error correcting code etc. are added to and are recorded on the recording medium such as CD etc.

In format encoded device 13, the video data of a GOP part is arranged with data array as shown in Figure 19 again, outputs to modulator 14 again.Here, the I image is divided into three zones as shown in Figure 18.When being set to I (1), I (2) and I (3), just constituted the I pictorial data, thereby data just are recorded in the fore-end of a GOP partial data string with the order of I (2), I (1) and I (3) corresponding to the visual data of the I in these zones 1 to 3.

In addition, the storage of each image region stores as header information in the address of GOP fore-end.The byte number that the data of tripartite each image region are occupied in form shown in Figure 19 also stores as header information.Therefore, according to the byte number that each zone of in header information, being write down is occupied, can regard each regional final position from the relative position of GOP front end as at playback time.Therefore, at special-purpose playback time, shaven head is the front end address that unit jumps to GOP with the certain hour, thereby according to reading each regional data from the header information of GOP front end.

Adopt general video signal recording and reproducing device, the record format during according to data record, the I image is that record is given in the unit with the frame.By contrast, in Figure 19, priority is arranged in the zone that is divided into three part I pictorial data picture cores, makes this zone be positioned at the front end of a GOP.Therefore, when high-speed replay, have only under the situation that this regional part of I image can decode within a certain period of time, the playback image of picture core can be exported at least.

The working condition of playback duration then is described according to Figure 16.Demodulator 21 carries out correction process, makes in format decoder 23, to be divided into moving vector and video data with the vision signal of format record shown in Figure 19 in buffer storage 22, exports prediction data decoding circuit 27 and variable-length decoder 24 respectively to.Here, the working condition when normal playback is identical with traditional embodiment, and the explanation of relevant its working condition is omitted.

When high-speed replay; For with a GOP unit record such as the data on the recording medium of CD etc.; Shaven head is the front end that the unit jumps to a GOP with the certain hour; Thereby according to the header information that is recorded in front end is the data division that the I image is read by unit with the zone, makes data in decoder 21, obtain decoding and input buffer storage 22.Here, under the situation such as the recording medium sense data of CD etc., even know the front end address of the GOP that is recorded on the CD, the time of waiting the CD rotation when jumping to the front end of GOP also can increase when high-speed replay.

Therefore, speed increases when high-speed replay, is used to read the time shortening of the last data of CD.Change owing to wait the time of CD rotation, so can not read all I pictorial data with stable manner.Therefore, when the speed of high-speed replay rose, after the data in the zone that only is positioned at the picture core 2 were read, shaven head just jumped to the front end of the GOP of back, thereby had only the data input buffer memory 22 in the zone 2 that can be read.In this case, the zone 2 of the I image that can be read of 23 decodings of format decoder.On the other hand, zone 1 that its data are not read and 3 usefulness grey data are sheltered, and output high-speed replay image, therefore, are set under the situation of 15 frames at a GOP, and the special use that can obtain 15 times of speed is reset visual.

Figure 20 is illustrated in through a playback and carries out the playback image under the high-speed replay situation from the zone 2 of the I image of nGOP to the n+3GOP of a GOP.In Figure 20, the

zone

1 and 3 at picture two ends is sheltered by grey data among the figure.In addition, when the amount of information of I image is few, CD rotation waiting time is short, and can obtain to read the time of data in the

zone

1 and 3 time, and the data in regional 1 and 3 are not just decoded.This is because when high-speed replay; All data of a picture part can not be read with stable manner; If only when the data in the

zone

1 and 3 can be read, just exporting shielding; These zones just can not be exported in certain time interval, thereby make the high-speed replay image seem unnatural.

As stated, because the special-purpose I image of resetting that is used for as shown in Figure 19 is arranged in the zone that priority is positioned at a picture center, the front end of I picture record GOP on recording medium.Even the speed of high-speed replay increases, in order to carry out high-speed replay, the data that also only are positioned at the zone 2 at center are read, and consequently see the content of resetting image easily.In addition,, therefore compare, can realize that more the special use of high speed is reset with the situation of reading the entire I image owing to have only the data in regional 2 scopes to read from recording medium.

In the foregoing description 1, the I image vertically is divided into three regional records as shown in Figure 18 and gets off, but image is not to be divided into three zones.Can use the section of the international standard MPEG defined of record data to be divided into n zone (n＜1) as unit.Here, the section has the one-dimentional structure of the macro block of any amount (more than) length, therefore when arriving the right-hand member of picture, just continues to show the next line left end.

Below just with reference to description of drawings embodiments of the invention 2.Figure 21 is in order to the schematic diagram of the special-purpose playback method under the situation of in embodiment 2, carrying out data extending to be described.In embodiment 1, the I image is divided into three zones as shown in Figure 18, and the feasible data that only are positioned at the zone 2 at picture center are read playback.Therefore for

zone

1 and 3, only export masking data.Yet the data that kind shown in figure 21 in zone 2 expands to the size of a picture.

In this case, when converting vision signal to behavior unit data, the numerical value in zone 2 is carried out interpolation, extend to the size of a picture part, export again with unformatted circuit 29.Under the situation of Figure 21, the size in zone 2 is 720 pixels * 288 row, vertically forms 144 row symmetries from the picture center.

Here; At special-purpose playback time, if as shown in Figure 21 A, be set at AR2a to the first half in zone 2, the latter half is set at AR2b; AR2a and AR2b just vertically expand 1.5 times respectively, make reset visual AR2a ' and the such ground symmetry of AR2b ' shown in Figure 21 B.For the method for expansion image, when the data definition with each behavior unit of AR2a is AT (l) (l: line number, 1≤l≤144), and the line data of the first half of expansion picture is when being set at DT (m) (1≤m≤240), and the expansion of being done is represented with formula:

DT(3n-2)＝AT(2n-1)

DT(3n-1)＝AT(2n-1)

DT (3n)=AT (2n) (n=1 to 80)

Simultaneously, when the data definition with each behavior unit of AR2b is AB (l) (l: line number, 1≤l≤144), and the line data of expansion picture the latter half is when being set at DB (m) (1≤m≤240), and the expansion of being done can be represented by formula.

DB(3n-2)＝AB(2n-1)

DB(3n-1)＝AB(2n-1)

DB (3n)=AB (2n) (n=1 to 80)

As stated, when high-speed replay, read-only fetch bit in the data in the zone 2 at picture center, make it to expand to the size of a width of cloth picture and as the image output of resetting.Therefore, because the image two ends are not masked when high-speed replay, so the playback image is no longer ugly.

In the foregoing description 2, picture is vertically expanded through inserting data with behavior unit simply.Line data vertical direction relatively carries out linear interpolation.

Embodiments of the invention 3 are described now.The structure of digital video signal recording and reproducing device register system and playback system and enforcement 1 identical (seeing Figure 15 and 16) in embodiment 3.

Just introduce the working condition of equipment below.Digital video signal from input terminal 1 input, is transported to formating circuit 2 with behavior unit.Here when moving compensation prediction, a GOP resembles and is set at 15 frames the conventional example shown in Fig. 6.So.GOP is through predictive coding, as frame I image, 4 frame P images (P1 to P4), 10 frame B images (B1 to B10).In this case, in formating circuit 3, resembling traditional instance with the video data of continuation mode input is that unit is arranged again with the frame with as shown in Figure 7 order, exports again.In addition; Data so that behavior unit was imported are that unit is arranged again with the piece with 8 * 8 pixels; To constitute as shown in Figure 17 macro block (adjacent four brightness signal Y pieces and two color difference signal Cr and Cb piece have six altogether), make that data are that unit exports with the macro block.Here bulk is the least unit of compensation prediction, and the moving vector of moving compensation prediction is unit and definite with the macro block.

In addition, in formating circuit 3, the I image is divided into five zones along the vertical direction of one-frame video data, and each zone is 720 pixels * 96 row.In this zone,, constitute macro block and export with 8 * 8 pixel piecemeals.Five zone definitions of being divided in this case, are

zone

1,2,3,4 and 5.Simultaneously, P image and B image be by piecemeal, but zoning not is that unit exports with the macro block.

The output of formating circuit 3 is input to subtracter 4 and moving compensation prediction circuit 11.The working condition that subtracter 4, DCT circuit 5, quantizer 6, variable length coder 7, reverse analog-digital commutator 8, reverse DCT circuit 9, adder 10 and moving repaid prediction circuit 11 is identical with traditional embodiment, and the explanation of these devices is omitted in the pass.

From the video data of variable length coder 7 outputs and the moving vector input buffer storage 12 of driven compensating circuit 11 outputs.The video data and the moving vector of a GOP part of record in buffer storage 12, above-mentioned data export format encoded device 13 in proper order to.The output of format decoder 13 is input modulator 14 then, thereby adds error correcting code etc. and be recorded on the recording medium such as CD etc.

In format encoded device 13, export the GOP partial data to modulator 14 through in data arrangement shown in figure 23, arranging vision signal again.The I image is divided into five zones shown in figure 22, makes that corresponding to the visual data definition of the I in zone 1 to 5 be I (1), I (2), I (3), I (4) and I (5).In Figure 23, the data structure of I image becomes with the journal of I (1), I (2), I (3), (I4) and I (5) the data flow front end at a GOP, makes priority be positioned at the zone at picture center.

In addition, in Figure 23, the address of storing the data of every width of cloth I image writes as header information.As header information, write down each regional data shared byte number in data format, the zone obtains through being divided into five parts to the I image.Therefore at playback time,, can regard each regional terminal position as relative address at playback time at playback time each regional byte number that is occupied with respect to the GOP front end according to being recorded in the byte number that each zone is occupied in the header information.As a result, shaven head is the front end address that the unit jumps to GOP with the certain hour, makes that the data of I image can be according to reading from the header information of GOP front end.

For general video signal recording and reproducing device commonly used, the data format during with record is that record is given by unit with the frame with the I image.By contrast; In Figure 23; Priority is arranged in the zone of being waited to be arranged in the picture core in five zones that the I image of a GOP front end obtained by division, even the result has only the subregion in the I image can be decoded, the playback image of picture core can be exported at least.

The working condition of playback time just is described according to Figure 16 then.In demodulator 21, be divided into moving vector and video data through correction process and with the vision signal of the format record among Figure 23 in buffer storage 22, moving vector and video data export prediction data decoding circuit 27 and variable-length decoder 24 respectively to.Here, the working condition during normal playback is identical with traditional embodiment, and its detailed introduction is just omitted.

When high-speed replay; For being recorded in such as the data on the recording medium of CD etc., shaven head is the front end that unit jumps to GOP with the certain hour, reads the data division of I image according to header information; Data are carried out demodulation in demodulator 21, with input buffer storage 22.Yet, under the situation that the amount of information of I image can not be read too greatly within a certain period of time, to have read half the data and read until the last term of data, shaven head jumps to down the front end of back GOP, the data input buffer memory 22 that only can read.In this case, in format decoder 23, the zone of the decoding I image that can read only, and export as the high-speed replay image.Therefore, when GOP was set at 15 frames, the special use that can obtain 15 times of speed was reset visual.

Figure 24 playback image under the situation of I image of a GOP of representing to reset.In this case, the amount of information of I image All Ranges is too big and can not read from recording medium, for the zone that can not read, front region is kept intact export, thereby it is visual to have synthesized high-speed replay.In Figure 24, under the situation that can not read n+1 GOP zone 5 and n+3

GOP zone

1 and 5, the last width of cloth playback image of this width of cloth playback image is just kept intact.

By this way; Be used for the special-purpose I image of resetting as shown in Figure 23 and be arranged so that priority is positioned at the zone that is recorded in the core of the picture of a GOP front end on the recording medium; Therefore; Even can not read entire I when image, also give the core of picture in the playback time priority, make the reset content of image understand easily.

In the foregoing description 3, in the time can not reading the entire I image, the image of resetting just is that unit carries out interpolation with the zone, if be not that unit carries out interpolation with the zone, can error correction block be that unit carries out interpolation then.

In this case, according to the data arrangement shown in Figure 23, demodulator 21 is divided into the packet of some byte longs to data and is added to error correcting code on each packet.It is that unit is divided into the instance under the situation of several packets with the data of continuation mode input with the error correction block unit that Figure 25 is illustrated in five zones.Figure 25 A representes the serial data before packet is divided.Figure 25 B representes the serial data after packet is divided.Five zones of I image are divided into the plurality of data bag of a constant volume, and area I (3) is divided into the packet from I to i, and area I (4) is divided into the packet from i to j, so that input.

When high-speed replay, for being unit record with GOP such as the data on the recording medium of CD etc., shaven head is the front end that unit jumps to GOP with the certain hour, is the data division that unit reads the I image with the zone according to header information.Data division carries out demodulation by demodulator 21, input buffer storage 22.Yet, can not read within a certain period of time under the situation of entire I image because of containing much information of I image, even when just when of reading of data is regional, shaven head also jumps to the front end of back GOP.In addition, the data that can read stand correction process, and making can be by the data input buffer memory 22 of error correction.In this case, format decoder 23 can be discerned the address of the I image region in the decoding way, makes that the data that can be read are that unit decodes with the macro block, and exports as the high-speed replay image.In this case, for the macro block that can not decode, just in statu quo keep last width of cloth picture to export.

In the foregoing description 3, the data in each zone of I image are divided into the plurality of data bag with continuation mode.Yet, can be divided into data and make that the data in zone can be not included in the packet more than two.In this case, the data of an area part are sealed with the integral multiple of error correction block, are right after consequently that data can be that unit is arranged again with the zone after the correction process.When each regional data was divided into packet unit, half inputed to the last packet in each zone data, makes remaining data need be placed on (for example, all data are all sheltered and are " 1 ") in the data masking.

In addition, in the foregoing description 3, the order accord priority with 3,2,4,1 and 5.Yet, have more than and be limited to this order.Order for example can be 3,4,2,5 and 1.

And in the foregoing description 3, I image along continuous straight runs is divided into five zones, and the record of giving shown in figure 22 suchly.Data not necessarily will be divided into five zones, but the section that can international standard MPEG be limited is that unit is divided into n zone (n＞1).Here, the section has the one-dimentional structure of the macro block of random length number (more than).When being an arrival picture right-hand member, the section arrives the band of next line left end continuously.

Below just according to explained embodiments of the invention 4.Figure 26 is the sketch map of special-purpose playback method among the expression embodiment 4.In embodiment 3, be to carry out special-purpose the playback with the playback method shown in Figure 24.Yet, also can carry out special use and reset, export the playback image of that kind shown in figure 26.In this case, each zone of the I image through such as shown in Figure 26 continuous five GOP that reset, format decoder 23 a just synthetic width of cloth pictures.For example in Figure 26 A; A visual picture part of playback is visual synthetic from the I of n to the n+4GOP, makes the I image of n+4GOP reset regional 5 at the I image of regional 4 playbacks and nGOP at the I image of I image 3 playbacks, the n+1GOP in the zone of regional 2 playbacks, n+2GOP at the I image of regional 1 playback, n+3GOP.In addition, with reference to Figure 26, regional 5 the time when noting, the I image of n, n+1, n+2......GOP is reset as the playback of video data.

In addition, in the time can not reading the entire I image within a certain period of time owing to containing much information of I image, the data of last width of cloth picture are kept intact and are exported, with synthetic high-speed replay image.

