CN1926862A

CN1926862A - Video information transmission

Info

Publication number: CN1926862A
Application number: CN200480041983.8A
Authority: CN
Inventors: 王如生; 拉吉普·库尔塞伦; 维克托·瓦萨; 基思·米勒
Original assignee: Nokia Oyj
Current assignee: Nokia Oyj
Priority date: 2004-02-23
Filing date: 2004-02-23
Publication date: 2007-03-07
Also published as: EP1719343A4; WO2005091632A1; EP1719343A1

Abstract

The present invention relates to a method for transmitting video information (fig. 5), in which a bitstream (510,520) is formed comprising a set of frames (512,513,515,516,518 and 522,523,525 and 528) comprising macroblocks. At least one switching frame (524) is formed into the bitstream and macroblocks of the switching frame are arranged into a first and a second group of macroblocks, each macroblock of the first group are encoded by a first encoding method (fig. 5, intra) to provide a switching point for continuing the transmission of video information with another bitstream formed from the video information.

Description

The video information transmission

Technical field

The present invention relates to a kind of method that is used for transmission of video information, wherein from the video information that comprises a framing, form at least one bit stream.The invention still further relates to a kind of encoder, decoder, transmission system, signal and computer product.

Background technology

Typical video flowing comprises the sequence of pictures that is commonly referred to frame.Frame comprises the pixel that is arranged in rectangular in form.In existing video encoding standard, for example, H.261, H.262, H.263, H.264 with MPEG-4 in, defined three kinds of main picture/mb-type: frame (I frame), predictive frame (P frame) and bidirectional frame (B frame) in the frame.Every kind of picture/mb-type is used dissimilar redundancies in image sequence, and therefore obtains different compression efficiency grades, and as described below, and different functions is provided in encoded video sequence.Frame is not use the video data frame of encoding from any information of past or future frame by the spatial coherence that only uses inner each pixel of frame itself in the frame.Frame is used as the basis of the decoding/decompression of other frame in the frame, and the access point to coded sequence is provided, and begins there to decode.

Predictive frame is to use from the called reference frame, just the frame that the motion compensated prediction of frame or predictive frame is encoded/compressed in available one or more front/rear frames in encoder or decoder.Bidirectional frame is the frame of encoding/compressing by from the prediction of frame or predictive frame in frame or predictive frame in the previous frame and/or the back frame.

Because consecutive frame is a height correlation in the typical video sequence, using two-way or predictive frame but not can obtain higher compression ratio in the frame in the frame.On the other hand, when service time predictive coding in encoded video streams, not correctly to be used to encode two-way and predictive frame all other before or after under the reference frame situation of decoding, can not decode to B frame and/or P frame.If owing to the mistake in the transmission or because some intentional actions of transmission ends, in the encoder in used reference frame and the decoder corresponding reference frame inconsistent, then can not rebuild from the subsequent frame of the prediction of this reference frame use, with the identical decoded frame of that frame that obtains and encode at coding side at first in decoding end.This not matching is not limited only to single frame, also can further propagate in time owing to the use of motion compensation encoding.

The type of Figure 1A-coding/compressed video frame that the 1C explanation is used in typical encoding and decoding of video system.For example, before coding, represent the picture of video sequence with these multiple bit digital matrixes, the brightness of a representative image pixel, each in two other is represented one of them of two colourities (color) composition respectively.Figure 1A has described the mode of only using the interior image information that exists of frame itself that frame in the frame 200 is encoded.Figure 1B has illustrated the structure of predictive frame 210.Arrow 205a representative uses motion compensated prediction to generate P frame 210.Fig. 1 C has described the structure of bidirectional frame 220.The B frame is inserted between I frame or the P frame usually.Fig. 2 has represented a picture group sheet with DISPLAY ORDER, and illustrates how the B frame is inserted between I and the P frame, and has represented the direction that motion compensation information flows.In Figure 1B, 1C and 2, arrow 205a has described and has rebuild the required forward motion compensation information of forecasting of P frame 210, and

arrow

215a and 215b have described at forward direction (215a) and back and go up the motion compensation information of rebuilding use in the B frame 220 to (215b).In other words,

arrow

205a and 215a represent when from DISPLAY ORDER early than the information flow when each frame of reconstruction frames is predicted predictive frame just, arrow 215b represents when be later than the just information flow when each frame of reconstruction frames is predicted predictive frame from DISPLAY ORDER.

In motion compensated prediction, the similitude in the video sequence between the successive frame is used to improve code efficiency.More specifically, so-called motion vector is used for describing pixel or the move mode of pixel region between the successive frame of sequence.Motion vector provides deviant and error data, and its expression has the past or the future frame of the video data of decoded pixel value, and wherein the decoded pixel value can be used from error data one given video data frame is carried out compressed/encoded or decompressed/decoded.

The ability of decoding/decompression P frame needs the availability of previous I or P reference frame, and, for the B frame of decoding, also need the availability of follow-up I or P reference frame.For example, if coding/packed data stream has following frame sequence or DISPLAY ORDER:

I ₁B ₂B ₃P ₄B ₅P ₆B ₇P ₈B ₉B ₁₀P ₁₁...P _n-3B _n-2P _n-1I _n，

Then corresponding decoding order is:

I ₁P ₄B ₂B ₃P ₆B ₅P ₈B ₇P ₁₁B ₉B ₁₀...P _n-1B _n-2I _n。

Decoding order is different with DISPLAY ORDER, and this is because the B frame needs following I or P frame to decode.Fig. 2 has shown the beginning of above-mentioned frame sequence, and for the dependence of understanding each frame as discussed previously can be with reference to this figure.It is available that the P frame requires previous I or P reference frame.For example, P ₄Require I ₁Decode.Similarly, frame P ₆Require P ₄Be available, with decoding/decompressed frame P ₆B frame, for example frame B ₃Need the past and/or following I or the P reference frame, for example, P ₄And I ₁, to decode.The B frame is the frame between I or the P frame during the coding.

Video flowing occurs as the important application in the fixed the Internet.Reckon with that further video flowing also will be important in the 3G in future wireless network.In stream was used, transmission server began by the video bit stream of transmission network to the receiver transfer pre-coding when the request that receives from receiver.Receiver is displaying video stream when receiving.Since the network condition that changes, the best endeavors characteristic of current network make the user can with effective bandwidth change.Change in order to adapt to these, transmission server can be adjusted the bit rate of compressed video.Being under the conversational services situation of feature, can reach this purpose by dynamically regulating the source encoding parameter with real-time coding and point-to-point sending.This customized parameter can for example be quantization parameter or frame rate.It is suitable to the feedback from transmission network to regulate.Under typical stream situation, when previous coded video bit stream will transfer to receiver, can not use above-mentioned solution.