Figure 27 is that the playback when n+1GOP zone 5 and n+

3GOP zone

1 and 5 can not be read is visual.In this case; Because data arrangement is to be recorded on the recording medium through will preferably weighing the zone that is positioned at picture core shown in figure 23 earlier; Under the situation that therefore even the entire I image can not be read according to the time; Playback time also awards priority the core of picture, and the result is the ugly sight of image of can not resetting.And even under the situation that the data in zone more than two can not be read, a width of cloth picture also is divided into five zones.Because the frame of in each zone, resetting is inequality, is difficult to detect the data that in the playback image, lacked.

Then introduce embodiments of the invention 5 with reference to accompanying drawing.Figure 28 is the sketch map of expression according to the arrangement architecture of the digital video signal data of embodiment 5.In embodiment 3, array is that kind is write as with the order in

zone

3,2,4,1 and 5 for the I image as shown in Figure 23.Array also can have the structure shown in Figure 28.In Figure 28, when the data of I image were recorded in a GOP partial data arrangement fore-end, the regional number on the front end of each GOP was by last volume.In other words, as shown in Figure 28, when the I pictorial data is recorded in nGOP when going up with the order of I (5), I (1), I (2), I (3) and I (4), the I pictorial data just is recorded among the n+1GOP with the order of I (1), I (2), I (3), I (4) and I (5).And I (2) at first gets into n+2GOP.Become n+3, n+4 etc. when GOP number ... the time, front end area goes up volume in proper order, with I (3), I (4), I (5), I (1) etc.. give record.

In addition, at the front end of GOP, store each I image data the address and write as header information in order to the information of identification front end area type.As header information, comprise the regional number that is recorded in front end and represent the byte number of shared data volume on each regional data format as shown in Figure 27.Therefore; At playback time, there has been the front end E territory that is recorded in the header information number and each zone each regional end position on the byte number that is occupied on the recording medium just can be the data sequence of I area of image and recording medium to regard relative position as about the GOP front end.Therefore at special-purpose playback time, shaven head is a unit with the certain hour, jumps to the front end address of GOP, just gets the I pictorial data and can both read according to header information for each zone from the GOP front end.

In this case, the position that record is divided into the I image region in five zones is to roll up in the unit with GOP, even therefore only some zone can be decoded at special-purpose playback time I image, zone that can not be decoded is not concentrated on the fixed position in the picture yet.

When high-speed replay; For the data on the recording medium that is recorded in CD for example etc., shaven head is the front end that unit jumps to GOP with the certain hour, is the data division that the I image is read by unit with the zone according to header information; In demodulator 21, carry out demodulation, input to buffer storage 22 again.Yet when the amount of information of I image can not be read the entire I image too greatly within a certain period of time; After being read by that the regional final data in being read on the way; Shaven head just jumps to the front end of next GOP, so that have only the data input buffer memory 22 in the zone that can be read.In this case, 23 decodings of format decoder are as the zone of high-speed replay image I image output, that can be read.Therefore, be set under the situation of 15 frames at a GOP, the special use that obtains 15 times of speed is reset visual.

Playback image when Figure 29 is illustrated in high-speed replay under the situation of the I image of a GOP of playback, in this case, the I image is recorded on the recording medium with the order shown in Figure 28.Here, under the containing much information of I image, situation that the entire I image can not in time be read, the data of last width of cloth picture are just kept intact and exports, so that synthetic high-speed replay is visual.Figure 29 representes the situation that n+1GOP zone 5 and n+

3GOP zone

1 and 2 can not be read fully.In this case, the data of last width of cloth picture are just kept intact.

As stated, to be used for the order of the special-purpose I image of resetting be to roll up in the unit with GOP for as shown in Figure 28 record.Therefore, even have only under some regional situation that can be decoded of I image at special-purpose playback time, zone that can not be decoded is not concentrated on the fixed position of picture yet.

Introduce embodiments of the invention 6 with reference to the accompanying drawings.Figure 30 is the sketch map of expression according to the data arrangement structure of the digital of digital video data of embodiment 6.In this case, I image and P image are divided into five zones, respectively are 720 pixels * 96 row, make that each zone is that unit carries out piecemeal with the macro block, and that kind are as shown in Figure 22 encoded.Yet the P image is divided into five zones.Mode so that the range of search of the reference pattern of moving compensation prediction is closed in the zone is moved compensation prediction and coding.Here, five zones being divided are defined as 1,2,3,4 and 5 from the top.In addition, for the B image, the zoning is not just moved compensation prediction and is encoded.

With format encoded device 13, arrange vision signal with the data arrangement shown in Figure 30 again with a GOP partial data, and export modulator 14 to.Here, for I image and P image, width of cloth picture that kind as shown in Figure 22 is divided into five zones.I image and P1, P2, P3 and P4 image be defined as I (1) to I (5) and Pi (1) to Pi (5) (i=1 to 4) in Figure 30; The front end of data structure Cheng Congyi GOP partial data string of I image, P1, P2, P3 and P4 image rises with 3,2,4,1 and 5 order gives record, makes priority be positioned at the zone of picture core.In addition, in Figure 30, each regional data volume is recorded in the front end of GOP as header information, thereby can discern the address of the data in each I image and the P image region.

When high-speed replay; For the data that are unit record with a GOP on the recording medium of for example CD etc., shaven head jumps to the front end of GOP, thereby is that I image and the regional data division of P are read by unit with the zone; Give demodulation by demodulator 21, and input buffer storage 22.Yet when the amount of information of I image and P image can not be read entire I image and P image too greatly within a certain period of time, the zone of reading in the way was read until last data.This moment, shaven head was the front end that unit jumps to GOP with the certain hour, made and had only the data input buffer memory of being read 22.In this case, decode in 23 zones visual to the I that can be read and the P image of format decoder, and dateout is as the high-speed replay image then.Therefore, be set under the situation of 15 frames at a GOP, the special use that can obtain the triple speed degree is reset visual.

In addition; Because priority is positioned at the zone of the core of picture; So that be arranged in the front end of a GOP in the I image that is divided into five zones,, at least also can export the core of picture even having only under the decoded situation of the visual part ability of I image or P.In addition, image is to be recorded on the recording medium with the order that I is visual, P1 is visual, P2 is visual, P3 is visual and P4 is visual.Therefore, even when all data all can not be read, can not survey can not the reset situation of reference data of data decode circuitry 27 yet.

It is visual that Figure 31 is illustrated in the playback of carrying out under the situation of high-speed replay of image through the I among the GOP that only resets image and P image.In this case; When figure I image can not be within a certain period of time read entire I image and whole P when visual from recording medium with containing much information of P image; The data of a last width of cloth picture are just kept intact and are exported, so that for the synthetic high-speed replay image in the zone that can not read.Figure 31 representes the situation that

zone

3,4 and 5 can not be read among the P4 of nGOP.In this case, the data of last width of cloth picture are kept intact.

As stated, as shown in Figure 30, be that used I image of special-purpose playback time and P image are collected and arranged in unit through zone with a GOP front end, and at special-purpose playback time playback I image and P image, the special-purpose image of resetting of output.In addition, under the situation that can not read entire I image and whole P image because of time restriction, the numerical value of a last width of cloth picture will interpolation, and is visual to allow output to reset.

In the foregoing description 6, under the situation that can not read entire I image and P image, be that unit carries out interpolation to the image of resetting with the zone.Yet, also can not be that unit carries out interpolation, but be that unit carries out interpolation with the error correcting code with the zone.

In this case, for the data arrangement shown in Figure 30, demodulator 21 is divided into the packet of some bytes to data, makes error correcting code be added on each packet.Figure 32 representes to be divided into the packet that several are made up of error correction block with the data in five zones of continuation mode input.Figure 32 A representes the serial data before packet is divided.Figure 32 B then representes the data after packet is divided.In Figure 32, data are divided into i to the j packet among the regional P1 (3).

When high-speed replay; For with GOP being the data of unit record on the recording medium of for example CD etc.; Shaven head is the front end that unit jumps to GOP with the certain hour; Thereby according to header information is the data division that the I image is read by unit with the zone, carries out demodulation at demodulator 21, and input buffer storage 22 again.Yet make containing much information of I image under the situation about can not read within a certain period of time of entire I image and whole P image, even in the way of the data that read an area part, shaven head also can jump to the front end of next GOP.In addition, the data that have been read stand correction process, through the slow reservoir 22 of leaving with of the data input of error correction.In this case, the I image in the format decoder 23 identification decoding ways and the address of P image make that the data that can be read are that unit decodes with the macro block, and export as the high-speed replay image.In this case, for the macro block that can not decode, the data of a last width of cloth picture are just kept intact and are exported.

In the foregoing description 6, when moving compensation prediction, detection range is set for and is enclosed in each zone, but always need not seal.

Introduce embodiments of the invention 7 with reference to Figure 33 below.Figure 33 is the sketch map of the special playback method of expression embodiment 7.In embodiment 6, carry out special playback with playback method shown in Figure 31.Yet, can carry out special use and reset, make the playback image export as shown in Figure 33 suchly.In this case, through from continuous five frames shown in figure 33 playback area one by one, format encoded device 23 a synthetic width of cloth pictures.In Figure 33 A, the playback image of a picture part is synthetic from I image, P1 to P4 image.In addition, in Figure 33 A, at regional 1 playback P4 image, at regional 2 playback P3 image, at regional 3 playback P2 image, at regional 4 playback P1 image, and at regional 5 playback I image.In addition, in Figure 33, note regional 5 elapsed times.The video data of resetting comprises the I image of nGOP, P1, P2, P3 and P4, and the I of n+1GOP image, and P1 image.

In addition, under the situation that all I are visual and the P image can not be read within a certain period of time that contains much information of I image and P image, the data of a last width of cloth picture are kept intact and are exported, synthetic high-speed replay image.Playback image under the situation that Figure 34 representes can not read in the

zone

1,4 and 5 of nGOP.In this case, as shown in Figure 30, for serial data, priority is positioned at the zone of core, so that be recorded on the recording medium, has not given the picture core ugly situation of image of resetting thereby playback time can not take place because of priority.In addition, even under the situation that data regional more than two can not read, a width of cloth picture also is divided into five zones, and the frame of being reset in each zone has nothing in common with each other, and the data that in the playback image, lacked just are difficult to see.

Then introduce embodiments of the invention 8 with reference to accompanying drawing.Figure 35 is the sketch map of expression embodiment 8 digital of digital video data arrangement architectures.In embodiment 6, data arrangement is write as with as shown in Figure 30

zone

3,2,4,1 and 5 order, but this arrangement also can have structure as shown in Figure 35.

In Figure 35, when the data of I image and P image were recorded in the front end of a GOP partial data arrangement, for each frame, the regional number on the front end was all gone up volume.In other words, as shown in Figure 28, the I pictorial data is with the order of P1 (2), P1 (3), P1 (4), P1 (5) and P1 (1) record in nGOP.In this external P2 image, P2 (3) has become front end.In P3 image and P4 image, front end area is by last volume, and resembles ground journal P3 (4) and the P4 (5).

In addition, at the front end of GOP, the address of the data that storage I image and P are visual and the type information of discerning the zone on each frame front end are noted as header information.Here, write down the byte number that the regional number that writes down at each regional front end and expression are divided into five data volumes in each zone of part as header information.Therefore, at special-purpose playback time, Optical devices are the front end that unit jumps to GOP with the certain hour, thereby can the zone be the unit sense data according to header information.

In this case; Because it is to roll up in the unit with the frame that I image and P image are divided into the position of five parts; Even having only under the regional situation that can be decoded of a part of I image and P image, the situation of not decoded regional centralized on the fixed position of picture can not take place yet.

At the high-speed dedicated playback time, be that the data of unit record on the recording medium of for example CD etc. are that unit is read with the zone according to header information with a GOP.At this moment, data are carried out demodulation by demodulator 21, and input buffer storage 22.Yet, when containing much information of I image and P image can not be read entire I image and P image within a certain period of time,, read final data for the zone of reading in the way.Shaven head jumps to the front end of GOP then, the data in the zone of only input ability input buffer storage 22.In this case, 23 of format decoder decoding I image and the visual zone of P, and export as the high-speed replay image.

Figure 36 is illustrated in the playback image that has only a high-speed replay under the situation that the I among the GOP is visual and the P image is reset.In this case, when the data volume of I image and P image greatly to not reading entire I image and P within a certain period of time when visual, keep intact just to export through the data that make a last width of cloth picture and can synthesize the high-speed replay pattern.Figure 36 representes the situation that

zone

3,4 and 5 can not be read among the P4 of nGOP.In this case, the data of a last width of cloth picture are kept intact.

As stated; Because the such as shown in Figure 35 ground of the order of record I image is to roll up in the unit with GOP; Even have only under the situation that a part of zone of I image can decode at special-purpose playback time, the zone that can not decode is not concentrated on the fixed position of picture yet.

Then introduce embodiments of the invention 9 with reference to Figure 37.Figure 37 is the block diagram that the record of the encoding digital video signals processing unit in expression digital video record and the reproducing device is surveyed, and wherein the DCT piece is divided into low frequency range and high frequency region level, makes and has only low frequency range to be positioned at the front end of GOP.In Figure 37, counter, 61 presentation format encoders and the 65 expression input terminals of event number and size of code are calculated in numbering 51 expression buffer storage, 52 expression subtracters, 53 expression DCT circuit, 54 expression digital conversion circuits, 55 expression variable length coders, the reverse digital conversion circuit of 56 expressions, the reverse DCT circuit of 57 expressions, 58 expression adders, the moving compensation prediction circuit of 59 expressions, 60 expressions.

Introduce the working condition of equipment below.Video data to be imported is a horizontally interlaced image, and its effective picture dimension is level 704 pixels and vertical 480 pixels.Here; The working condition of subtracter 52, DCT circuit 53, digital conversion circuit 54, variable length coder 55, inverse quantizer 56, reverse DCT circuit 57, adder 58 and moving compensation prediction circuit 59 be with traditional embodiment in corresponding intrument identical, relevant their explanation is just omitted.

The existing working condition that variable length coder 55 is described with reference to Figure 38.Figure 38 representes the data arrangement of DCT coefficient in the DCT piece.In Figure 38, low frequency component is positioned at upper left, and the data of the DCT coefficient of high fdrequency component are positioned at lower right-most portion.In the data of the DCT coefficient of being arranged in the DCT piece, be encoded into variable-length code (VLC),, and export format encoded device 61 to as the variable-length zone until the data (the for example dash area among Figure 38) of ad-hoc location (end of incident) coefficient.So, the data of the DCT coefficient after the data of the DCT coefficient of above-mentioned position are carried out variable length code.In other words, the data in spatial frequency zone are encoded through cutting apart.

Boundary line between low frequency range and the high frequency region is called flex point.Flex point is set at the predetermined code amount of the supposition low frequency range that shaven head can arrive when special playback.Variable length coder 55 is divided into low frequency range and high frequency region to the DCT coefficient according to flex point, to export format encoded device 61 to.

Place, boundary line in incident confirms the code area.It goes without saying that can make with other method definite.For example, can confirm the code area at the boundary of the incident of some.Use by quantizer 54 and reach the difference value that carefully quantizes with thick quantification, data are divided through the thick quantized data that quantizes.In addition, can be to buffer storage the difference image between the image of the thinning visual and original resolution of dredging that the image of half degree is encoded and resolution being recovered from half degree of its resolution being encoded and data being divided.In other words, data are divided the division that is not limited to frequency zones.It goes without saying that, can divide the high efficient coding data of image with the division of quantification and spatial resolution.

More important data as image during at this moment, by frequency partition are low-frequency datas.When division is that in order to encode, data stand thick quantification with quantification agent timesharing.When data are divided with spatial resolution, the thinning image of dredging (thinned) is encoded.Through only those significant datas just being decoded, can obtain the decoded picture that people just feel easily.In this way, a kind of coded data efficiently can be divided into more basic significant data and other data (this process is called stratification), adds the error correction mistake, modulates; So that be recorded on the CD.

In this way, owing to only divide the low frequency component of I image and P image, when special playback, only read and these low frequency components of resetting have just reduced the data volume that special-purpose playback time is read widely.As a result, just shorter from the time of medium sense data, when jumping search, can realize the high-speed replay that steadily moves.And, when having only I image and P image to arrange, just can read the data of the low frequency component of I image and P image easily from CD, so that decoding with continuation mode.In this case, compared with the whole zone of arranging I image and P image at the GOP front end, just can form more effective data structure as long as extract and arrange low frequency component.

Just introduce the working condition of format encoded device 61 below.Figure 39 is the block diagram of presentation format encoder ruuning situation.When coding work starts, judge whether coding mode is in layered model at the beginning.When pattern was not layered model, information was inserted into system flow, and this information representation pattern is non-layered model, to obey traditional flow structure.Under the situation of layered model, confirm to have set the order header.Confirm that specifically order can survey the data of extension.Correctly write fashionablely when data,, make the I image be divided into low frequency component data and high frequency component data, with detection data length separately with the P image with regard to the front end of recognition image.

Simultaneously, be the length of the data of each Image Detection B image.In addition, arranging under the low frequency component data conditions of I image and P image, making the packet of a recording address information immediately following the front end of GOP.In this packet, comprised the high fdrequency component and the visual address information of ten width of cloth B of low frequency component, I image and four width of cloth P image of I image and P image, thereby write down the data length of each data.

Therefore, according to this data length, can obtain the front position of each data flow, as relative address with respect to GOP header front end.The low frequency component and the remaining data that contain packet, I image and four width of cloth P image of this address information are sequentially arranged, to format.

Wherein, the affirmation of the pattern surveyed on the extension surveyed of above-mentioned order header relates to the affirmation of the description of priority flex point on determined affirmation of surveying mode initialization in the sentence structure of MPEG2 among Figure 40 and the section header.The priority flex point is arranged in the predetermined quantity incident place (corresponding to above-mentioned flex point) of Figure 40, and relates to the low frequency component that expression divides and the data of the boundary's warp between the high fdrequency component.

When scrambling mode is assumed to be " 00 ", represent that follow-up bit stream is the bit stream that data are cut apart, also the expression bit stream that is divided into low frequency component and high fdrequency component is continuous.When the B image was made up of low frequency component so that can not be produced high fdrequency component, the B image just will not be divided.

An instance of the bit stream that produces by this way is shown in Figure 41.Figure 41 A representes the bit stream without layering.When bit stream during with the circuit layering shown in Figure 37, bit stream is just cut apart and layering shown in Figure 41 B suchly.When considering this special-purpose the playback data arrangement being become array, I image and the visual low frequency component of P just are arranged in the front end of GOP shown in Figure 41 C.

Figure 41 D is illustrated in address information data arrangement under such situation that is included in dedicated packet shown in the flow chart among Figure 39.In this case, as stated, address information can be with representing with respect to the relative address of GOP header front end.Yet address information can be that the mode of each visual front end is represented with which byte of which packet.It goes without saying that address information also can be represented with the sevtor address on the CD.

Figure 42 representes to comprise in the dedicated packet instance of address information.When elementary bit stream (the being called PES) packet through the data packetize was used as dedicated packet, bit stream ID was expressed as BF (sexadecimal notation).After having described data packet length, byte MSB is set at 1, and bit thereafter is set at 0, makes code and all initial codes (initial code of packet and the initial code of bit stream) seem inequality.So with remaining six bits describe layered model, layering kind, special-purpose playback time the visual kind that will use and initial address number etc.

, describe the address information of 21 bit long, thereby can represent GOP data volume up to the maximum length of 2M byte thereafter.Yet 100 (representing with binary law) were inserted the first tribit of 21 Bit datas, made preceding 24 bits 000001 (sexadecimal notation) of these data and above-mentioned initial code seem inequality.Here, initial address comprises the initial address of I image low frequency component, the initial address of the low frequency component of four P images, the initial address of the high fdrequency component of the high fdrequency component of I image, four P images, and the initial address of ten B images.In addition, record the sevtor address on the CD of data of forward and backward GOP in the interpolation, so that when special playback, shaven head is beated.

When the parity check bit of I bit is added to 21 bit addresses, just improved the reliability of data.In this case, 10 (binary representations) can be added to the front end with respect to 21 bits+1 bit.In addition, consider the special-purpose high speed multiple of resetting, during the address of and back one GOP regional and last when the sector that adds some front and back GOP, just can enlarge the special-purpose high speed multiple of resetting with regard to change.In addition, this is illustrated in special use 2 packets of packet of PES and has described address information.It goes without saying that address information can write on other user area of the special-purpose descriptor (descripter) of programming flowchart etc. for example etc.

Introduce the playback side of embodiment 9 now according to Figure 43 and 44.Figure 43 is the block diagram of decoded digital video signal processing section.In Figure 43, number 71 representation programs stream header detector, 72 expression PES packet header detectors, 73 expression video bits flow-generators; 74 expression data rearrangement devices, 75 presentation address memories, 76 expression mode conversion switch, 77 expression variable-length detectors; 78 expression switches, 79 expression inverse quantizer, the reverse DCT circuit of 80 expressions, 81 expression adders; 82 expression prediction data decoding circuits, 83 expression frame memories, and 84 expression decodable code determinants.Figure 44 is the sketch map of the operation principle of expression Figure 43.

The working condition of Figure 43 is described according to Figure 45 below.Figure 45 is the flow chart of expression playback time format decoder working condition.Stand the detection of program flow header from the bit stream of ECC output, so that be divided into each PES packet.In addition, detect the PES packet header, contain the dedicated packet and the video packets of data of address information with difference.

Under the situation of dedicated packet, address information contained in the packet extracted store.Simultaneously, under the situation of video packets of data, then extract the bit stream of video data.Here, under the situation of normal playback,, from the bit stream of video data, extract the data of low frequency component and high fdrequency component, thereby array data and output playback are visual again for I image and P image.Simultaneously, at special-purpose playback time, only extract the low frequency component of video data and reset.Here, behind the playback low frequency component, allow shaven head to jump to the front end of back GOP.

In this case, when describing these addresses at video flowing, address information just extracts and storage after converting bit stream to.Therefore, under situation about describing with the special-purpose descriptor of programming flowchart, with the extraction of detection level and the storage address information of programming flowchart header when address information.It goes without saying that address information can be relative address or absolute address`.

In fact, the mode signal input pattern change over switch of resetting continuously etc. such as the search and normal of jumping 76.Simultaneously, amplify by amplifier, thereby use the phase locked clock replay signal of process by outputs such as PLL from the replay signal of CD etc.Then carry out discriminated operant for digital demodulation.So after carrying out correction process, program flow header detector 71 has just obtained the data message after the head.

In addition, PES packet head detector 72 for example detects the address information in each visual address information described in special use 2 packets (private2packet) of PES packet and special-purpose data of resetting, and this information deposits addressed memory 75 in.Here, will be used for the PES packet of audio frequency, the PES packet that is used for the PES packet of character for example etc. and is used for video is classified, so that is had only video packets of data to export video bits flow-generator 73 to.

Here, video bits flow-generator 73 is wiped the additional information from the PES packet, constitutes bit stream.Specifically, give cancellation such as the data of all kinds of control codes and time mark (stamp).After this, according to the address information that from addressed memory 75, is obtained,, arrange bit stream when the normal playback again by data rearrangement device 74 with the output of mode conversion switch 76.

The output of mode conversion switch 76 (control signal) is transported to data rearrangement device 74 and decodable code determinant 84.Through obtaining control signal, data rearrangement device 74 is from being divided and the low frequency component of layering and the data before the high fdrequency component reconstruction division.Otherwise, only to variable-length decoder 77 output low frequency components.In other words, when normal playback, each low frequency component and high fdrequency component combine, and the equipment that makes is worked with the mode of the order data rearrangement of common image.At special-purpose playback time,, only export the low frequency component of I image or the low frequency component of I image and P image according to the high speed multiple.

Special-purpose playback time only allowing through low frequency component does not adopt time mark.By contrast, variable-length decoder 77 extracts the border by the incident in the represented low frequency component district of the priority flex point of section header together with decodable code determinant 84, makes decodedly until the data on border, and exports switch 78 to.This switch 78 connects into and when normal playback, can not insert zero.Simultaneously, at special-purpose playback time, with decodable code determinant 84 control switchs 78, the high fdrequency component district when making the zero insertion special playback after the priority flex point.

According to Figure 44 above-mentioned working condition is described now.With reference to Figure 44, when cutting apart flex point and be E1 to E3, E1 to E3 deposits low frequency component stream in.E4 to EOB then deposits high fdrequency component stream in.In low frequency component stream, there are the low frequency component data in the back DCT piece behind the E3.

Here, when normal playback, data rearrangement device 74 flows extracted data E1 to E3 from low frequency component, and from high fdrequency component stream extracted data E4 to EOB.And data rearrangement device 74 is extracted data respectively, and order is rebuild the DCT data.By contrast, at special-purpose playback time, data rearrangement device 74 extracted data E1 to E3; Then carry out length-changeable decoding by variable-length decoder 77; Decodable code determinant 84 detects the priority flex point, makes the dash area among zero insertion Figure 44, so that only constitute the DCT piece with low frequency component.

The data based moving vector that converts the DCT piece to is decoded.Here, because decoding is identical with conventional example, so omitted the explanation of relevant moving vector decoding.Yet the used benchmark of decoding P image is decoded with the I image or the P image of low frequency component decoding through only adopting when special playback.

With the data block is the data input frame memory 83 that unit decodes.Here, frame memory 83 recovers image with the former orderliness of gop structure, and through exporting from the conversion of block scan to raster scan.By the way, frame memory 83 usually can with the memory adapted that is contained in the prediction data decoding circuit 82.

The code area is limited to the boundary of incident, still it goes without saying that and also can use other method limited boundary, and in other words, except frequency zones is divided, the high efficient coding data of image can also be used and quantize to divide, perhaps spatial resolution division.

At this moment, are low frequency range data as the prior data of image under the frequency partition situation.Quantizing under the situation of dividing, data relate to the coded data through thick quantification.Under the situation with the spatial resolution dividing data, data relate to through with the thinning thin data that picture coding obtained.In this case, in the playback image of only decoding with these data item, the zone definitions that people aware easily is a significant data.In other words, high efficient coding data are divided into basic significant data and reach not significant data (this process is called layering) so, when making data from optical disk data reproducing, have only basic significant data to reset at special-purpose playback time.

Embodiment 9 has described the situation corresponding to the record side of playback side.This considers that also record and reentry are combined into one group, for example under the situation of hard disk etc., only considers playback side, and the principle that presupposes data and be according to general obtainable laser disk etc. writes down.

Introduce the embodiment of the invention 10 below.Figure 46 is the block diagram of expression according to the register system of the digital video signal recording of the embodiment of the invention 10 and reproducing device.Same parts or corresponding component among Figure 46 in the identical number table diagrammatic sketch 37.Device and 61 presentation format encoders are reset in numbering 65 expression input terminals, 51 expression buffer storage, 52 expression subtracters, 53 expression DCT circuit, 54 expression quantizers, 56 expression inverse quantizer, the reverse DCT circuit of 57 expressions, 58 expression adders, the moving compensation prediction circuit of 59 expressions, 55 expression variable length coders, 62 expression zones.

Figure 47 is the circuit block diagram of expression according to the playback system of the digital video signal recording of the embodiment of the invention 10 and reproducing device.Identical parts or corresponding components in the identical number table diagrammatic sketch 43 among Figure 47.Number 71 representation programs stream header detector, 72 expression PES data and comprise that detector, 73 expression video bits flow-generators, 85 expression zones reset devices, 75 presentation address memories, 76 expression mode conversion switch, 77 expression variable-length decoders, 79 expression inverse quantizer, the reverse DCT circuit of 80 expressions, 81 expression adders, 82 expression prediction data decoding circuits and 83 expression frame memories.

Introduce the working condition of embodiment 10 below.Digital video signal from input terminal 65 is imported with behavior unit, is transported to buffer storage 51.Identical with above-mentioned instance from the working condition of buffering memory 51 to variable length coder 55, relevant its explanation is just omitted.

For being the I image the bit stream of vision signal of unit output with GOP from variable length coder 55, device 62 data rearrangements are reset in the zone, make to be positioned at the front end of the area arrangement of picture core at bit stream.Here, the I image is divided into three zones as shown in Figure 48.Data definition corresponding to regional 1-3 in the I image is I (1), I (2) and I (3).Yet each zone shown in Figure 48 is the set of a plurality of MPEG section layer.In Figure 48,

zone

1 and 3 is made up of 6 sections, is made up of 18 sections for regional 2.

In fact, the section header that device 62 detects the I image of bit stream is reset in the zone, divides into three districts to each sheet, and is shown in figure 48, thereby every district is formed a bit stream, for resetting bit stream in each district.In other words, shown in figure 49, by being reset, like this, bit stream is with I (2), the front that is arranged in order on GOB of I (3) and I (1) in this unit, district for bit stream.The bit stream of this rearrangement is exported the format encoded device 61 of row GOP unit.

Next, explain the operation of format encoded device 61 according to Figure 50.Figure 50 is a flow chart, and the algorithm that turns to video data format the PED bag in the GOP unit is described.Under screen center's part mode of priority situation, bit stream image header and pictorial information that detection will be imported.Under the situation of I image, the core of drawing I (2) shown in Figure 49, I (3) and I (1) is extracted, and detects each data length, and the data length in each district of detection changes the binary number of 24 bit wides into, processes address information.On the other hand, data length is to be detected with respect to P and B image in visual unit, and like this, data length is converted into the binary number of 24 bit wides (3 hytes), processes address information.

Formatting unit is collected INADD information and video data bit stream, is divided into two types of PES bags.In other words, PES bag that only has address information and the PES bag that only has audio frequency have been constituted.

Therefore, as shown in Figure 6 under the situation that a GOP is made up of 15 frames, 17 types of images are arranged as address information, for example, 3 kinds of I images, 4 kinds of P images, 10 kinds of B images.And, the two types of address informations (absolute address` on the dish) that guide GOP and follow GOP are arranged as special playback address information constantly on dish.Being collected in the bag of the address information of these clauses and subclauses, and be formatted as the PES bag.In fact, these address information projects are collected in the bag, and are formatted as the special I bag of the PES bag shown in Figure 51.In Figure 51, guiding GOP is arranged in the front of wrapping data with the absolute address` of following GOP on dish.Then, each visual address information is arranged successively.Yet, each address information has been specified the long address information of 3 hytes (24), so the length of bag is 57 hytes.

During this period, with regard to a GOP part of the bit stream beyond the data of address, bit stream is formatted as PES bag (video packets), and method is to be divided into a plurality of packets to bit stream, and adds header information such as synchronizing signal.

In addition, format encoded device 61 is divided into many PES bags to the bit stream of the voice data of input, constitutes the MPEG2-PS system flow with video data PES bag.In fact, shown in Figure 52, the bit stream of GOP of video data part and audio data bitstream are cut apart and are arranged in a plurality of packets of same group.In this case, one represent the bag of aforementioned addresses information to be arranged in the preceding bag of system flow of Figure 52.Then, constituent apparatus as follows promptly partly contains the formation that encapsulates of bit stream in the screen center of I image.