A kind of solution that obtains the bandwidth adjustability under the situation of precoding sequence is to produce stream a plurality of and that independently have different bit rates and quality.Transmission server then dynamically switches between stream, to adapt to the variation of available bandwidth.Following example illustrates this principle.The let us hypothesis is used different coding parameters corresponding to same video sequence, for example, and quantization parameter.Allow { P _{1, n-1}, P _{1, n}, P _{1, n+1}) and { P _{2, n-1}, P _{2, n}, P _{2, n+1}) represent decoded frame sequence respectively from bit stream 1 and 2.Because coding parameter is different for two bit streams, at the frame of synchronization from their reconstructions, for example, P _{1, n-1}And P _{2, n-1}, be inequality.If suppose that now server sends the coded frame from bit stream 1 at first, until time n, it begins to send the coded frame from bit stream 2 afterwards, decoder received frame { P _{1, n-2}, P _{1, n-1}, P _{2, n}, P _{2, n+1}, P _{2, n+2}).In this case, owing to do not receive its reference frame P _{2, n-1}, P _{2, n}Then can not carry out correct decoding.On the other hand, replace P _{2, n-1}The frame P that receives _{1, n-1}With P _{2, n-1}Different.

Therefore, owing to not matching between the reference frame that is used for motion compensated prediction in different sequences, the switching on the optional position between the bit stream causes visual artifacts.These visual artifacts are not limited only to the frame on the switching point between the bit stream, and owing to the continuous motion compensation coding in the video sequence remainder, also propagate in time.

The video quality that video flowing/delivery system is subjected to causing owing to transmission error inevitably worsens.Transmission error can be divided into random bit error and frame deletion mistake (packet loss) roughly.A lot of error control and concealing technology attempt to use that forward error is hidden, reprocessing and interactive error concealment are avoided this problem.Prediction video coding mechanism is little to the tolerance of packet loss, wherein loses the mistake that piece causes and will propagate, thereby generate disagreeable vision distortion.The intra-frame macro block of hiding based on forward error inserts (intra macroblock insertion) can be by introducing self-contained intra-frame macro block and the concealing errors piece stops error propagation.The problem of the intra-frame macro block of being introduced is that this macroblock encoding has increased the amount of information of bit stream, thereby, reduced code efficiency, and it can not adjust (scalable).

In the time can not carrying out the transmission again of lost packets, it is extremely important that good mistake is recovered (resilience) instrument.(AIR) system (" MotionAdaptive Intra Refresh for MPEG-4 " that Worral delivers) is upgraded in the self adaptation intraframe coding of describing in the MPEG-4 standard on the Electronic Letters in November, 2000, Worral mentions, and inserts the part of intra-frame macro block as the Motion Adaptive scheme on the position after a while of successive frame.Judge that when inserting macro block (when bandwidth is available for that frame) shows and can benefit from use high motion recognition image zone.Worral notices, its method and standard backward compatibility (not needing standards change).Encoder moves down frame coding intra-frame macro block, until the macro block number that sets in advance is encoded.For next frame, encoder begins at same position, and begins intra-frame macro block is encoded.

The purpose of inserting intra-frame macro block is to attempt the caused artefact of minimise false macro block and propagates, and stops this propagation.Another kind of optional method is to upgrade (RIR) at the intra-frame macro block at random that H.264 JM61e uses in the reference software, there, inserts intra-frame macro block randomly.But intra-frame macro block one inserts, and just cannot replace with common much smaller prediction piece.In other words, code efficiency for based on the self adaptation frame inner updating or immediately the system of frame inner updating fix.For wireless connections, packet loss rate is different constantly, and wherein, schemes such as AIR can not reflect packet loss rate so that best performanceization.In other words, the error protection of AIR can not improve (scalable).Under good condition of contact, because piece in the frame that is inserted, quality is not best.

The video flowing server can be adapted to different condition of contacts and different network types, and for example, wired and wireless network is very important.The bit stream handover scheme that uses a plurality of bit streams provides a kind of to server and has allowed server adapt to the condition of contact that changes and need not the low-complexity method of recompile video content, and wherein the recompile video content needs high computing power.But, occur in predictive frame if switch, then flow to the switching generation pixel drifting problem of another bit stream from a bit.Because reference frame is taken from another bit stream, does not match and will propagate, thereby reduced video quality.

The problem that bit stream switches is that switching point must be a frame (key frame) in the frame, otherwise the pixel that all will occur reducing video quality in next frame before the frame does not match.During the video flowing session, wish that switching can appear on any one frame.But, realize this system and do not cause the remarkable reduction of code efficiency and be not easy.

Frame can be used to provide switching point in the frame of rule.But the frequency of frame is high more in the frame, and needed bit is many more, and this will reduce video quality.A kind of scheme provides additional bit stream, and wherein framing bit is in the specific period in all frames, and for example one second, and between transfer period, frame will be used for switching in the frame, this will minimize prediction error.Another simple technology is to switch on any frame, and this causes very significantly pixel drift usually.

Compressed video frame by forming particular type and on allowing to switch to from a bit stream position of another bit stream the frame of particular type being inserted into the video bit stream can be realized correct (coupling) switching between the video flowing.Patent application WO02054776 has described and has been used for making system to carry out switching to another bit stream and need not at switching position frame in bit stream inserts frame from a bit stream.The compressed video frame of particular type will be called as S frame (switching) usually.More specifically, the S frame can be divided into SP frame (handoff predictions), it is to use the motion compensated prediction that draws from the frame that uses the motion vector information decoding to form in decoder, and the SI frame, space (in the frame) prediction that its neighbor that is use has been decoded in decoder from the frame of decoding draws forms.Generally speaking, the S frame is to form on one one basis, and can comprise interframe encode (SP) piece and intraframe coding (SI) piece (in the switch frame).

Switching between the frame permission bit stream of particular type not only appears at the position of I frame, also appears at the position of SP frame.The code efficiency of SP frame is far superior to the code efficiency of typical I frame, wherein needs bandwidth still less will to have the bit stream of SP frame on the position that be used to be transmitted in the I frame.Can on the position in the SP frame being placed on coded bit stream, carry out the switching that a bit flows to another bit stream.