Below, explain the operation of playback time in conjunction with Figure 47.In Figure 47, because the operation of program flow header detector 71, PES packet head detector 72, video bits flow-generator 73 and mode converter 76 is identical with conventional method, so omit explanation.

In the video bit stream of decoding, be positioned at the front of bit stream in screen center's partial data of I image.The data length of device 85 according to the bit stream I (2), I (3) and the I (1) that export from addressed memory 75 reset in the district, and presses the order reorganization I pictorial data of I (1), I (2) and I (3).The bit stream of reorganization is input to variable-length decoder 77, is decoded into blocks of data, moving vector etc.Because behind the length-changeable decoding to operate in normal playback constantly identical with conventional method, therefore, its explanation is omitted.

In high-speed replay, as stated, be to distribute to group system stream to a GOP data division; So think a kind of method arranged; When shaven head when coiling reading of data, just jump to the address, front of each GOP, pick and place pictorial data with read-only at the I of system flow front; Like this, shaven head jumps to the front of follow-up GOP.In this case, detect have one be arranged in the address information recording of system flow front the PES bag, come the control panel driving through the address on the dish of the follow-up GOP that decodes and the address information of I image.

Under the situation of Fig. 6, when all the I images among each GOP can read, can realize 15 times high-speed replay in a frame.When the I image in each GOP reads, can realize 7.5 times high-speed replay in two frames.In this mode, during high-speed replay, the time of reading the last data of dish is shorter.

From recording medium, for example CD etc. is gone up under the situation of reading of data, even know preceding address, and also rotation stand-by period of existence dish, in this time, shaven head will the jump to physical record position of data.When vision signal during with variable rate encoding, the amount of information of I image is indefinite, and the time of reading the I image also is variable.Therefore, the speed of high-speed replay is high more, and the time of reading the last data of dish is short more.Because the stand-by period of disc spins is indefinite, can not stably read the whole data of I image.

In embodiment 10, at high-speed replay constantly, for being recorded in the upward data of GOP unit of recording medium (CD etc.), shaven head jumps to this GOP data recorded in finite time unit.Like this, read the data division of I image from dish.As if in this case, the whole data of I image can not be read, shaven head jumps to the front of follow-up GOP.In other words, shaven head jumps to the preceding address of each GOP in a special time unit, to read data as much as possible from the front of system flow, jump to the front portion of follow-up GOP then.

In this case, the PES that comprises the PES bag of dish data address and follow-up GOP address and comprise I image center partial data encapsulates the previous section that is arranged in system flow.Therefore, even constantly can not read under the whole data conditions of I image when resetting in special use, address and I image center partial data can be decoded as address and the data that control panel drives on the follow-up at least GOP dish.

Reset constantly in special use; Have only under the decoded situation of screen center's part ability; Have only and can be outputed to variable-length decoder 77 by the data of zone rearrangement device 85 decodings, like this, the video data of length-changeable decoding is imported into frame memory 83 through inverse quantizer with reverse DCT.Simultaneously, district's information that 85 of zone rearrangement devices can not be decoded is input to frame memory 83.About district that can not be decoded, the dateout former state of former frame keeps and is exported.

Figure 53 shows an example of the image of resetting, and wherein, high-speed replay is to accomplish to the I image among n+4 the GOP through n the GOP that only reset.Figure 53 A shows a kind of situation, and the entire I image can be decoded.The situation of Figure 53 B is that

district

2 and 3 can be decoded.In district 1 that can not be decoded, the value of former frame is preserved and output by former state.In addition, the situation of Figure 53 C is to have only district's 2

abilities decoded.In district

1 and 3, the value of former frame is preserved in the same old way.

In general video signal recording and replay device, the form of taking is that the I image is recorded in the record frame unit constantly.On the contrary, at Figure 52, the district that is positioned at outside screen center part, the I pictorial data (being divided into three parts) is arranged in the front of GOP because of having priority.Therefore, even can under situation about reading from dish with definite time is constantly reset in special use, be exported in the playback screen of screen center's part at least in the district that has only the I image area.

As stated; In embodiment 10 (Figure 52); About being used for the special-purpose I image of resetting, the data placement that is positioned at central region is in the front of a GOP, and this district has priority and is recorded on the recording medium; Even make the district 2 that is positioned at screen center's part under the very high situation of the speed of high-speed replay, have the playback priority, be easy to see the content of high-speed replay image.Accomplish special-purpose playback time, shaven head jumps to the front of GOP in confirming time quantum, the result, and the picture of output is with predetermined being updated at a high speed.

By the way, the preceding embodiment of chatting can constitute as follows, resets the data in district that constantly can be decoded by all output in special use, for data can not be decoded the district, the data of former frame intactly preserve.Yet, have only screen center's part to reset and reset constantly in special use.

In this case, the data of the I area of image 2 that 85 decodings of device read from dish are reset in the

zone.District

1 and 3 about its data can not be decoded for example, covers the high-speed replay image with output frame memory 83 with grey data.

Figure 54 has also shown a playback image, wherein, has only n GOP to be used for high-speed replay to the district 2 of the I image of n+4 GOP.In Figure 54, the

district

1 and 3 at screen two ends is covered by grey data.Even the stand-by period very little, that dish rotates of the amount of information at the I image is very short, there is time enough to read under the data conditions in district 1 and 3.Do not decode with the data in district 3 in district 1 yet.

This is because can be read if distinguish 1 and 3 data, and on screen, export; Under the situation that

district

1 and 3 is not updated within a certain period of time; The high-speed replay image nature that becomes, therefore, only in screen center's part of I image at special-purpose playback duration playback time; The district that upgrades is constant, the visual nature that just becomes of resetting.

In the above embodiments, have only district's (these districts can be decoded at special-purpose playback time) of I image center part to be shown to cover the two ends of screen.Yet the core of screen can expand to a screen size and exported.

In this case, in frame memory 83, the data in the district 2 of decoding are expanded the size of a screen, shown in Figure 55.Yet under the situation of Figure 55, the core in the district 2 that is surrounded by dotted line expands to 2 times of sizes by means of the linear interpolation of level and vertical direction.In other words, under the situation of Figure 55, the size of the part that is centered on by dotted line is vertical 240 lines of level 360 pixel X.This imaginary point part is expanded to the size of a screen, vertical 480 lines of level 720 pixel X by linear interpolation.

Therefore, have only when screen center's partial data decodedly when expanding to the size of a screen to it in special repetition the time, the district of dateout diminishes.In this way, can eliminate the part that the screen two ends cover, this part only is only significant when screen center's part is exported.

In the above-described embodiments, have only the screen center of I image partly to have priority to be arranged on the bit stream.Yet other structure also is possible, and for example, screen center's part of P image and screen center's part of I image all have priority.In this case, the data placement of the screen cover core of P image is after the bit stream of I image.

In the above embodiments, pictorial data converts bit stream earlier into, and then in the unit, district, resets.Yet pictorial data possibly needn't be reset after converting bit stream into and before converting bit stream into, reset.

Figure 56 shows the flow chart of the replayed portion of embodiment 10.The process front of flow chart was described, omitted at this.

The recording section of embodiment 10 is similar to replayed portion and is described.Also has a kind of situation, the thing of the record and a pair of similar hard disk of formation of resetting.It is also conceivable that a kind of situation supposes the similar present pocket dishes of record data (CD) situation when replayed portion.Obviously, the data of the lower upright position, 8 lengthy motion picture districts of the section opening code through using the section head can realize in prediction data decoding circuit 82 and frame memory 83 that the on-screen data of unit, district resets.

Embodiment 11

Below, embodiments of the invention 11 will be described.Figure 57 has shown the encoding digital video signals processing unit in a digital video signal recording and the replay device, and wherein, frame DCT data hierarchy is low frequency range and high frequency region.Block diagram when Figure 57 is record, at this moment, screen is divided into a plurality of districts, and the screen center of low frequency range partly has priority to be arranged in the front of GOP.In Figure 57, device is reset in reference number 62 expressions one zone.Similarly partly all use identical label among Figure 57, do not state at this with corresponding part.

Below, explain the operation of this device.The video data of importing comprises effective screen size, vertical 480 pixels of level 740 pixel X.Motion compensation and DCT are used for high efficient coding is added to pictorial data.Here, it is all identical with embodiment 9 with layering operation before that data are divided, and do not repeat at this.

Embodiment 11 is with the identical point of embodiment 9, can with restrictive condition, spatial resolution and with divide the relevant frequency zones of layering and cut apart data.In embodiment 11, by data rearrangement device 62 further divide and the significant data of layering as embodiment 10 to screen each differentiation cut, make screen center partly have priority to be arranged in the front of GOP.In other words, data are divided into significant data and inessential data, and it can be recorded on the dish by the priority that is predetermined in the district.

In this method, the low frequency component of I image and p image is cut apart, and makes screen center that the priority in the arrangement partly arranged.Have only the screen center's part when these low frequency components to be read constantly and playback time in special playback, the data volume that constantly reads in the special use playback just can reduce greatly.Therefore, the speed of reading & recording medium can allow for tens times even tens times and realize the retrieval of leapfrog as quick as thought.

Here; Screen center's portion arranged of the low frequency of screen is in the front at GOP; The data of P image are followed the data back of screen peripheral part in the low frequency range of I image; As a result, because the screen center's part in the low frequency of the I image only reset, so can realize the high-speed replay of tens times even tens times.And, screen center's part of the I image and the low frequency component of P image, concerning special playback, data volume is very little, so can be easily read and decode from coiling in screen center's partial data.Like this, can realize high-speed replay with several times of speed.In other words, because as far as I image and the visual low frequency component of P, screen center's partial data amount is less than the data volume of whole low frequency component, so can be to realize resetting than embodiment 9 faster speed.

Below, the operation that device 62 and format encoded device 61 are reset by the district is described, Figure 58 is its flow chart.Beginning, i.e. coding beginning, the section header of the I image of detection low frequency component part, each sheet is divided into three districts, and is shown in figure 48.Then, prepare the bit stream in each district, with the bit stream in each district that recombinates.In other words, to each district's recombination data, with the order of low frequency range I (2), low frequency range I (3) and low frequency range I (1) bit stream is arranged in the front with the visual relevant GOP of low frequency range I, like Figure 49.

Then, under the situation of screen center's part mode of priority, the visual header of the bit stream that detection will be imported is to detect pictorial information.Here, under the situation of low frequency range I image, extract the low frequency range part I of screen center (2), low frequency range I (3) and low frequency range I (1) with the detection data length, thereby prepare address information from the data length in each district., under the situation of P image and B image, in image cell, detect data length therebetween, thus the preparation address information.In non-picture core priority pattern, operation is like embodiment 9.

Below, judge hierarchical pattern.In the non-graded pattern, represent that this pattern is the information insertion system stream of non-graded pattern, follows traditional flow structure.Under the situation of layered model, confirm the setting of sequence header.That is, the affirmation sequence can be measured expanded data.Under the situation that data correctly limit, with the front of visual header identification image, like this, be extracted in the I image and the visual low frequency range data of P of recombinating in the screen area, and the length of inspection data.Simultaneously, to the visual data length of every frame image inspection B.

Advance you; Preparation one bag; Wherein, in screen center's portion arranged of the visual low frequency range of I image and P under the situation of GOP front, a recording address information; This bag comprises I. image and visual low frequency range screen center's part partly, the visual high frequency region part and the visual address information of B of peripheral part, I image and P of screen of P, to write down the data length of each data.Therefore, the anterior locations conduct of each data flow obtains with relevant address, GOP header front.

In Figure 59, show the bit stream of preparation by this way.Shown in Figure 59 c, I image of in the unit, district, recombinating and P image low frequency range are rearranged the front at GOP.The situation that Figure 59 D representes is that the low frequency range data of reorganization are arranged in the bag, so address information is arranged in special 2 bags, shown in the flow chart of Figure 58.In this case, address information can replace with the relevant relative address in front aforesaid and the GOP header.In addition, address information can represent like this that also promptly, which byte in which bag drops on each visual front.

It goes without saying that address information also can be represented with the sector address that is added on the dish.

Figure 60 shows address information and is included in the situation in special 2 bags; In this case; The PES bag is set up as special 2 bag stream ID; So described layered model, hierarchical categories, at special use the reset visual classification used constantly and number number of start address, wherein, start address refers to: the start address of the start address of the screen peripheral part of the start address of screen center's part of the low frequency range of I image, the low frequency range of I image and the B image of reservation.

On dish, add the sector address of guiding and follow-up GOP, shaven head is reset in special use constantly jumped over.In this case, consider special-purpose playback high speed constantly doubly, except guiding and follow-up GOP address, also add the sector address of several forward and backward GOP, then the variation doubly of the high speed of special-purpose playback will strengthen again.Regulation in also explanation, address information special 2 bags in the PES bag.Obviously, address information can be at programming flowchart, stipulates in the special definitional part such as other user districts.

The existing replayed portion of describing the device among the embodiment 11 according to Figure 61.Figure 61 is the block diagram of decoded digital video signal unit.Similarly partly all mark among the figure, do not repeat at this with identical label with corresponding part.

Below, explain Figure 61 according to Figure 62.Figure 62 is a flow chart, shows the flow chart of format decoder constantly of resetting.Program flow header in the bit stream of detection ECC output, and, each PES bag separates bitstream.The header that detects the PES bag of bit stream comprises the special bag of address information and video packets with differentiation.

Under the situation of special bag, extracting also, storage package is contained in the way address information in the bag.Simultaneously, under the situation of video packets, extract the bit stream of video packets.Under the situation of special bag and normal playback, or under the situation of video packets, from the bit stream of the I image and the video data of P image, extract the data of low frequency component and low frequency component, thereby recombination data is visual to export playback.

Therebetween, under the situation of special bag and normal playback, the time that when beginning, just the judges whether whole low frequency I that is suitable for resetting is visual.Under the situation that is suitable for resetting in the time, the time that further the judges whether P image that is suitable for resetting.Accomplish above-mentioned two or a judgement.Therefore, under the time is suitable for resetting the situation that low frequency I is visual and P is visual, playback I image and I image.Reset the low frequency P image of only resetting under the situation of low frequency P image at time be suitable for resetting whole low frequency I image but be not suitable for.And then, under the time is not suitable for resetting the situation of entire I image, screen center's part of playback low frequency I image.Under above-mentioned three kinds of situation, let shaven head jump to the front of next GOP.

Under the situation about in bit stream, stipulating in these addresses, after bit stream forms, extract and storage address information.Under situation about stipulating in the special definitional part of these addresses at programming flowchart, in the level of trace routine stream header, extract and storage address information.Obviously, address information both can be the relative address of program, also can be the absolute address` of program.

In fact, shown in Figure 61, jump over scanning, normal playback isotype signal continuously is input to mode converter 76 from microcomputer.Simultaneously, be exaggerated device from the replay signal of coiling and amplify, and, make Phase synchronization with the clock replay signal that PLL exports.Then, with the signal digital modulation, error correction, recovery routine stream.So the program flow header detector 71 that detects each program flow head obtains following the information of the data of header.