Summary of the invention

The object of the present invention is to provide a kind of new method and system that is used for transmitting video image.The present invention is based on such idea: use intra-frame macro block or SI macro block to replace some macro block of SP frame.Successive frame is repeated this process, make after the continuous SP frame of transmission and some number of having decoded, use intra-frame macro block to replace all basically macro blocks in frame zone (image).This expression uses intra-frame macro block or SI macro block that the entire image zone is upgraded.In an example implementation mode of the present invention, to be that a segment a slice is disconnected carry out the replacement process, until the frame of having revised enough numbers.In other words, in the SP frame that will revise, replace the macro block of first segment, in second frame, replace the macro block of second segment with intra-frame macro block with intra-frame macro block (SI segment), or the like.In the n frame, replace the macro block of last segment with intra-frame macro block.But in different implementations, the replacement order can be different.Also may use the present invention, make replacement order be not fix but variable.In addition, superseded macro block number also may not need more than one, just, in some cases, replace a macro block with another macro block.For example, this replacement can only comprise under the situation that a macro block and that macro block replace with the macro block of another type in segment and uses.

Use intra-frame macro block systematically to upgrade entire image according to said system frame inner updating of the present invention (SIR) scheme, thereby stoped owing to the error propagation that causes is lost in grouping.The bit stream that this scheme can also be used on any frame switches.When switching, the SI segment is used for switch frame first frame of some number afterwards, makes all macro blocks that are used for entire image all be updated, and promptly has inter-coded macroblocks (piece in the frame).

According to a first aspect of the invention, provide a kind of method that is used for transmission of video information, wherein form at least one bit stream from the video information that comprises a framing, frame comprises macro block, and wherein, this method comprises:

-at least one switch frame of formation in described bit stream;

-macro block of described switch frame is divided into first and second groups of macro blocks;

-by first coding method each macro block in described first group of macro block is encoded, to be provided for using another bit stream that forms from video information to continue the switching point that video information is transmitted; And the macro block in described second group of macro block is encoded by another kind of coding method.

According to a second aspect of the invention, provide a kind of encoder, be used for video information coding is entered at least one bit stream, video information comprises one group of frame that comprises macro block, and this encoder comprises:

-be used for forming the device of at least one switch frame at described bit stream;

-be used for the macro block of described switch frame is divided into the apparatus for grouping of first and second groups of macro blocks;

-the first code device is used for by first coding method each macro block of described first group of macro block being encoded, to be provided for using another bit stream that forms from video information to continue the switching point that video information is transmitted; And

-the second code device is used for by another kind of coding method the macro block of described second group of macro block being encoded.

According to a third aspect of the invention we, a kind of transmission system that is used for transmission of video information is provided, this system comprises and being used for the encoder of video information coding at least one bit stream, be used for bit stream is transferred to the transmitter of receiver, and be used for decoder that the bit stream that transfers to receiver is decoded, video information comprises one group of frame that comprises macro block

This encoder comprises:

-the second code device is used for by another kind of coding method the macro block of described second group of macro block being encoded;

This decoder comprises

-the first decoding device is used for by first coding/decoding method corresponding to first coding method each macro block of described first group of macro block being decoded; And

-the second decoding device is used for by second coding/decoding method corresponding to second coding method each macro block of described second group of macro block being decoded.

According to a forth aspect of the invention, but provide a kind of computer program that comprises the machine execution in step that is used for transmission of video information, wherein from the video information that comprises a framing, form at least one bit stream, frame comprises macro block, wherein, but this computer program further comprises the machine execution in step, is used for:

-at least one switch frame of formation in described bit stream;

-by first coding method each macro block in described first group of macro block is encoded, to be provided for using another bit stream that forms from video information to continue the switching point that video information is transmitted; And

-by another kind of coding method the macro block in described second group of macro block is encoded.

According to a fifth aspect of the invention, provide a kind of method that video information is transmitted the influence of transmission error that is used for reducing, wherein form at least one bit stream from the video information that comprises a framing, frame comprises macro block, and wherein, this method comprises:

-carry out predictive coding by macro block to frame, in described bit stream, form at least one SP coded frame;

-usefulness adopts the part in the coded macro block replacement SP coded macroblocks of inner frame coding method; And

-transmission comprises the coded frame of predictive coding macro block and inter-coded macroblocks, but not transmits described SP coded frame.

According to a sixth aspect of the invention, but provide a kind of computer program that comprises the machine execution in step of the influence that is used for reducing video information transmission transmission error, wherein from the video information that comprises a framing, form at least one bit stream, frame comprises macro block, wherein, but this computer program further comprises the machine execution in step, is used for:

According to a seventh aspect of the invention, provide a kind of signal that is used for transmission of video information, wherein form at least one bit stream from the video information that comprises a framing, frame comprises macro block, and wherein, this signal comprises:

-at least one switch frame;

-be divided into the macro block of the described switch frame of first and second groups of macro blocks;

Usually be better than the AIR scheme of prior art according to the code efficiency of the inventive method, this is because SP macroblock encoding efficient is better than intra-frame macro block usually.Also measure, the feasible recovery from packet loss of the method according to this invention is usually faster than AIR.The method according to this invention can also be used for bit stream to be switched, and AIR hereto purpose be not very suitable.

Compare with handover scheme in the frame of prior art, it is then less that frame has the SP frame of a SI segment very greatly in each frame, so between transfer period, frame need increase transmission rate in the frame, and SI segment occupied bandwidth more equably.

The present invention can provide a kind of scalable (scalable) error protection that is used for bit stream, and it improves the video quality in the transmission under any packet loss condition usually.The present invention also provide a kind of be used for drifting in extremely small pixel carry out the bit stream device for switching on any frame.

The present invention improves mistake usually and recovers, and compares with described prior art solution, has quickened the recovery from packet loss.

In conjunction with the accompanying drawings with reference to following description, these and other characteristic of each execution mode of the present invention, aspect and advantage will become more obvious.But, should be appreciated that accompanying drawing only is designed for illustration purpose, be not to limit the invention, restriction of the present invention should be with reference to appended claims.

Description of drawings

Below, will with reference to the accompanying drawings the present invention be described in further detail, wherein

Figure 1A-1C and Fig. 2 are the schematic diagrames of the frame of video coding/compression of expression prior art,

Fig. 3 a is the explanation that the exemplary frames that the method according to this invention encodes is used in expression,

Fig. 3 b is the explanation that expression comprises the frame sequence example of the frame that use the method according to this invention is encoded,

Fig. 4 is the explanation that another exemplary frame sequence that the method according to this invention encodes is used in expression,

Fig. 5 is that the expression use is being used the explanation of switching between two different bit streams of SP/SI frame according to the present invention,

Fig. 6 is the block diagram of the encoder of an example embodiment according to the present invention, and

Fig. 7 is the block diagram of the decoder of an example embodiment according to the present invention, and

Fig. 8 is the system block diagram of the example embodiment according to the present invention.

Embodiment

Below, will an example embodiment of the present invention be described in further detail with reference to the system 1 of figure 8 and the frame of Fig. 3 a and 3b.In encoder 3, always the vision signal from video source 2 forms one or more bit streams.Vision signal can be to comprise a plurality of images, i.e. any digital video signal of image sequence.If formed a plurality of bit streams, then wherein each is used to the different coding parameter of small part and is encoded from same vision signal.For example, can change bit rate, like this, can form bit stream with different bit rates by selecting different coding parameters.For example, coding parameter can be that frame rate, quantization parameter, spatial resolution or another influence the factor of image size.Encoder 3 also inserts frame 10 at least one frame in each bit stream.Usually, first frame of each bit stream is suitably frame in the frame at least.This makes decoder 8 can begin the reconstruction video signal.3 pairs of I frames from vision signal of encoder, P frame, B frame, SP frame and SI frame are encoded.

Encoder 3 also inserts the frame (P frame and optional B frame) that uses motion compensated predictive coding to encode in bit stream.Encoder is also allowing to insert SP frame 11-19 on the position of carrying out switching between the different bit streams in each bit stream.The SP frame can will insert in art methods on the position of intracoded frame and use, and use when perhaps using intracoded frame in video sequence.Different bit streams for example transfers to streaming server 5 from transmitter 4.In streaming server 5, bit stream can be stored in the memory 6 for future use.But, also may after coding, carry out immediately basically to the transmission of receiver 7, wherein do not need to store whole video sequence, but the reference frame of storage necessity is just enough.The transmission of encoded video streams can for example be undertaken by streaming server 5.Transmission server 5 can also have a kind of device that is used for that bit stream transferred to the transmission network (not shown) and/or directly transfers to receiver 7.

In the example of Fig. 3 a, the QCIF image uses as an example of coded image.The size of QCIF image is the 176*144 pixel, is arranged as the row of 11 macro blocks of 9 every row.Row can also be called as segment or macro block group (GOB).In this example, each macro block comprises the 16*16 pixel.The frame 11-19 of Fig. 3 a can form in encoder 3 in the following manner.Encoder 3 is encoded to the macro block of first segment 11.1 of a P frame 11 by using intraframe coding, and wherein first segment 11.1 comprises piece (being 11 macro blocks in 99 macro blocks in this case) in the frame.Encoder 3 is encoded to the macro block of other segment 11.2-11.9 by using some predictive coding, to form the prediction piece, for example, P or B piece.When 3 pairs of next frames 12 of encoder are encoded, another segment, for example, second segment 12.2 is intraframe codings, and comprises that other segment of first segment 12.1 uses predictive coding method to encode.Next, the 3rd frame 13 is encoded, make another segment again (just, being not first segment 13.1 and second segment 13.2) of the 3rd frame 13 carry out intraframe coding, and all other segments use the predictive coding method to encode.Repeat this process, all carried out an intraframe coding at least until all segments of image basically.In the QCIF example images, this needs 9 repetitions, just forms 9 frame 11-19, and one of them segment is intraframe coding, and other segment uses predictive coding method to encode.By doing like this, can be for the QCIF image in 9 frame inner updating entire image.By using the method according to this invention, have only the minority of frame partly to need the intraframe coding of higher bit rate, and the major part of frame is carried out predictive coding.In fact this expression, the present invention does not increase the size of coded frame in the bit stream significantly, and can recover faster than the art methods execution error.

Encoder 3 is also encoded to frame 10 at least one frame, and is inserted in the bit stream, and making can decoding bit stream, and can be at the receiving terminal reconstructed image.Encoder 3 can further add P frame, B frame, SP frame and SI frame system as prior art system in bit stream.Fig. 3 b has illustrated the coded frame sequence example that comprises the frame of encoding according to the present invention.Sequence comprises frame 10 in one or more frames, and a plurality of predictive frame 11-19 are arranged after it, and its coding is made that all macro blocks of a segment all are inter-coded macroblocks in this frame.Frame 10 can be used as switching point and uses in the frame, for example changing bit rate, to provide and carry out the tram that scene changes, or the like.

According to the present invention, needn't revise all predictive frame.For example,, make decoder 8 decoding bit stream correctly, then can revise if other element of some of network, streaming server 5, decoder 8 or system finds to lose or to have damaged one or more transmission groupings.The element of discovery mistake is for example notified this situation to streaming server 5, and 5 of streaming servers begin to transmit the predictive frame 11-19 of correction, and it comprises the segment that has inter-coded macroblocks.If there be not (for example, encoder 3 do not encode this frame) in this frame in memory 6, streaming server 5 notice encoders 3 then, and ask its according to the present invention correction predictive frame.When all segments are all upgraded, promptly transmitted the correction predictive frame 11-19 of enough numbers from streaming server 5 (perhaps from encoder 3), and receiver 7 has received them, and decoder 8 is when having carried out decoding to them, the visual artifacts that packet loss causes is eliminated or eliminates basically, and normal coding/decoding process can continue.

To carry out the order of intraframe coding be not as mentioned above must be from frame head (first segment) to postamble (last segment) to each segment in the frame, but can be different therewith.In some implementation, order even can be at random for example, is used at the arbitrary form of the order of the flexible macro-block described in the standard (FMO) H.264.Order can also change during cataloged procedure.For example, in first group of correction frame, order is from first to last, in second group of correction frame, order is like this: in first frame of second group of correction frame, second segment comprises inter-coded macroblocks, and in second frame, the 3rd segment is carried out intraframe coding, the rest may be inferred to the frame second from the bottom of second group of correction frame, wherein last segment is carried out intraframe coding, and in the last frame of this second group of correction frame, first segment is carried out intraframe coding.