Each visual address information of regulation and special-purpose replay data (data that low-frequency data and screen area are arranged) detected by PES packet head detector 72, and information are stored in the addressed memory 75 in special 2 bags of PES bag.Judge whether that the PES bag is audio frequency PES bag, character PES bag or video PES bag here.Header under 73 cancellations of video bits flow-generator move from the PES bag, output bit flow.After this, according to the address information that obtains from addressed memory 75, the bit stream that 74 reorganization of data rearrangement device are exported from mode converter 76, and in normal playback, export this bit stream.

The output signal (control signal) of mode converter 76 is added to data rearrangement device 74 and decodable code determinant 84.Here, data rearrangement device 74 and layering and synchronous to the low frequency component and the high fdrequency component of each district's reorganization, and export this synchronous component.Simultaneously, only with the low frequency component data or only the low frequency component data of screen center's part are reset in special use and output to variable-length decoder 77 constantly.In other words, at normal playback constantly, the low frequency component of I image and P image is pressed the order reorganization in screen Shang Ge district.Then, low frequency component and high fdrequency component are synchronous, and device carries out work with by visual original order recombination data.Reset constantly in special use, changed output in the low frequency component district of the I of screen center's part image and P image.Reset constantly in the special use that only utilizes low frequency component, do not use PTS and DTS time mark.

On the contrary, the border that variable-length decoder 77 and decodable code determinant 84 extract incident in the low frequency component districts, these districts are by preferential discontinuous point mark, and are all decoded up to the data on border like this, and output to switch 78.The effect of switch 78 is not insert O constantly at normal playback.Reset constantly in special use, decodable code determinant control switch 78 makes in the high fdrequency component after O is inserted into the preferential point of interruption.

Decoding operation principle such as Figure 44 of low frequency component.Its explanation is omitted at this.At this constantly, identical to the reorganization of screen area with the explanation of embodiment 10.It is explained in this omission.

The code area is limited to the border of incident, need not say, event boundaries also can be limited additive method.For example, the code area also can be by the end point of giving the determined number incident separately, can also be through letting data carry out the coarse data conversion and obtain rough and meticulous digital translation difference and come partition data quantizing 54.Cutting apart of data is also available to the spatial resolution image half the because of thinization is reduced to, resolution from half level return to former level image and with the completion of encoding of the poor image of image with former resolution.In other words, the pictorial data of high efficient coding is cut apart to cut apart through (except that the dividing frequency district) quantize to distinguish the rate completion with the space.

Carve at this moment, under the situation of frequency division, the more important data of an image are the data of low frequency range.Under the situation of quantization, more important data are coded datas of rudenss quantization.Under the situation of spatial resolution partition data, more important data are that thinization picture coding is obtained data.In this case, only use the data playback image of these clauses and subclauses, the zone that is easier to accept of behaving has constituted than significant data.In other words, the data of a high efficient coding are divided into more basic, important data and not too important data, and in the moment of resetting from dish, the important data of fundamental sum are reset.

Corresponding to the recording section of replayed portion embodiment 11 is described.Has a kind of situation, the device of the record and a pair of similar hard disk of formation of resetting.Consider such a case, tentation data is coiled by record as CD.About the component reorganization to each district of screen, obviously, the method for Figure 54 of embodiment 10, the output screen shown in 55 is available.Obviously, if use the data of the section vertical position in the header of section, the reorganization on screen in the unit, district also can enough prediction data decode circuitry 82 realize in embodiment 11, having only the master data of I image to be cut apart by screen area with frame memory 83.Can use low frequency or other mode partition datas of P image.

Embodiment 12

Explain embodiments of the invention 12 with reference to Figure 63.Figure 63 is a block diagram, and encoding digital video signals processing unit in digital video signal recording and the replay device is shown.In Figure 63,

reference number

101 and 104 is preprocessors, and 102 and 105 is moving vector detectors; The 103rd, resolution converter, 106 and 107 is subtracters, 108 and 109 is DCT circuit; 110 and 111 is quantizers, and 112 and 113 is variable length coders, and 114 and 115 is inverse quantizer; 116 and 117 is reverse DCT circuit, and 118 and 119 is adders, and 120 and 121 is video memories; 122 and 123 is rate controllers, the 124th, and the resolution reverse converter, the 125th, the data reconstruction device is as data repacking device.Figure 63 shows first code device and second code device as an example.Specifically, subtracter 106 output first code device and the difference components of second code device in two cataloged procedures.

Below, explain the operation of embodiment 12.Video data is input to resolution converter 103 with the order of interlacing grid scanning.The video data filtering of 103 pairs of inputs of resolution converter and thinization, the repetitive noise of removal high frequency region.Figure 64 is the principle explanation of image being carried out conversion of resolution.For example, be under the situation of level 704 pixels and vertical 480 pixels in data, first filtering, thinization promptly has original half the resolution respectively, thereby changes the low resolution on-screen data into to 352 pixels of level and vertical 240 pixels then.

The on-screen data of this low resolution is input to preprocessor 104, transfers block scan to from grid scanning.The order of sending in the block scan device is the order of DCT piece.The coding of I image need not utilize the output signal of the frame memory of intraframe coding to accomplish interframe and calculate.

Under the situation of I image, as the not output of video memory 121 of the input of subtracter 107, so vision signal is through subtracter 107.These data are frequency component by 109 orthogonal transforms of DCT circuit.The data of this orthogonal transform are input to digital quantizer 111, from the digitlization of low frequency range sawtooth mode scanning sequency.The pictorial data that numeral transforms converts entropy coding into through variable length coder 113 and outputs to data repacking device 125.

Simultaneously, by digital quantizer 111 digitized data with the 115 reverse digitlizations of inverse quantization device.Pictorial data is reversed DCT circuit 117 and is converted into spatial component from the frequency component data back.The decoding of I image need not be used the intraframe coding output signal of frame memory to accomplish interframe and calculate.Therefore, under the situation of I image, because there is not input, so data are through adder 119 from the video memory 121 of adder 119.Another output of adder 119 is as the data that deposit video memory 121 in.At least, I pictorial data or I image and P pictorial data will deposit video memory in.This is because it is reference data that I image and P pictorial data are used at MPEG1 and MPEG2 normal decoder.

Video memory 120 input is from the output signal of the decoded data of adder 118, and the number result of the recovery pixel of accomplishing with resolution reverse converter interpolation pixel, to store the image decoding data with a weighted average.About weight, for for simplicity, a kind of situation is arranged, weight 1 is used the output as the resolution reverse converter, and weight O uses the output as adder 118.

The vision signal of output cushions in preprocessor 101, with from raster scan-convert into block scan.Then, subtracter 106 vision signal from video memory 120, carry out aforementioned low resolution and handle the signal data of (referring to the resolution residual components) and deduct.The resolution residual components is the frequency zones that can be quantized device 110 appropriate conversion from the low frequency range orthogonal transform by DCT circuit 108.These data are encoded to the entropy sign indicating number for variable length coder 112, and output to data rearrangement device 125.

During this period, digital quantizer 110 data converted reverse conversion in inverse quantizer, and reverse conversion is the data of space region in reverse DCT circuit 116.Adder 118 is from video memory 120; Through low resolution reverse conversion data of handling and the translation data addition that comes self-reversal DCT circuit 116; Obtain the decoded result of these data; As the example of data (rather than low resolution), it is formed by two-layer, i.e. high-resolution data and residual components data.This layer is by the frequency decision of conversion of resolution.Might use two conversion of resolution with three layers of this layer formation.The data that might prepare in the same way, any amount aspect with similar method.

About normal mpeg encoded, I image and P image are decoded, and are stored as decoded data, through the two-way prediction coding B image to I image and P image.In this method, I image and P image are encoded, and carry out the B image processing then.

Before chat I image, P is visual and B is visual encoding process are differentiated component and the high-resolution component is accomplished to low.In this mode, can constitute a sequence, wherein, low resolution components R (claiming the R component later on) and resolution residue difference component S are arranged side by side.Operated by data rearrangement device 125, like this, make the front of data placement at GOP, its position makes shaven head be convenient to seek.For example, data can be arranged to the sequence of Figure 65.When data were arranged by Figure 65, the L component occupied the half the of district, the low resolution of resetting component.The data volume of resolution residue difference component is less than non-resolution residual components, and data are layering effectively.In other words; Here, be provided with according to first code device and second code device that give the fixed condition coding and carry out effective layering, second code device is encoded to the remainder of first code device coding; For example, be to the video information coding beyond the video data of first code device coding.

Figure 65 shows data structure result's a example.In Figure 65, sequence a is the sequence that the encoding process process of embodiment 12 produces.Sequence b is the sequence that the cataloged procedure of another embodiment produces.Sequence c is the sequence that the cataloged procedure of an embodiment again produces.In sequence b, symbol L is a low frequency component, and H is a high fdrequency component.In sequence C, symbol C is the coding component of rudenss quantization, and A is the residual components of rudenss quantization.As shown in the sequence a of Figure 70, aforesaid operations is only to I image and the completion of P image.Have only these components can be arranged in the front of GOP regularly.

In this method, having only low frequency component to be arranged in regularly under the situation of GOP front, the L component accounts for overall ratio and descends greatly, and like this, reading speed is just share from the medium, easily realizes the retrieval of jumping.In addition, sequence a for example, when having only the R component of I image and P image to be arranged in the front of GOP regularly, executable operations like this, has only the visual low-resolution data of I image and P decoded.In the aforementioned embodiment, explain that thinization ratio is to 1/2 times of level, vertical 1/2 times in the situation about being directed against.Obviously, ratio value can be set at different values, but artificial preset proportion value can be added among the embodiment.

Coding mode comprises MPG1, MPG2 and JPEG etc.In resolution demixing, and needn't adopt the universal coding technology.This be because, reducing under the situation of resolution to the data coding, might be fully corresponding to the MPEG1 pattern-coding.In addition, in jpeg mode, a frame has constituted motion video to the stack of another frame.Therefore, even occupied in data under the situation of specific position of GOP, also possibly correctly decode.In addition, be presented for the explanation of two-level resolution, but, obviously, can utilize a large amount of layerings.Can encode to difference component as follows: with the digital coding of first code device of Figure 63 to the low resolution component; Output signal interpolation to this first code device; Before thinization pixel and interpolative data, the difference component of image obtains from subtracter 106; With the difference component code device difference component is encoded.

The frame of reading from video memory obtains from the prediction reference frame.Owing to there is low-resolution frames, will there be memory (comprising storage address) in data through regulating time shaft.Obviously, massaging device can be set, audio signal, header etc. and error correction signal additional information are added on the difference component.

Embodiment 13

Embodiments of the invention 13 make an explanation with Figure 66.In embodiment 13, the DCT piece is divided into low frequency range layer and high frequency region layer, has only low frequency range to be arranged in the GOP front.Figure 66 is the block diagram of an encoding digital video signals processing unit.In Figure 66,

label

126 and 127 is represented first variable length coder and second variable length coder respectively.Among Figure 66 with Figure 63 in similar or corresponding part with mutually label mark, do not repeat at this.

Below, the interpretation work process.This interpolate video data is one and has the for example data item of level 704 pixels and the effective screen size of vertical 480 pixels.Because the decoding of I image need not utilize the intra-encoded output of frame memory to carry out interframe and calculate, so video passes through and output.Video data orthogonally transforms into frequency component by DCT circuit 108, and transforms to block scan from low frequency region.Then, video data is transformed into block scan from low frequency region, is suitably quantized by quantizer 110.

The data recombination of the DCT coefficient in the DCT is shown in Figure 67.In Figure 67, low frequency component is positioned at top, the left side, and high fdrequency component is positioned at bottom, the right.Be arranged in the DCT coefficient data in this DCT piece; Low frequency range DCT coefficient data before the one ad-hoc location DCT coefficient data (for example; The dash area of Figure 67) carries out entropy coding through first variable length coder 126 (low frequency range extraction element), and output to data rearrangement device 125.The DCT coefficient data of 127 pairs of above-mentioned specific position DCT coefficient datas of second variable length coder back carries out variable length code.That is to say that in this mode, data are divided and encoded.

About the coding of moving vector and DC component, can only accomplish coding with first variable length coder.Second variable length coder does not need.This is because at normal playback constantly, the dateout of first and second variable length coders 126,127 can synchronously and be encoded.

The judgement of code area is to accomplish in the fixed position of DCT coefficient, also can use other modes to accomplish.For example, the also available fixing event number of coding is judged.In other words, the unit of variable-length code (VLC)-Huffman code being provided is an incident.The also available fixed number purpose incident (for example three is a unit) of giving in code area is set.In an example, at data rearrangement device 125 output bit flows, have the reorganization of the sequence b of Figure 65, have only when the first half of low frequency range is read, the low frequency range image can be reset.In the such sequence of the sequence b that arranges Figure 70, can judge the code area with the mode of variable-length.

Simultaneously, the data of quantizer 110 quantifications are carried out inverse quantization.Then, these data by reverse DCT circuit 116 reverse conversion to the space region data.The coding of I image need not utilize the output signal of the frame memory that has stood intraframe coding to carry out interframe and calculate.Therefore, under the situation of I image, there is not input from the video memory 120 of adder 118.So, allow data through adder 118.Adder 118 is used to being stored in the data in the video memory 120.

At least, I image will be stored in the video memory with the P image, that be because, will be under I image and the visual normal condition of P as in MPEG1 and 2, describing the visual reference data of B.

After constituting in this way, the ratio of L component reduces greatly, allows reading of data from the medium, makes and realizes the retrieval of jumping.Shown in the back, to have only when I image is arranged with the P image is fixing, apparatus of the present invention can works, and the data ability quilt that has only low frequency component is easily decoded.Because the data volume of high frequency region is littler than the data in other districts, so exist with the data of extracting low frequency range and these data before the data in all districts than realizing data formation effectively.

When allowing the I image to finish the visual coding of I through subtracter 106, according to the time with last P image of GOP with two-way prediction to the B picture coding.The data of the output of preprocessor 101 and reference frame storing device (arrow among the figure omits) each other relatively detect moving vector, judge prediction pattern and frame structure.With the judged result is the basis, and the data of the output of preprocessor 101 and reference frame storing device are the most consistent with each other, and the data that read the reference frame storing device are as direction of advance part and direction of retreat partial data from frame memory 120.Then, the data that read are by this way subtracted each other (this result is the P image and visual both the time residual components of I) with the output result of the preprocessor 101 of B image in subtracter 106.This time residual components is carried out DCT calculate, its result quantizes and variable length code.

Embodiment 14

Embodiments of the invention 14 are that explain on the basis with Figure 68.In embodiment 14, data are divided into DCT coefficient rudenss quantization component and the difference component layering (example of rudenss quantization component) of rough residue, thereby, be arranged in digital translation component roughly the front of GOP.Figure 68 is a block diagram, shows the encoding digital video signals processing unit.In Figure 68, reference number 128 is subtracters, the 129th, and adder.Among Figure 68 with Figure 63 in similar or corresponding part with identical label mark, do not repeat at this.

Below, explain the operation of embodiment 14.Effective screen size of the input imagery of inserting in it for instance, is level 704 pixels and vertical 480 pixels.The decoding of I image need not utilize the output signal of the frame memory of intraframe coding to carry out interframe and calculate.Therefore, under the situation of I image, video memory is input not, and subtracter 106 is just not input also, the result, and vision signal is through subtracter 106.These data are converted into frequency component by DCT circuit 108 quadratures, and are transformed into block scan from low frequency range.Then, quantizer 110 is accomplished suitable rudenss quantization, is reduced to the coded data amount less than half the.The data of this quantification are encoded to the entropy sign indicating number through variable length coder 112, output to data repacking device 125.

Simultaneously, quantizer 110 data converted are carried out inverse quantization (its result is as the result of rudenss quantization).The data of inverse quantization are delivered to another coding processing unit (in Figure 68, showing with the dot-dash frame table).Therebetween, data are reversed and are transformed into the space region data.The coding of I image is described here.Though not output of video memory 121, under normal circumstances, the coding result of this coding processing unit is stored in the video memory 121, like this, carries out moving vector in 102 pairs of data of motion vector detection and detects, and cover half formula and DCT block mode are given in decision.The position data that is suitable for judgment model is delivered to the input side that subtracts each other of subtracter 107.

DCT is carried out in output to subtracter 107, with the definite residue of the result of rudenss quantization and subtracter poor (being that the rudenss quantization residue is poor).Rudenss quantization residue difference is by fine quantization (consider the control of encoding amount, fine quantization is the identical level of normal encoding).With in inverse quantization, reverse DCT and decoding to accomplish variable length code when existing in the video memory 121.The position that the coding result of this coding result and rudenss quantization has determined necessary data and the header that will add etc.

As the example of this dateout, adopted the sequence C shown in Figure 65.Only through reading the decoded result that first half of GOP just can obtain through the image of rudenss quantization.In addition, because compare with the fine quantization data, rudenss quantization residue difference data is smaller, so compared with before the thickness quantized data, storing the rudenss quantization data of extracting, can obtaining more, data effectively constitute.

In another example, can accomplish variable process to the arrangement of the sequence C shown in Figure 70.Such formation, C component (carrying out the coding component of rudenss quantization) account for whole ratio and reduce greatly, thereby satisfy from the medium reading speed, enable the retrieval etc. of jumping.As hereinafter described, have only when I image and P image are arranged regularly, the operation of apparatus of the present invention makes that to have only the data of the visual process rudenss quantization of I image and P decoded.

When making I image finish the coding of I image through subtracter 106, according to the time with last P image of guiding GOP with two-way prediction mode to the B picture coding.The output signal of preprocessor 101 and the data of reference frame storing device (arrow among Figure 70 omits) relatively like this, detect moving vector, judge prediction pattern and frame structure.On the basis of judged result; The output signal of preprocessor 101 the most well with the reference frame storing device in data consistent; Data in the reference frame storing device of reading are as the forward part and reverse part of video memory 120, and the output result of the preprocessor 101 of data of reading like this and B image subtracts each other (this result remains difference component at P image and B both times of image) in subtracter 106.From video memory 121, read forward part and reverse partial data, the output signal of these data and preprocessor 101 subtracts each other in subtracter 107, is used for the orthogonal transform entropy coding.Identical process is also accomplished the coding of P image.

Figure 69 shows the coded data statistic.This illustrates the number of codes when the number of the frame of GOP is following: N=15, the cycle of I image and P image: M=3.Also illustrate at Figure 69, I image and P image account for 50% of integral body.When with resolution, frequency, when carrying out layering about the quantification of this part I image at least, the sign indicating number amount that reset further reduces, and the journey time of shaven head can shorten, thereby helps function such as realization jump retrieval.

Figure 70 shows the processing sequence in the above-mentioned situation.In Figure 70, the I of original picture image, the arrangement that P is visual and B is visual are encoded, and only the I image and the P image of aforementioned image are handled by

embodiment

12,13,14, and the B image are encoded, without layering.According to the processing procedure of embodiment 12 sequence visual to I and that the P image is handled is sequence b, and the sequence of and P image processing visual to I according to embodiment 14 processing procedures is a sequence C.

In each sequence, the formation of data be through data rearrangement device 125 each low resolution component (R), accomplish GOP was fixed and be arranged in to the I image component of high fdrequency component (L) and rudenss quantization component (C) and P image component front.About sequence a, the low resolution image of I image and P image can only be decoded with the low resolution component (in the sequence a of Figure 70, the core space part that data recombination becomes) of I image and P image, the result, and apparatus of the present invention can easily realize jump scanning.In fact, the data in non-core district and do not require arrangement by Figure 70.Obviously, the sequence arrangement data that constantly can press frame number at coding.

About sequence b, can use low frequency component (in the sequence b of Figure 70, the core space part after the data recombination) to form I image and the visual low frequency component of P, so device of the present invention can easily be accomplished jump scanning.About sequence C, only just can be to the rudenss quantization image decoding of I image and P image, so device of the present invention can easily be realized jump scanning with the rudenss quantization component (the core space part after the sequence C data recombination of Figure 70) of I image and P image.

For example, in the structure of Figure 63, comprise that the coding collar of preprocessor 104 is not used for the B image, therefore, the mode of operation of apparatus of the present invention is only to carry out digital coding with the coding collar that includes preprocessor 101.In the structure shown in Figure 66, all frequency components all use first variable length coder 126 to encode.In the structure shown in Figure 68, in digital quantizer 110, accomplish fine quantization, so that data are encoded.

Most desirably, resemble the front that master datas such as low frequency part data can be collected in GOP.Obviously, these data can be shifted a little.Overlapping with the front, unit that constitutes error correcting code.Arrange master data and error correcting code unit correspondingly also can on other embodiment, realize in the same manner by this way.

Embodiment 15

Embodiments of the invention 15 will be consulted Figure 71 and 72 and make an explanation.Figure 71 illustrates the structure arrangement of DCT piece, and the structure outline of the frequency component of bit stream is given an example in a piece.Figure 71 A illustrates macro block of whole DCT block structured, is made up of the DCT piece Y1-Y4 of macroblock header, luminance signal, the DCT piece U1 of color difference signal (B-Y) and the DCT piece V1 of color difference signal (R-Y).Figure 71 B illustrates a low frequency component about the macro block of the structure of low frequency component DCT piece.By macroblock header, the DCT piece UIL of luminance signal and DCT piece Y1-Y4, color difference signal (B-Y) and the DCT piece VIL of color difference signal (R-Y) form.

High frequency macro block shown in Figure 71 C is the structure about high fdrequency component DCT piece, is made up of luminance signal DCT piece Y1H-Y4H, color difference signal (B-Y) and DCT piece U1H and color difference signal (R-Y) and DCT piece V1H.Figure 71 D shows the principle of the bit stream medium frequency component data arrangement of a piece.The block diagram of Figure 72 A and 72B shows a decoded digital video signal processing unit and an operation principle thereof.In Figure 72 A, reference number 130 is the mode converters as mode conversion device, the 131st, and as the data rearrangement device, the 132nd of data repacking device, decodable determiner, the 133rd, variable-length decoder, 134 is switches.Decodable determiner 132 and switch 134 composition data operating means.Reference number 135 is inverse quantizer, the 136th, and reverse DCT circuit, the 137th, video memory, the 138th, adder, the 139th, reverse scan transducer.

Below, interpreter operation.The data of Figure 71 are the for example code structure combinations of 8 (bytes) of vertical direction.In each macro block, use the macroblock header provisioning information.This information refers to such an extent that be incremental address, quantitative calibration sign indicating number, moving vector, flag bit, macro block model etc.

The coded data of each DCT piece is followed after macroblock header.The method of these data of embedding is to constitute byte with bit stream, with each byte of sequence arrangement.Because each DCT piece has variable code length, the border between block boundary and header and data is not done in byte units.The border is in a situation in the middle of the unit byte and often occurs.The data of each piece have variable-length, and in a side that compares near macroblock header low frequency range are set.

These data are divided into low frequency component (L) and high fdrequency component (H), come the coded data shown in pie graph 71B and Figure 71 C through setting a sign indicating number amount with the irrelevant regular length of incident as maximum.(this incident provides the unit of a variable-length code (VLC), and under the situation of DC component, the DC component constitutes an incident, and under the situation of AC component, an incident that constitutes the executable operations length code of the DCT coefficient of non-O and working length.The completion of an incident is represented at the end of piece with a sign indicating number EOB).

Below, the operation of Figure 72 is described.Beginning from the mode signal input pattern transducer 130 of microcomputer etc., is being carried out the signal of indication jump scanning, or carries out and normally reset continuously.Simultaneously; Amplifier amplifies the replay signal from dish; Thereby and through dateout completion difference operation being carried out digital demodulation completion error correction; Said dateout is accomplished reproducing signals and is being obtained from outputs such as PLL through the clock signal of synchronised, then, from a certain system that forms by video signal data and audio signal layer isolate audio signal.Extract the bit stream of vision signal and be input to reset device 131 thereafter.

The output of mode converter 130 (control signal) is added to data rearrangement device 131 and decodable code determiner 132.Data rearrangement device 131 controlled signals connect these data again, are divided into L component shown in Figure 71 and H component then, only output to variable-length decoder 133 to the L component.Variable-length decoder 133 extracts the event boundaries in the L component district together with decodable code determinant 132.Part before the border is decoded and output to switch 134.The connection of switch 134 makes when normal playback, do not have O to insert.Switch 134 is by 132 controls of decodable code determinant, to the low frequency component of DCT piece input decoding.Simultaneously, constitute whole DCT piece, make O insert the high frequency side of DCT piece.

When decoding, the DCT blocks of data that constitutes with preceding method will carry out reverse DCT processing.Then, according to the reading of each visual situation control video memory 137 that will carry out addition by adder 138.Under the situation of I image, the output of adder 138 is passed through.Under the situation of P image, the P image is only corrected by the moving vector of the I image of desiring addition and P image.Under the situation of B image, the B image is only corrected by the moving vector of the I image of wanting addition and B image.

Information Control DCT pattern and the prediction pattern moving vector at this moment that obtains according to decoding to the header sign indicating number.According to the aforementioned processing process, the data of motion-compensated prediction are decoded, and are stored in the image memory 137.Image recovery is to the order of the original formation of GOP.Reverse scan transducer 139 is transformed into buffering and block search the grid scanning of visual output order.

Embodiment 15 proposes length and fixes, and has both made this embodiment be shorter than the jump of macro block or gives the data of measured length.Yet when being shorter in length than regular length, through each detection EOB, the L component is confirmed and can be removed.Obviously, both made at the L component data to be connected under the situation of subsequent block, and also can not produce what problem.Surpass when giving measured length in length, EOB can be added to the incident boundary, as the L component data.Obviously, also be provided with additional information adding devices such as audio signal, header etc. and connection sign indicating numbers, to be added to it on data of high frequency region, this point does not specify in the figure of previous embodiment.

Embodiment 16

Next, with reference to Figure 73 and 74 explanation embodiments of the invention 16.Figure 73 A and 73B are the block diagrams of encoding digital video signals processing unit.Its operation principle is described.In Figure 73 A, reference number 140 is rate controllers.Here, first variable length coder 126 and second variable length coder 127 are as code device.Adopt identical reference number with similar part among Figure 63 or corresponding part.

Below, explain its course of work.The interior input image data of inserting is transformed into block scan to raster scan with preprocessor 101 bufferings.The decoding of I image need not utilize the output signal of intra-encoded frame memory to carry out interframe and calculate.Therefore, under the situation of I image, the video memory of importing as subtracter 106 120 does not have input signal, and therefore, vision signal is through subtracter 106.

These data are transformed into frequency component by DCT circuit 108 quadratures, and are transformed into block scan with low frequency range, are suitably quantized by quantizer 110.The DCT coefficient data of the ad-hoc location in the data of this quantification outputs to data rearrangement device 125 according to preceding low frequency range data by entropy coding and through first variable length coder 126.

DCT coefficient after 127 pairs of above-mentioned specific positions of second variable length coder carries out variable length code.About moving vector and DC component coding, can only use first variable length coder at least.Even require EOB to be added on the L component and H component in a piece, so the border of L component is not there to be the rate variation of sign indicating number restriction.Through temporarily being arranged in the EOB sign indicating number boundary section of L component and H component, the border of L component is changed with a certain speed.

Simultaneously, the data that quantizer 110 quantizes are by inverse quantizer 114 inverse quantization, then by reverse DCT circuit 116 reverse conversion to spatial component.

The decoding of I image need not utilize the output signal of intra-encoded frame memory to carry out interframe and calculate.Therefore, under the situation of I image, owing to there is not signal to be input to adder 118 from video memory 120, data are through adder 118.The output variable of adder 118 is stored in the data in the video memory 120.At least the data of I image, perhaps the I image will deposit video memory 120 in the visual data of P.This is because it is B image normal decoder reference data that I image and P pictorial data all will be used at MPEG1 and MPEG2.

When the end-of-encode of I image, with last width of cloth P image of GOP front the B image is carried out two-way prediction and encode.Then, the data of the output of preprocessor 101 and reference frame storing device (arrow among the figure is omitted) compare, and with the retrieval moving vector, and judge prediction pattern and frame structure.According to judged result, from video memory 120, read with the optimal reference frame storing device of the output signal of preprocessor 101 in data, these data are divided into forward part and reverse part.Data are subtracted each other (time that this result refers to P image and B image remains difference component) with the output result of preprocessor 101 in subtracter 106.This time residue difference component carries out DCT processing, quantification and variable length code to be handled.

Be divided in data under the situation of low frequency range and high frequency region, speed frequency component aspect becomes uncertain.Therefore, since uncertain at the data rate of low frequency range, so the controlled range of bare headed driver can not be compensated fully.Here, it is variable that rate controller 140 makes the low frequency component district.Speed control 140 control speed are so the size of low frequency range is variable about targeted rate, shown in Figure 73 B.

In other words, when keeping watch on 126 outputs of first variable length coder, rate controller 140 has reduced the size of the area that the low frequency region data occupy, and at this moment, the output of being monitored is greater than the targeted rate of this application settings.When the sign indicating number amount of first variable length coder 126 hour; Rate controller 140 enlarges the area of low frequency region; In this mode; When keeping watch on the sign indicating number amount, rate controller 140 suitably changes the low frequency region area occupied of setting in first variable length coder 126 and second variable length coder 127.

In addition, for example, can carry out instantaneous coding, which zone to have big yardage according to, which zone has less yardage, sets the shared area of low frequency region to confirm a standard, thus target setting speed.

Figure 74 is the block diagram of decoded digital video signal processing unit.Illustrate as above coded data is carried out the process of decoding process.In Figure 74, reference number 141 is EOB retrieval units.Wherein, with Figure 72 category-A like or identical part with identical reference number.The mode signal that expression is being jumped retrieval or carried out the normal state of resetting continuously is input to mode converter 130 from microcomputer etc.Simultaneously, amplified by an amplifier from the replay signal of dish, replay signal is carried out digital demodulation with the clock difference of PLL.Carry out error correction, extraction will be input to the video bit stream in the data recombination device 131, thereby audio signal is separated from a system layer.Output as the mode converter 130 of mode conversion device is added to data rearrangement device 131 and decodable code determinant 132.Data rearrangement device 131 obtains this control signal, is connected to data before separating of L component and H component.In addition, have only the L component to output to variable-length decoder 133 (decoding device), and be free of attachment to the H component.

In theory, the L component will never be cut apart in the centre of incident.Can consider a kind of situation, when the signal of the retrieval of jumping etc. is good inadequately, confirm the border of incident with variable-length decoder 133 and variable determinant 132, like this, the former part in border is all decoded and output to switch 134.The mode of operation of switch 134 is, when the signal quality of replay data is the same when good during with normal playback, it is in " on " all the time.Here, decodable code determinant 132 has constituted the data manipulation device with switch 134.