The present invention can also realize together with switching to another bit stream from a bit stream.The present invention also makes transmission system can regulate frame inner updating speed adaptively.According to H.264 the SP picture and the SI picture of standard is special coded frame, and wherein they can be rebuild fully with another SP or SI frame.This specific character makes the present invention can regulate frame inner updating speed adaptively.The present invention uses update scheme in the above-mentioned system-frame.With reference to figure 4 and Fig. 5, two bit streams 410,420 are encoded, one is used the SP segment that whole sequence is encoded, and another uses the SI segment to encode, and wherein this SI segment is accurately duplicating of all SP segments.Under the above-mentioned sample situation of using the QCIF image, a QCIF image comprises the 176*144 pixel, is arranged as the macro block of 16*16 pixel.Therefore, the QCIF image comprises 9 segments, wherein has only one or some segments to use SP/SI macro block according to the present invention to encode.Generally speaking, the intra-frame macro block that comprises the SI macro block needs more bits to encode than the predicted macroblock that comprises the SP macro block.Compare with the intra-coded slice phase failure, the SP coded slices is littler.Because each SP macro block can replace with SP or SI macro block, and can not cause any pixel drifting problem, so during the words that fail to be convened for lack of a quorum, can be used for stream with the bit stream of SP fragment encoding and deliver to client (receiver 7), when streaming server 5 detected packet loss, the SI segment can replace the SP segment to hide mistake.Normally, under the situation of QCIF, can after 9 frames, recover by the SI segment to any damage of image.May dispose the SI segment randomly based on the speed of lost packets.The present invention is that than the advantage of AIR during the good network condition, the code efficiency of SP segment is better usually, and for relatively poor network condition, update scheme can be recovered mistake usually quickly in the system-frame.

Below, description is according to the method that is used for this realization of an example embodiment of the present invention.Encoder 3 forms two different coded frame 411-415,421-425 (only having represented some frames among Fig. 4) from same pictorial information.First framing 410 uses SP to encode, and promptly each segment of describing in Fig. 4 among the frame 411-415 is the SP coded slices, segment of every in the drawings frame.The coding of second framing 420 makes that for example each frame among the 421-425 segment is carried out the SI coding, and other segment of these frames is carried out the P coding.For example, two framings 410,420 can be stored in the memory 6 of streaming server 5, with when client device (receiver) is asked immediately or after a period of time after a while, be delivered to client (receiver).Encoder 3 may be that P and/or B frame coding enters in the bit stream also with frame in one or more frames.In normal transmission, if transmission SP coded frame 411-415 and streaming server 5 detect between the frame transmission period mistake takes place, then it begins to transmit each frame (the coded frame 421-425 that just comprises one or more SI coded slices) of second framing 420 but not each frame of first framing.When the frame in second framing 420 that has transmitted enough numbers, streaming server 5 can switch to each frame of transmission first framing 410.

Use the problem of the SIR of fragment encoding in the frame to be, the observer may feel a kind of disturbing effect, and a scrolling segment is over and over again rolled to afterbody from the head of image exactly.This problem also can exist for the SP coded frame that comprises the SI coded slices, still, influences not obvious and it takes place for a SP/SI frame.For SP/SI segment scheme, preceding 9 frames of QCIF sized images will show the influence similar to the SIR situation, and still, it will can not show any visual artifacts usually after that.A kind of ways of addressing this issue is exactly in frame frame (normally scene change frame) the SP frame of encoding at once afterwards.

The bit stream of encoded video streams needs a lot of key frames (generally being frame in the frame), to allow fast forword/back to operation and index.With reference to figure 5, scene changes may use the interior frame 510,519 of frame to encode, and between the frame, can insert a plurality of SP frames 514,517 in these two frames, to carry out fast speed playback, search, bit stream switching and error concealment, this is because can use the SI frame to replace the SP frame where necessary.Between it, may be placed with SP/SI frame 512,513,515,516,518, to carry out error concealment and urgent the switching.

In example embodiment of the present invention, at first SP segment and SP frame are encoded, be SI segment and SI frame then.The added bit stream that comprises SI segment and frame can be stored with main SP bit stream.Every group of bit stream comprises main bit stream and SI bit stream, and with different bit rates all main bit streams of every group encoded, to be used for different connection speeds.

Fig. 5 has described the part of first bit stream 510 that forms and the part of second bit stream 520 in encoder 3.Some frames in each bit stream have only been represented.Particularly, shown in first bit stream 510 comprise I frame 511,519, SP frame 514,517 and SP/SI frame 512,513,515,516,518, and shown in second bit stream 520 comprise corresponding I frame 521,529, SP frame 524,527 and SP/SI frame 522,523,525,526,528.Here should be noted that for clear, do not show all the SP/SI frames between the SP frame.Suppose, two bit streams 510 with 520 corresponding to for example same sequence by using different frame rate, different spatial resolution or different quantization parameters to encode with different bit rates.Suppose that further first bit stream 510 transfers to decoder 8 (Fig. 7) from transmission server 5 by the transmission network (not shown), and transmission server 5 receives the request of the bit rate that will change the video flowing that is transmitting from transmission network.

During cataloged procedure, the SP frame is placed on and is arranged in those permissions of video sequence switch to another bit stream from a bit stream position in the bit stream.When being encoded to the video sequence frame of SP frame 514 in transmission server 5 arrival first bit stream 510, it can begin operations necessary continues video flowing with the coded frame of using second bit stream 520 transmission.In that, transmission server 5 has transmitted each frame before the SP frame in first bit stream 510 514, decoder 8 corresponding each frame that received and decoded.Thereby those frames have been stored in the frame memory 750 of decoder 8.Frame memory 750 comprises enough memories, rebuilds P frame or the required all that frame of B frame, the necessary information of promptly rebuilding all required reference frames of present frame with storage.

Transmission server 5 is carried out following operation continues video flowing with the coded frame of using second bit stream 520 transmission.For example by checking the type information of frame, the present frame that discovery will be transmitted is the SP frame to transmission server 5, so may switch between bit stream.Certainly, only, just switch receiving some other reason that this request that will do or existence are switched.Transmission server 5 is imported the corresponding SP frame 524SP of second bit stream, and SP frame 524SP is transferred to decoder 8.SP frame 524SP is a predictive frame, uses frame 513 conducts with reference to frame, to rebuild SP frame 524.After transmission SP frame 524SP, transmission server 5 continues the coded frame of transmission second bit stream 520, just the SP/SI frame 525,526 after the SP frame 524SP, other SP frame 527 or the like.

Next, the S frame be positioned at bit stream is described, for example, S frame 513 and 523 coding.

Usually, on one one basis, construct according to SP/SI frame of the present invention, for example, the frame 512,522,513,523 among Fig. 5.As previously described, a chunk, for example segment is encoded in a kind of like this mode of the spatial coherence between image encoded (in the frame or SI piece) pixel of just having utilized.The coded system of other piece is encoded in the mode of the temporal correlation between a kind of like this block of pixels of utilizing video sequence (interframe or SP piece) successive frame.

Will be with reference to the coding of figure 6 descriptions according to S frame of the present invention, Fig. 6 is the block diagram according to the frame encoder 3 of first embodiment of the invention.