Switch 134 is by 132 controls of decodable code determinant, and like this, O is inserted into high frequency one side of piece, leaves the low frequency component that success is decoded, thereby constitutes the DCT piece.Then, these data are carried out reverse DCT, and under the situation of I image, the output of adder 138 is passed through.On the contrary, under the situation of P image, data are corrected, with the moving vector part addition with reference to the I image.About the B image, be controlled for reading of video memory 137, and in adder 138 additions, like this, the B image is wanted the I image of addition and the moving vector of P image partly to correct.Sign indicating number that solves through to shaven head is controlled DCT pattern and prediction pattern moving vector.The data of motion-compensated prediction are decoded by this way and are deposited video memory 137 in.Then, image reassembles into the order of original formation.Reverse scan transducer 139 buffered datas, and an output data transaction in proper order of pressing piece is to raster scan.

In above-mentioned explanation, be that example is illustrated with the size of controlling the DCT coefficient.Also can use the control of event number to replace.In this case, occur the L component sometimes and do not reach and give fixed number purpose incident, and EOB situation about being coupled with.Yet, because EOB retrieval unit 141 is kept watch on the appearance of EOB, so the L component can be retrieved with certainty.Specifically, come composition data according to the data of low frequency region, the data and the EOB of high-frequency region respectively.In other words, data recombination device 131 has constituted a data reconstruction device with EOB retrieval unit 141.

Because the energy behind the DCT is very little, obviously, the L component is encoded with the best mode identical with non-encoding block of H component.About the H component, the data except that the L component are the working length coding ideally.Through being set at 0 to the L component, the H component can be encoded.Because might use the same structure of accomplishing of variable-length decoder with normal MPEG, so can on circuit, simplify.

Embodiment 17

Explain embodiments of the invention 17 with reference to Figure 75.Figure 75 shows the block diagram of decoded digital video signal processing unit.In Figure 75, reference number 142 is multiplexers, the 143rd, and switch; 144 is first variable-length decoders, and 145 is second variable-length decoders, and 146 is first inverse quantizer; 147 is second inverse quantizer; 148 and 149 is adders, and 150 and 151 is video memories, the 152nd, and the resolution reverse converter.Figure 75 also shows a low resolution decoding unit as decoding device.Indicate with identical reference number with part similar or identical among Figure 72 A, its explanation is omitted.

Below, the course of work of explanation embodiment 17.Content shown in Figure 75 can think at encoded data record that Figure 68 describes under the situation on the CD etc., the process that the replay signal block of video data of dish is handled.

For example the jumping mode signal of retrieval or the normal state of resetting continuously of indication is input to mode converter 130,130th, mode conversion device from microcomputer etc.Simultaneously, amplified by an amplifier from the replay signal of dish, replay signal is carried out difference with the clock of PLL, is used for digital demodulation.Then, isolate audio signal, to extract video bit stream from system layer.

The video bit stream of this extraction is input to multiplexer 142.Multiplexed road 142 is delivered to second variable-length decoder 145 to the low resolution component data, delivers to first variable-length decoder 144 to other data through switch 143 simultaneously.

Switch 143 is by mode converter 130 controls.As a kind of pattern, though when the retrieval of jumping waits, only need the playback image output of low resolution, to reset midway at resolution residue difference component, switch 143 will move and send redundant data to suspend.

The decoding of 145 pairs of Huffman codes of second variable-length decoder and working length sign indicating number is second inverse quantizer, 147 inverse quantization then, is transformed into spatial component by reverse DCT circuit 136 from frequency component thereafter.

For the I image, data converted deposits video memory in through adder 149.Under the situation of P image, the I image from video memory reads first frame P image, and second frame P image or later image are meant the follow-up image that exists in the video memory and partly corrected by moving vector, to carry out the motion compensation prediction in adder 149.Under the situation of B image, carry out identical operations according to I image and P image.

In Figure 75, the quantization parameter and the prediction pattern of variable-length decoder output moving vector, inverse quantization.These moving vectors, quantization parameter and prediction pattern are all identical with Figure 74.The ring shown in the chain-dotted line frame is a formation unit to the decoding of low resolution component among Figure 75.Because decoded result is inserted between the pixel in the resolution reverse converter 152 (interpolation video generating means), the decoded result as resolution residue difference component is compensated decoded result input image memory 150.

The resolution residual components decoded result when normal playback and the result combinations of low frequency component (or carrying out when time-division operation under the situation of decoding of resolution residual components according to the time-division result) are exported by the reverse scan transducer as image.Data can be decoded as frequency component by first variable-length decoder through switch 143.146 pairs of data inverse quantization of first inverse quantizer, reverse DCT circuit 136 is data decode the resolution residual components data of area of space.

Video memory 150 is relevant with the pixel interpolative data of low resolution component; The P image is relevant with the I image, and the B image is relevant with the P image with the I image, therefore; Data are in position corrected by the moving vector part; As a result, data are read from video memory 150, the decoding that adder 148 is accomplished the motion compensation prediction.

In the situation of retrieval of jumping, to reset in order to prevent resolution residue difference component midway, switch 143 forbids that redundant data from reverse DCT circuit 136 outputs, promptly interrupt the output of resolution residue difference.Therefore, have only the low resolution component data of in pixel, inserting to export (operating process is identical, as the importation at video memory 150 switch is set) through video memory 150 and reverse scan transducer 139.

Embodiment 18

Explain embodiment 18 with reference to Figure 76 and 77.The block diagram of Figure 76 is the GOP address-generation unit and the block diagram of dish control unit, and this figure is illustrated in especially and is recorded in the processing block diagram under the situation on the header in succession to aforementioned rate information.In Figure 76, the 153rd, register, the 154th, the GOP address calculator, the 155th, shaven head/dish rotates the control transformation device and rotates the control transformation device as head position conversion equipment and dish.Figure 77 is a block diagram, shows GOP address-generation unit and dish control unit that comprising resets handles, and this figure shows a kind of structure especially, and the GOP that is used for the completion dish resets, and on several position on this dish, collecting has aforesaid rate information.Consult Figure 77, the 156th, playback amplifier, the 157th, digital demodulator, the 158th, error-corrector, the 159th, system layer processor, the 160th, speed data memory.System layer processor 159 and speed data memory 160 composition data rate information extraction elements.The 161st, the GOP counter.GOP address calculator 154 has constituted the positional information calculation device with GOP counter 161.

Below, explain the course of work of embodiment 18, through letting each GOP rate-compatible, as convention, can optimize the whole speed of a program, result, the quality of image can think and improve.Yet, be not that before observing data content, data all drop on the front of GOP.Also have, halfway once more under the situation of that playback, only way is to detect original position through the data on the retrieval dish constantly at software that hope has been mated.

In this case, in the incipient stage, the rate controlled that variable bit rate is set is discrete speed target, and like the 1M bit, 1.5M bit, 2M bit, 2.5M bit, 3M are than top grade, thereby each rate information in all GOP all is recorded on the dish.Specifically, should be the most effective when the rate information of each GOP is recorded on TOC (contents table: recording areas is assigned to beginning most of dish, like this, quilts such as title, writing time record), half TOC etc.

Rate information about GOP can combine with the sequence header of video bit stream.For example, two hours software 14.4K GOP (twohoursoftware14.4kpiecesofGOP).When rate information can be divided into 5 kinds of speed, the rate information of this moment can be used 3 bit representations.Therefore, all GOP speed can be recorded in dish upward (14.4K * 8/byte of 3 ÷) with 5.4K.

Through depositing the rate information of each GOP in the speed data memory 160 of Figure 77 in, and add and can realize high speed access corresponding to the message length of this value.

Explain device of the present invention with reference to Figure 76.To Huffman code, the decoding of working length sign indicating number, the decoding header like this, is judged moving vector and visual kind.

Simultaneously, sequence header is decoded, and like this, rate information is imported into GOP address calculator 154.In addition, the address information of the GOP that was just being visited at that time deposits register 153 in, like this, calculates the next preceding address of GOP that will visit, and deposits register 153 in.Simultaneously, rotate the front of control transformation device 155, confirm the position of shaven head according to the address to the next GOP that will be visited through making shaven head/dish.Then, calculate the control signal of next visit from the GOP that is visiting with the difference of the preceding address that will visit.According to this control signal, realize the Position Control of bare headed driver and the control of dish rotation.

With reference to Figure 77 playback procedure is described.Address (behind assigned rate information explanation address, visiting this address part with reading speed information) is read in the rotation of control shaven head and shaven head directly or indirectly from the zone of toc region territory or corresponding TOC.Then, with the replay signal that playback amplifier 156 amplifies from shaven head,, be the digital signal that digital demodulation is used with this signal differential with digital demodulator 157 these signal waves of inspection.

Digital demodulation is that the replay signal of digital signal is input to mistake in the error-corrector 158 correction replay signal.Data after the error correction are isolated audio bitstream in system layer processor 159, video bit stream and other data clauses and subclauses.

For example, judge this signal belongs to which kind of data (binary data of AV (video and audio frequency) data, file data and program etc.), and circulation road is divided in cutting in view of the above.In such processing, aforesaid rate information writing rate data storage 160.

On the contrary, the information that produces about the GOP number that will handle with GOP counter 161.According to the address that GOP address calculator 154 calculates, control bare headed driver and dish velocity of rotation.

In above-mentioned explanation, what provide is that the GOP counter receives the signal from system layer meter reason device 159 for example.Replacing processing or operation under the situation of the jump scanning of resetting with I/F (like microcomputer), the processing of carrying out video bit stream from input data such as variable-length decoders is more effective.

Embodiment 19:

By Figure 78,79 and 80 embodiments of the invention 19 are described.Its work will be explained as follows.Figure 78 illustrates a signal processing unit, wherein carries out cutting apart of carrying out according to the frequency of digital signal playback unit and cutting apart of carrying out according to quantizing, and this figure is a block diagram, and in the case, data rate information is collected and is recorded in the several places on the CD.With part identical or corresponding in the accompanying drawing in the previous embodiment by identical symbolic representation.

Shaven head or shaven head rotate and to be controlled, and therefore, data can be directly or from toc region or the district corresponding with toc region, read indirectly.Replay signal is amplified by playback amplifier 156.This signal is by 157 detections of digital demodulator subsequently.As a result, the replay signal that has become numerical data is imported into the mistake that 158 pairs of error-corrector are included in the replay signal and carries out error correction.Error-free data are by system layer processor 159 composition audio bitstream and video bit streams, and other data are also so handled.

For example, judge that this signal is that which kind of data in AV data, lteral data or the binary data such as program are with flow path separately.Aforementioned data rate information is deposited in the data transfer rate data storage 160.The information corresponding with the GOP number that need be processed on the contrary, produces through using GOP counter 161.Subsequently, by 154 calculated addresses of GOP address calculator.On the basis, address of being calculated by GOP address calculator 154, the rotating speed of bare headed exciter and CD is controlled.

In this way, jumping over playback time, on CD, jumping over to find with the address of being visited.When jumping over when visiting needed GOP, use the low frequency range data that obtained like embodiment 16 said structures to be reset and on screen, describe when calculating next address in the same manner.

A representative data is jumped over mode signal search or that carry out the state of normal playback and is input to the mode switch 130 from microcomputer.As stated, the video bit stream that is imported in the data rearrangement device 131 is extracted out.The output of mode switch 130 is added on data rearrangement device 131 and the decodable code determinant 132.Data rearrangement device 131 obtains such control signal and works, and is therefore linked up again by the data before cutting apart in L component in Figure 71 and the H amount.Otherwise data rearrangement device 131 only outputs to the L component on the variable-length decoder 133, and the L component is not linked on the H component.

In embodiment 19, theoretically, the L component can not occur and cut off midway.But, consider that to have undesirable semaphore that jumps over search etc. decoded, the border of incident confirmed by variable-length decoder 133 and decodable code determinant 132, so that can be decoded and outputed on the switch 134 until the part on border.

Switch 134 is to be controlled by the output of decodable code determinant 132, therefore, from the low frequency component that is constituted the DCT piece by complete decoding, will zero be input to the front end of piece.Subsequently, make reading of video memory 137 is controlled, and addition is provided by adder 138; Like this; Data will be carried out reverse dct transform according to following mode, the time output of adder 138 passed through at the I image, and data are proofreaied and correct by dynamic vector part that will addition when the P image; When B resembled, data were partly proofreaied and correct from the dynamic vector of I image and P image and by addition.

In addition, carve at this moment by the decoding of header sign indicating number being controlled DCT pattern and predictive mode dynamic vector.In this way, it is decoded and be stored in the video memory 137 to constitute according to former formation order and resemble to move the data of compensation prediction.In reverse scan transducer 139, data are buffered, to convert the block scan data to the raster scan data by the output order of elephant.In addition, switch 134 is UNICOM not, makes zero being inserted into when normal playback, but is controlled the back work and the replay data of only resetting.

After data are cut apart and are encoded into low frequency range and high frequency region, following situation will appear: focus on the quantization table of low frequency end, the quantization table that focuses on front end and no matter frequency field and quite average thin quantification with respect to a quantization table., two groups of variable-length decoders and inverse quantizer can realize this situation when being used as this machine decoder shown in Figure 68.At this moment, data rearrangement device 131 will be a multiplexer.

Below, will be on the basis of Figure 79 the work of illustrative embodiment 19.Figure 79 illustrates signal processing unit; The bit length of wherein cutting apart according to the digital signal playback unit carries out; This diagrammatic sketch is a block diagram, is used for handling explanation embodiment with respect to the playback that aforementioned data rate information is collected and records under the situation in several specific places of coiling.For example, beginning most of optical disk data reproducing, shaven head or shaven head rotate controlled, thereby data can be directly or indirectly read from toc region or the district corresponding with toc region.Replay signal from shaven head is amplified by playback amplifier 156, and therefore, available digital decoder 157 detects this digital signal that difference is become to be used for digital demodulation.

As a result, the replay signal that has become numerical data is imported in the error-corrector 158 in order to the replay signal error correction.Error-free data are divided into audio bitstream and video bit stream, and other data are also so handled.

For example, whether these signal determining data are the binary data of video/audio data, lteral data or program and so on, to separate flow path.From these data, aforementioned data rate information is deposited in the data transfer rate data storage 160.Simultaneously, the GOP number of handling as need by GOP counter 161 generation information, and remove to control the rotating speed of exciter and CD by 154 calculated addresses of GOP address calculator.

In this way, search through jumping on CD jumping over the GOP address that playback time will access.When required GOP is accessed to, then next address will be calculated in an identical manner, simultaneously, with as the low frequency range data that obtained of embodiment 15 described structures reset out, with these data of reproduction on a screen.

Representative data is jumped over mode signal search or that carry out the normal state of resetting continuously and is input on the mode switch 130 from microcomputer etc.Video bit stream is extracted out and is imported in the data rearrangement device 131.The output of mode switch 130 is added on data rearrangement device 131 and the decodable code determinant 132.Data rearrangement device 131 obtains such control signal and works, and is linked up again by the data before cutting apart in L component in Figure 71 and the H component like this.Otherwise data rearrangement device 131 only outputs to the L component on the variable-length decoder 133 and the L component is not linked on the H component.

Be extracted in the border of the incident among the L component by variable-length decoder 133 and decodable code determinant 132, like this, can be decoded and output on the switch 134 until the part on border.Switch 134 is controlled by the output of decodable code determinant 132, thereby is inserted into the front end of piece from fully decoded null value with the low frequency component that constitutes the DCT piece.These data are carried out reverse dct transform.When being the I image, the output of adder 138 is passed through.When being the P image, this resembles the dynamic vector part that will be coupled with in the benchmark I image and proofreaies and correct.The reading of video memory 137 is controlled and by adder 138 additions, like this, data can be proofreaied and correct by the dynamic vector part.