The frame of video of encoding is piecemeal at first, is piece in SP piece, SI piece or the frame then with each block encoding.Switch 690 is used for carrying out correct switching between SI and SP coding mode, and promptly switch 690 is a kind of structures of using in the present invention describes, must not be physical equipment.Under the SP coding mode, switch 690 is used to current block 670 to obtain motion compensated prediction.Motion-compensated prediction block 670 with to used similar mode from the known motion compensated prediction of prior art for current block in the frame of encoding form prediction P (x, y).More specifically, motion-compensated prediction block 670 is by determine describing in the current block pixel and remain on the motion vector that concerns between the pixel value of the reconstruction reference frame in the frame memory 646, for current block in the frame of encoding form prediction P (x, y).

Under the SI coding mode, switch 690 is used to current block acquisition prediction from the frame that intra-frame prediction block 680 is encoded.Intra-frame prediction block 680 with to used similar mode from the known infra-frame prediction of prior art for current block in the frame of encoding form prediction P (x, y).More specifically, intra-frame prediction block 680 is used the spatial prediction from neighbor encoded in the frame of encoding, for current block in the frame of encoding form prediction P (x, y).

Under SP and SI coding mode, prediction P (x, y) form of employing pixel value piece.Forward transform, for example discrete cosine transform (DCT) are applied to pixel value P in the piece 660 (x, prediction piece y), and resulting conversion coefficient is expressed as c _Pred, in quantize block 650, quantize subsequently, to form quantized transform coefficients I _PredAlso raw image data is carried out corresponding operation.More specifically, the current block with the pixel value of the original image of encoding is applied to transform block 610.Forward transform (for example, DCT) is applied to the pixel value of original picture block, to form conversion coefficient c here, _OrigThese conversion coefficients are sent to quantize block 620, there, it are quantized, to form quantized transform coefficients I _OrigSummator 630 receives two groups of quantization transform coefficient I from each

quantize block

650 and 620 _PredAnd I _Orig, and generate one group of prediction error coefficient I that quantizes according to following relationship _Err:

I _err＝I _orig-I _pred

With quantitative prediction error coefficients I _ErrBe sent to multiplexer 635.If current block is encoded under SP form/pattern, then multiplexer 635 also receives the motion vector that is used for the SP encoding block.If current block is to encode, then the information that is used for forming for the SI encoding block in the intra-frame prediction block 680 the relevant intra prediction mode of prediction is sent to multiplexer under SI form/pattern.Aptly, variable length code is applied to quantitative prediction error coefficients I _Err, and be applied to intra prediction mode information in motion vector or the multiplexer 635, multiplexing forming bit stream by the information of various ways is carried out, thereby and the bit stream that will form transfer to corresponding decoder 8 (with reference to figure 7).

S frame encoder 3 according to the present invention also comprises the local decode function.The predictive transformation coefficient I of the quantification that will in quantize block 650, form _Pred Offer summation component 640, it also receives and quantizes error coefficients I _ErrThe predictive transformation coefficient I that summation component 640 reorganization quantize _PredWith quantitative prediction error coefficients I _Err, to form one group of quantized transform coefficients I that rebuilds according to following relationship _Rec:

I _rec＝I _pred+I _err.

The quantized transform coefficients of rebuilding is sent to inverse quantisation block 642, and it carries out inverse quantization to the quantized transform coefficients of rebuilding, to form the reconstruction conversion coefficient d of inverse quantization _RecFurther the reconstruction conversion coefficient with inverse quantization is sent to inverse transform block 644, there, its will carry out with piece 660 in the performed corresponding inverse discrete cosine transformation of conversion (IDCT) or any other inverse transformation.Like this, rebuild the pixel value piece for the image block of discussing forms, and it is stored in the frame memory 646.Owing to experienced foregoing coding and local decode operation with follow-up each piece in the frame of S coded, decoded version with present frame concentrates in the frame memory gradually, can conduct interviews and can be in same frame uses in interframe (motion compensation) prediction of subsequent frame in the infra-frame prediction of follow-up each piece or the video sequence it therefrom.

The operation of the interchangeable frame decoder of an example embodiment according to the present invention is described referring now to Fig. 7.

Before the bit stream that generates of the frame encoder of describing in conjunction with Fig. 6 was received by decoder 8, and was its part by demodulation multiplexer 710 demultiplexings.Decoder is rebuild the decoded version of SP/SI frame on one one basis.As mentioned above, the SP/SI frame comprises SP coding and SI coded picture block.For SP format-pattern piece, the information in the bit stream that is received comprises the kinematic coefficient information of VLC coding and the prediction error coefficient I of the quantification that VLC encodes _ErrFor carrying out the image encoded piece with the SI form, the information that is received in the bit stream comprises the quantitative prediction error coefficients I that VLC coded message and VLC encode _Err, wherein this VLC information encoded is relevant with the intra prediction mode that is used to form SI encoding block infra-frame prediction.

When decoding SP encoding block, demodulation multiplexer 710 is at first used correct length-changeable decoding (VLD) to the bit stream that is received, to recover the prediction error coefficient I of motion vector information and quantification _ErrIt is with the prediction error coefficient I of motion vector information and quantification then _ErrSeparate.Motion vector information is offered motion-compensated prediction block 760, and the prediction error coefficient of the quantification that will recover from bit stream is applied to an input of summator 720.The pixel value of the previous reconstruction frames that motion vector information keeps in frame memory 750 in motion-compensated prediction block 760 uses, with encoder 3 in used similar forms form prediction P (x, y).

When decoding SI encoding block, the prediction error coefficient I of 710 pairs of intra prediction mode information that received of demodulation multiplexer and quantification _ErrUse length-changeable decoding.Then intra prediction mode information is separated from the prediction error coefficient that quantizes, and provided it to intra-frame prediction block 770.With the prediction error coefficient I that quantizes _ErrOffer an input of summator 720.The pixel value of the early decoding of the present frame that intra prediction mode information keeps in frame memory 750 in intra-frame prediction block 770 uses, think the current block of decoding form prediction P (x, y).Once more, performed in the infra-frame prediction process of in decoder 8, carrying out and the encoder 3 with previously described similar.

In case formed the prediction of current block in the frame of decoding,, made that (x y) offers transform block 790 to the prediction P that will comprise predicted pixel values with regard to console switch 780.Once more, switch 780 is the abstract structures that use in description of the invention, needs not to be physical equipment.Under the situation of SP encoding block, switch 780 is used to connect motion-compensated prediction block 760 and transform block 790, and under the situation of SI encoding block, it is used to connect intra-frame prediction block 770 and transform block 790.

In piece 790, with forward transform, for example, discrete cosine transform (DCT) is applied to pixel value P, and (x, y) prediction piece is and with resulting conversion coefficient c _Pred Offer quantize block 795, there, it is quantized to form quantized transform coefficients I _PredThen with quantized transform coefficients I _PredOffer second input of summator 720, there, they are added to prediction error coefficient I _ErrIn, to form the quantized transform coefficients I that rebuilds according to following relationship _Rec:

I _rec＝I _pred+I _err.

Further with the quantized transform coefficients I that rebuilds _RecOffer inverse quantisation block 730, there, it is carried out inverse quantization, to form the reconstruction conversion coefficient d of inverse quantization _RecThen with the conversion coefficient d of inverse quantization _RecBe sent to inverse transform block 740, there, its will carry out with piece 790 in the performed corresponding inverse discrete cosine transformation of conversion (IDCT) or any other inverse transformation.By such manner, form the pixel value piece of rebuilding for the image block of discussing.The pixel value of rebuilding is offered video output and frame memory 750.Because follow-up each piece has experienced foregoing decode operation in the S frame of decoding, decoded version with present frame concentrates in the frame memory 750 gradually, can conduct interviews and uses in interframe (motion compensation) prediction of subsequent frame in the infra-frame prediction of follow-up each piece or the video sequence in same frame it therefrom.

First embodiment of the invention has been described the 26S Proteasome Structure and Function of S frame encoder, may understand now according to S frame of the present invention and how carry out the switching between the bit stream and can not produce the mistake that do not match, for example, in existing encoding and decoding of video system, run into.Refer again to the bit stream in Fig. 5, described and switch example, appear at from the switching of first bit stream, 510 to second bit streams 520 on the

SP frame

514 and 524 position of each bit stream.As mentioned above, when carry out switching, one group of SP/SI coded frame is encoded and transmitted, so that the quick renewal to frame to be provided.

Except transmission network, also may initiate by other parts of transmission system to the request of bit flow transmission characteristic variations.For example, receiver can change parameter because of certain reason request transmission server.This request is sent to transmission server by transmission network.

Though H.264 use as an example of standard, each execution mode of the present invention and any distortion and correction are considered to be located within the scope of the present invention.

It not is the adaptable unique application of the present invention that bit stream switches.If one of them bit stream has lower temporal resolution, for example, 1 frame/second, then this bit stream can be used to provide the fast forword function.Particularly, decode, and switch to bit stream then this function will be provided with normal frame speed from bit stream with low temporal resolution.Fig. 8 illustrates two bit streams, wherein second the S frame of interval from predicting each other that includes only with the frame recurrence interval that is higher than first bit stream.In addition, " fast forword " can begin on any position of bit stream and stop.Below, some other application of the present invention is described.

Previously described bit stream switches the bit stream that example considers to belong to same image sequence.But, be not all situations that needs bit stream to switch all be like this.Example comprises: from catching same incident but from different perspectives different cameral or place switching between the bit stream that obtains with the camera that monitors around building; In television broadcasting, video connection or the like, switch to this locality/national program or insert advertisement.The generic term of process that is used to connect coded bit stream is for engaging.

The above-mentioned adaptive error that the invention provides a kind of SP/SI of use coding mode recovers instrument and bit stream handover scheme.Clearly, the present invention is not limited to above-mentioned execution mode, but can revise within the scope of the appended claims.For example, can replace a plurality of chunks in the SP frame with the SI coded macroblocks.

Claims

1. a method that is used for transmission of video information wherein forms at least one bit stream from the video information that comprises a framing, and this frame comprises macro block, and wherein, this method comprises:

In described bit stream, form at least one switch frame;

The macro block of described switch frame is arranged in first group of macro block and the second group of macro block;

By first coding method each macro block in described first group of macro block is encoded, to be provided for using another bit stream that forms from this video information to continue the switching point that video information is transmitted; And

By another kind of coding method the macro block in described second group of macro block is encoded.

2. method according to claim 1 comprises and uses inner frame coding method that described first group of macro block encoded.

3. method according to claim 2 comprises and uses the predictive coding method that described second group of macro block encoded.

4. method according to claim 1, comprise that the described macro block with described switch frame is arranged into the segment group, and the macro block of one of them segment of described segment group is arranged to described first group of macro block, and the macro block of other segment of described segment group is arranged to described second group of macro block.

5. method according to claim 1 comprises

Form at least the first switch frame and second switch frame in described bit stream, this switch frame is divided into similar macro block group mutually, each macro block of this first switch frame has the space correspondence in described second switch frame macro block;

The macro block of described first switch frame is arranged in first group and the second group of macro block;

The macro block of described second switch frame is arranged in the 3rd group and the 4th group of macro block, make macro block in described the 3rd group of macro block be with described first group of macro block in the spatially different macro block of macro block;

By first coding method each macro block in described first group and described the 3rd group of macro block is encoded, continue the switching point of video information transmission to be provided for using described other bit stream that forms from video information; And

By another kind of coding method the macro block in described second group and described the 4th group of macro block is encoded.

6. method according to claim 1, comprise that a frame from a described framing forms intracoded frame, form the handoff predictions coded frame from the frame of following after described intracoded frame, and form described at least one switch frame from the frame of following after described handoff predictions coded frame.

7. one kind is used for the encoder of video information coding at least one bit stream, and this video information comprises one group of frame that comprises macro block, and this encoder comprises:

Be used for forming the device of at least one switch frame at described bit stream;

Be used for the macro block of described switch frame is arranged into the apparatus for grouping of first group and second group macro block;

First code device is used for by first coding method each macro block of described first group of macro block being encoded, to be provided for using another bit stream that forms from this video information to continue the switching point that video information is transmitted; And

Second code device is used for by another kind of coding method the macro block of described second group of macro block being encoded.

8. encoder according to claim 7, wherein said first coding method is an inner frame coding method.

9. encoder according to claim 8, wherein said second coding method is the predictive coding method.

10. encoder according to claim 7, comprise a kind of device, this device is used for the described macro block of described switch frame is arranged to the segment group, and be used for the macro block of one of them segment of described segment group is arranged to described first group of macro block, and be used for the macro block of other segment of described segment group is arranged to described second group of macro block.