In addition, carve at this moment by decoding control DCT pattern and predictive mode dynamic vector the header sign indicating number.In this way, it is decoded and be stored in the video memory 137 to constitute with the order according to former formation GOP and resemble to move the data of compensation prediction.In reverse scan transducer 139, data are buffered, so that the block scan data transaction is become the raster scan data.In addition, switch 134 not conductings make zero being inserted into when normal playback, after the conducting, and the replay data of only resetting.

Below, explain the work of Figure 80.Figure 80 illustrates the signal processing square frame; Wherein data are cut apart according to the resolution of digital signal replayed portion; This sketch map is a block diagram, is used for handling explanation embodiment with respect to the playback that aforementioned data rate information is collected and records under the situation in the several specific places on the CD.Begin most shaven head or shaven head at optical disk data reproducing rotate controlledly, and therefore, data can be directly or indirectly read from toc region or the district corresponding with toc region, are amplified by playback amplifier 156 from the replay signal of shaven head.Like this, available digital decoder 157 detects this digital signal that difference is become to be used for digital demodulation.

Representative data is jumped over mode signal search or that carry out the normal state of resetting continuously and is input on the mode switch 130 from microcomputer etc.Video bit stream is extracted out and is imported on the multiplexer 142.Multiplexer 142 is added to the low resolution component data on second variable-length decoder 145, simultaneously other data is added on first variable-length decoder 144 through switch 143.Switch 143 is controlled by mode switch 130.Need be although have only the playback of low resolution component to resemble output as a pattern in jumping over search etc., switch 143 breaks off in all the other components of resolution are reset the way.In addition, when playback work is carried out, 143 conductings of switch, the signal transmission quality that wherein can obtain is to carry out normal playback.

145 pairs of Huffman sign indicating numbers of second variable-length decoder and the decoding of trip distance sign indicating number.The data of being carried out inverse quantization by second inverse quantizer 147 will transfer it to spatial domain from frequency domain by reverse DCT circuit 136.When data are I when image, the data through adder 149 will be stored in the video memory.When data are the P image, the P image is proposed partly to be proofreaied and correct on the tram by dynamic vector subsequently from video memory, to set out to compensate prediction by adder 149 decodings.When data are that B is when resembling, with respect to I image and the same operation of P image execution.

In Figure 80, from variable-length decoder, export the quantization parameter and the prediction pattern of dynamic vector, inverse quantization.Because information flow is identical with Figure 74's, thereby no longer explain.At a loop of Figure 75 bottom loop for the decoding of low resolution component.Decoded result is undertaken inserting in the pixel by resolution reverse converter 152, and with the decoded result of compensation as resolution residue difference, these data are imported in the video memory 150.

Embodiment 20:

Explain embodiments of the invention 20 with reference to Figure 81,82 and 83.Figure 81 and 82 is the block diagram of decode procedure.In Figure 81 and 82, label 162 representatives are as the video coder of code device.163 is that audio signal encoder, 164 and 167 is that memory, 165 and 168 is Memory Controller.Memory 164 and Memory Controller 168 composition data generators.In addition, video coder 162 constitutes one yard amount comparison means with storage controller 165.Label 166 representative system layer bit stream generator.169 represent variable length coder, and 170 represent the signal processor of the decoding after the variable

length coder.Encoder

169 and 170 is as data deciphering device.The data rearrangement device 131 of Figure 82 is as the data reconstruction device.

At first, explain the structure shown in Figure 81.Between video coder 162 and systematic bits flow-generator 166, settle memory 164.Be embedded between each GOP of vision signal of coding in data after; Each GOP is imported into system layer bit stream generator 166, and the audio signal of encoding with audio signal encoder 163 is input in the systematic bits flow-generator 166 with the vision signal that adds operations such as header subsequently.

At this, the data of describing in the memory 164 are imbedded operation.Memory Controller 165 is used for the vision signal of control of video signal data coding as the control circuit of memory 164 with control, thereby between each GOP, is embedded into the space.Explain signal processing with reference to Figure 83 below.Figure 83 is a sketch map, is illustrated in the processing notion on digital video signal recording and the replay device.For example; When nGOP produces redundant data with respect to the control unit of the access address of shaven head and wrong control (n+1) GOP with respect to the control unit of the access site of shaven head or error correction midway under situation with the space that produces the data field; Device of the present invention is shown in Figure 83 A; Make nGOP redundancy as certificate imbed (n+1) GOP space segment afterwards; And in the same manner, make because of nGOP and imbed in the space segment that a small amount of remaining data of (n+1) GOP that can not imbed and (n+2) GOP be embedded to (n+3) GOP (imbed direction and see it is from left to right) from paper.

In addition, as another kind of control method, redundant data can not oppositely transmit as stated.Shown in Figure 83 B; When (n+2) GOP is some above access site; (n+3) GOP finishes with respect to the access site of the shaven head and the control unit of wrong control midway, and device of the present invention is controlled, like this; Redundant (n+3) GOP data are imbedded space segment after (n+2) GOP data; In this way, the remainder data that is embedded in (n+2) GOP that can not imbed because of (n+3) GOP is imbedded in the space segment of (n+1) GOP, and (n+1) GOP that can not imbed partly imbed the space segment (imbed direction and see it is from right to left from paper) of nGOP.

The work of structure shown in explained later Figure 82.Memory 167 is controlled by Memory Controller 168, and the data of resetting according to earlier figures 83A and the described rule of 83B revert to previous status.For example after data were resumed shown in Figure 83 A, device work of the present invention made the GOP data revert to original state, followed the nGOP that (n+1) GOP is arranged partly to be linked to the left part of (n+1) GOP, and subsequently, (n+1) the GOP data are linked on (n+2) GOP data.

This reordering rule will limit with the format rule of medium; Like this; This rule is being recorded such as the flag information in the good zone of the rearrangement in toc region territory down as following, and under the not limited situation of form, rule must clearly be described out in the somewhere on medium.

Embodiment 21

Explain embodiment 21 with reference to Figure 84,85 and 86 below.Figure 84 is a block diagram, and the representation signal processing unit is wherein carried out cutting apart or cutting apart according to what quantize according to frequency at the digital signal replayed portion.Figure 85 is the block diagram of representation signal processing unit, wherein carries out cutting apart or cutting apart according to what quantize to carry out according to position progress row at the digital signal replayed portion.Figure 86 is the block diagram of representation signal processing unit, wherein carries out according to what resolution was carried out at the digital signal replayed portion and cuts apart or cut apart according to what quantize to carry out.In Figure 84,85 and 86, label 171 is represented the IP selector marker, and 172 represent determinant, and 173 represent switch.In Figure 86, counterpart shows makes first decoding device, second decoding device and the 3rd decoding device.In Figure 84,85 and 86, similar part is represented with same numeral, and is omitted the description to it carried out.

The work of explained later embodiment 21.In Figure 84 and 85, bare headed or bare headed rotation is controlled, and like this, data are read from toc region or the zone corresponding with toc region directly or indirectly.Replay signal from shaven head is amplified by playback amplifier 156.This signal is become to be used for the digital signal of digital demodulation with difference by digital demodulator 157 retrievals.As a result, become the replay signal of numerical data to be imported into error-corrector 158 with to wrong error correction contained in the replay signal.Error-free data are divided into audio bitstream and video bit stream by system layer processor 159, and other association is also so handled.In Figure 84,85 and 86, control signal is input in the IP selector marker 171 from the mode switch 130 as mode-changeover device, with switching encoding/decoding device on the basis of special-purpose playback speed.Embodiment 21 is controlled, and like this, embodiment 21 can and jump over sweep speed with this control signal and between the pattern of the pattern that only shows the I image or demonstration I image and P image, switch.

When jumping over search speed and be 100 times, if I image and P image are exported on screen, GOP is bound to export with thinization of a great deal of.As a result, with respect to moving of playback screen, resemble as if quite unnatural.In the case, for eliminating this not nature, need mode switch on the pattern of the I image of only resetting.Decodable determinant 132 (among Figure 86 172) not only is used for B is resembled decoding but also to P image decoding (switch 173 at Figure 86 as this function).Simultaneously, video memory 137 (among Figure 86 150 and 151) is controlled only to show the I image.

Speed when the screen display of I image and P image prefers 15 times of speed usually, but only the screen display of I image prefers 15 times and above speed.This is because resemble when under 15 times of speed, showing as entire I image and whole P, owing to the GOP that in subsequent treatment, can reset be positioned at from addition screen to every frame update and the 5th GOP position of the GOP of current demonstration, continuous motion extreme variation.In addition, when the image of frame number N=15 and I in GOP was 3 with the P image in the cycle, all P images were decoded and only have visual second frame (the 3rd and the 9th frame in the GOP) quilt of I image and P to be exported, and can carry out the thinner search of jumping over.

As stated; On the basis of the predetermined condition of following situation through data mode being cut apart and the record data state; These situation are: through the dividing frequency district data are recorded from each GOP precalculated position that recording medium is read, be used for record data situation, by the resolution partition data make data by the situation under recording and through the quantization level that will be write down the situation of partition data.When playback is come out from concentrate the data of settling for first data of master data in the replay data and second data except that basic first data; Coming from following situation, to obtain playback with a decoding device one of resembles: the situation that all first and second data are decoded, and the data that obtained are and I image and the low frequency range of P image or the corresponding situation of number of picture elements of thinization.Subsequently, can switch on the basis at special-purpose playback speed at the used decoding device of special-purpose playback time.

Need not speak more, show that the mode of I image and P image can change forward and negative sense at playback time.Cause can be decoded to only at forward the time in P, existing screen before the decoding when therefore being stored in the reverse reproduction of P image.As a result, need on that part of, use redundant memory.Need not this redundant memory for the ease of reverse reproduction, the I image of when forward jumps over search, can resetting resembles with P, is oppositely jumping over the I image of only resetting when searching for.

Embodiment 22:

Explain embodiment 22 with reference to Figure 87,88,89 and 90.Figure 87 is a block diagram, and the representation signal processing unit is wherein carried out cutting apart or cutting apart according to what quantize according to frequency at the digital signal replayed portion.Figure 88 is the block diagram of representation signal processing unit, wherein carries out cutting apart or cutting apart according to what quantize to carry out according to position progress row at the digital signal replayed portion.Figure 89 is for explaining the diagrammatic sketch of when jumping over search, handling notion.In Figure 87 and 88, label 174 is represented a display controller.System layer processor 159 is as the video data extraction element.In addition, Figure 87 and 88 illustrates the video data decoding of counterpart and the instance of replay device.Label 130 representatives are as the mode switch of mode-changeover device.Same or analogous part use with previous embodiment in the representative of same or similar label.

The work of explained later embodiment 22.In Figure 87 and 88, bare headed or bare headed rotation is controlled, and like this, data are read from toc region or the zone corresponding with toc region directly or indirectly.Replay signal from shaven head is amplified by playback amplifier 156.This signal is become to be used for the digital signal of digital demodulation with difference by digital demodulator 157 retrievals.As a result, become the replay signal of numerical data to be imported into error-corrector 158 with to wrong error correction contained in the replay signal.Error-free data are divided into audio bitstream and video bit stream by system layer processor 159, and other association is also so handled.

When jumping over search, outputed to continuously on the screen such as I image and P image, screen is then exported according to the order of arrow representative among Figure 89.At this moment, when jumping over search, the odd field of the even field of I image and P image gets up continuously, and four barnyards are arranged between them in coded data.In other words, the playback speed in coded data is between the parity field of I image in the space five times of playback speed.Therefore, because playback speed changes to 5 times of speed from 1 times of speed concerning one, then the motion change of I image is unnatural.

This can be substituted by the odd field or the even number place of I image on same screen.Otherwise screen can prepare through the mean value that lower tracer is imbedded in the boundary of considering output.Video memory 137 shown in Figure 87 and 88 is with 174 controls of field display controller, in follow-up P image, to constitute same screen.As a result, on every of playback elephant when record for space between the field of coding between the amount of jumping over can accompany or follow and jump over (not proper motion) and change to unconspicuous result and can evenly obtain.

In addition, Figure 90 is a sketch map of handling notion when explaining reverse reproduction, the field sequential when wherein showing reverse reproduction especially.Below, the field sequential in the time of will on the basis of Figure 90, explaining reverse reproduction.When reverse reproduction, reverse reproduction is carried out in the frame unit that constitutes a pair of parity field.Specifically, when operation when odd field fades to even field, reset as on the image time execution (replay operations process a describes in Figure 90) on same direction.When operation when even field fades to odd field, two field parts are reversed and see off (Figure 90 A illustrate and jump in process b) with the direction in the opposite direction with on image the time.

But, when aforementioned playback procedure is carried out, the result that the triple speed degree is reset, making resets to resemble with unaccommodated mode moves, and like this, it is steady that motion sequence can not felt.Receive field display controller 174 controls of Figure 87 and 88 when video memory 137 after, resetting to resemble on a screen shows with the opposite direction of order strange, even, strange, even shown in Figure 90 B.Because the amount of jumping between can obtain almost evenly, beating then becomes not obvious.But,, need to keep the relation of normal odd number and even field and will field sync signal not reverse with respect to the synchronizing signal of this moment.

Video memory 137 does not receive the output of adder 138, because it has adopted the display packing that is different from each, but the output of adder 138 can receive from video memory 137 individually.For carrying out this operation,, the buffer as separating device changes order through being provided.Can use the memory with 3 ports, address control can be set up individually in 3 ports.Aforementioned operation can in addition when data with can when the memory of sense data under the fast speed very is multiplexing, still can realize.In addition, because reverse scan transducer 139 adds a jointing positions at least one at least one memory is provided, needn't speaks more, the sort buffer function can be introduced in the reverse scan transducer 139.

In addition, about the output of playback slowly that special use is reset, when same frame is repeated to export, jumps over and become not obvious.As a result, frame is by reconstruct and output, and this playback is identical at interval.For example, slow playback time under 1/3 times of speed B frame that decoded I frame decodes after with three times of outputs can not occur with three times of outputs yet.The substitute is, first frame is that the odd field by the I frame constitutes.In even field, can adopt the mean value that rolls off the production line.

In this formation, shift out at present because the row of intersection elephant moves not have on the vertical direction that makes screen to resemble, thereby can obtain stable resembling.The I frame that next frame output is original, next frame constitutes a frame (at the even field end, the mean value of desirable up-downgoing) with the odd field of I frame again.Subsequently, the original B frame of back one frame output, next frame constitutes a frame (at the odd field end, the mean value of desirable up-downgoing) with the even field of B frame again.As a result, in the period interval, can uniformly-spaced realize resetting slowly at this moment.

Claims

1. digital video signal replay device; Be used for replaying recording medium; Record the digital video signal that comprises block of video data on this recording medium; This block of video data comprises through the I image that utilizes motion compensated prediction and orthogonal transform coding, P is visual and B is visual, and in said recording medium:

Said digital video signal is divided into access unit and wrong control unit;

The header of the said block of video data of said digital video signal is corresponding to the header of each access unit or wrong control unit;

When white space when the end of said access unit or wrong control unit produces, predetermined data are embedded in this white space;

Said replay device comprises:

Be used for said embedding data are separated from the data of the block of video data of each access unit or wrong control unit, and the equipment of the said data video signal of reconstruct; With

Be used for equipment to the decoded digital video signal of institute's reconstruct.