11. encoder according to claim 7 comprises

Be used for forming at described bit stream the device of at least the first switch frame and second switch frame, this switch frame is divided into similar macro block group mutually, each macro block of this first switch frame has the space correspondence in described second switch frame macro block;

Be used for the macro block of described first switch frame is arranged into the device of first group and second group macro block;

A kind of device, this device are used for the macro block of described second switch frame is arranged into the 3rd group and the 4th group of macro block, make macro block in described the 3rd group of macro block be with described first group of macro block in the spatially different macro block of macro block;

A kind of device, this device are used for by first coding method each macro block of described first group and described the 3rd group of macro block being encoded, and continue the switching point of video information transmission to be provided for using another bit stream that forms from video information; And

Be used for the macro block of described second group and described the 4th group of macro block being carried out apparatus for encoding by another kind of coding method.

12. encoder according to claim 7, comprise a kind of device, this device is used for forming intracoded frame from the frame of a described framing, be used for forming the handoff predictions coded frame, and be used for forming described at least one switch frame from the frame of following after described handoff predictions coded frame from the frame of following after described intracoded frame.

13. a transmission system that is used for transmission of video information, this system comprises: encoder is used for video information coding at least one bit stream; Transmitter is used for this bit stream is transferred to receiver; And decoder, being used for the bit stream that transfers to this receiver is decoded, this video information comprises one group of frame that comprises macro block,

This encoder comprises:

Be used for macro block with described switch frame and be arranged into apparatus for grouping in first group and the second group of macro block;

First code device is used for by first coding method each macro block of described first group of macro block being encoded, to be provided for using the switching point that continues the transmission of video information from another bit stream of this video information formation; And

Second code device is used for by another kind of coding method the macro block of described second group of macro block being encoded;

This decoder comprises

First decoding device is used for by first coding/decoding method corresponding to this first coding method each macro block of described first group of macro block being decoded; And

Second decoding device is used for by second coding/decoding method corresponding to this second coding method each macro block of described second group of macro block being decoded.

14. system according to claim 13, wherein said first coding method is an inner frame coding method.

15. system according to claim 14, wherein said second coding method is the predictive coding method.

16. system according to claim 13, comprise a kind of device, this device is used for the described macro block of described switch frame is arranged to the segment group, and be used for the macro block of one of them segment of described segment group is arranged to described first group of macro block, and be used for the macro block of other segment of described segment group is arranged to described second group of macro block.

17. system according to claim 13 comprises

A kind of device, this device are used for by first coding method each macro block of described first group and described the 3rd group of macro block being encoded, and continue the switching point of the transmission of video information to be provided for using another bit stream that forms from this video information; And

18. system according to claim 13, comprise a kind of device, this device is used for forming intracoded frame from a frame of a described framing, be used for forming the handoff predictions coded frame, and be used for forming described at least one switch frame from the frame of following after described handoff predictions coded frame from the frame of following after described intracoded frame.

19. but computer program that comprises the machine execution in step that is used for transmission of video information, wherein form at least one bit stream from the video information that comprises a framing, this frame comprises macro block, wherein, but this computer program further comprises the machine execution in step, is used for:

In described bit stream, form at least one switch frame;

The macro block of described switch frame is divided into first group and second group of macro block;

20. computer program according to claim 19, but comprise and be used to the machine execution in step of using inner frame coding method that described first group of macro block encoded.

21. computer program according to claim 20, but comprise and be used to the machine execution in step of using the predictive coding method that described second group of macro block encoded.

22. computer program according to claim 19, but comprise following machine execution in step, be used for the described macro block of described switch frame is arranged to the segment group, and the macro block of one of them segment of described segment group is arranged to described first group of macro block, and the macro block of other segment of described segment group is arranged to described second group of macro block.

23. computer program according to claim 19, but comprise following machine execution in step, be used for:

Form at least the first switch frame and second switch frame in described bit stream, this switch frame is divided into similar macro block group mutually, wherein, and each macro block of this first switch frame has the space correspondence in described second switch frame macro block;

By first coding method each macro block in described first group and described the 3rd group of macro block is encoded, continue the switching point of video information transmission to be provided for using another bit stream that forms from this video information; And

24. computer program according to claim 19, comprise that a frame from a described framing forms intracoded frame, form the handoff predictions coded frame from the frame of following after described intracoded frame, and form described at least one switch frame from the frame of following after described handoff predictions coded frame.

25. one kind is used for reducing the method that video information is transmitted the influence of transmission error, wherein forms at least one bit stream from the video information that comprises a framing, each frame comprises macro block, and wherein, this method comprises:

Carry out predictive coding by macro block, in described bit stream, form at least one handoff predictions coded frame each frame;

With the part that adopts in the coded macro block replacement handoff predictions coded macroblocks of inner frame coding method; And

Transmission comprises the frame of predictive coding macro block and inter-coded macroblocks, but not transmits described handoff predictions coded frame.

26. computer program, but comprise and be used for reducing the machine execution in step that video information is transmitted the influence of transmission error, wherein from the video information that comprises a framing, form at least one bit stream, each frame comprises macro block, wherein, but this computer program further comprises following machine execution in step, is used for:

With a part that adopts in the coded macro block replacement handoff predictions coded macroblocks of inner frame coding method; And

Transmission comprises the coded frame of predictive coding macro block and inter-coded macroblocks, but not transmits described handoff predictions coded frame.

27. a signal that is used for transmission of video information wherein forms at least one bit stream from the video information that comprises a framing, each frame comprises macro block, and wherein, this signal comprises:

At least one switch frame;

Be arranged to the macro block of the described switch frame in first group and the second group of macro block;

28. signal according to claim 27 wherein uses inner frame coding method that described first group of macro block encoded.

29. signal according to claim 28 wherein uses the predictive coding method that described second group of macro block encoded.

30. signal according to claim 27, wherein, the described macro block of described switch frame is arranged to the segment group, and the macro block of one of them segment of described segment group is arranged to described first group of macro block, and the macro block of other segment of described segment group is arranged to described second group of macro block.

31. signal according to claim 27 comprises

At least the first switch frame that forms in described bit stream and second switch frame, this switch frame are divided into mutually similar macro block group, wherein each macro block of this first switch frame macro block of having the space correspondence in described second switch frame;

The macro block of described first switch frame is divided into first group and second group of macro block;

The macro block of described second switch frame is divided into the 3rd group and the 4th group of macro block, makes, the macro block in described the 3rd group of macro block be with described first group of macro block in the spatially different macro block of macro block;

32. signal according to claim 27, wherein, a frame from a described framing forms intracoded frame, form the handoff predictions coded frame from the frame of following after described intracoded frame, and form described at least one switch frame from the frame of following after described handoff predictions coded frame.