CN106993237A - Dynamic self-adapting code rate selection method based on MPEG DASH agreements - Google Patents

Abstract

The invention discloses the dynamic self-adapting code rate selection method based on MPEG DASH agreements, the HTTP connections of server and client are initially set up；User end to server sends HTTP Get requests, and request is downloaded MPD file and stored；Server is received carries out response transmission MPD file to client after the request of client, client is downloaded to MPD file and the MPD file of download is parsed；The URL addresses obtained according to parsing MPD are first downloaded part section with minimum code rate and are put into buffer area, and network bandwidth when being cut into slices to foradownloaded video is calculated, and count and record and network bandwidth when downloading next video slicing is predicted；The code check for downloading next video segment is made a choice with reference to the remaining situation of current cache area video segment, until terminating.Realize and improve video quality as far as possible on the premise of smooth playing, greatly enhance user's viewing experience.

Description

Dynamic self-adapting code rate selection method based on MPEG-DASH agreements

Technical field

The invention belongs to technical field of video transmission, it is related to streaming media video based on MPEG-DASH agreements or audio A kind of transmission, and in particular to dynamic self-adapting code rate selection method based on MPEG-DASH agreements.

Background technology

At present, being continuously increased with the network bandwidth, people are also constantly increasing to the demand of media information.According to Cisco Company shows that, by 2017, media flow will occupy the 86% of whole network flow, all kinds of matchmakers in the report that May in year is issued Body is applied as a part indispensable in people's informatized life.Current various multimedia application layer go out not thoroughly, they Collection audio is provided for people, video, picture and text are experienced in the comprehensive of one.It is intended that obtaining more rapidly, more smooth, higher quality Media services experience, how to meet this demand turn into current stream media technology a big difficult point.

Stream media technology refers to the technology using streaming manner transmission media data, and its main feature is by servicing Device to user it is continuous, transmit media data in real time.In the system using stream media technology, user is without waiting for total data Downloading completion can play out, that is, realize the effect played when downloading, and greatly reduce user and start the time and to user The demand of side cache size.Traditional RTP/RTSP stream media technologys are carried out data transmission using udp protocol, therefore are only capable of providing Insecure transmission.Additionally due to RTP/RTSP stream media technologys need specific streaming media server, cause lower deployment cost compared with It is high.For these problems, HTTP dynamic self-adaptings Streaming Media (DASH, Dynamic Adaptive Streaming over HTTP) technology turns into the new lover of Streaming Media circle in recent years.The technology uses HTTP/TCP protocol infrastructures, provides the user reliable Data transfer.The technology uses 80 ports simultaneously, therefore can easily pass through fire wall.Furthermore, DASH technologies can be weighed With common Web server, lower deployment cost is low.Finally, the media table of the same media content different quality rank of the technical support Show, especially moving end subscriber can be according to capacity of equipment, buffer size and the real-time adjust request of real-time network bandwidth situation Media quality rank, the characteristics of possessing dynamic self-adapting.Therefore, DASH technologies have obtained all kinds of manufacturers and standardization body Extensive concern, is expected to turn into main flow stream media technology of future generation.

But MPEG-DASH agreements define only MPD file structure and Streaming Media section form, and client is not provided The flow or mechanism of end selection section file, to a certain extent dynamic self-adapting code rate selection method based on DASH agreements Quality directly determines the height of streaming media efficiency and the viewing experience of user, so the dynamic based on DASH agreements is adaptive Code rate system of selection has very big research and development space, and does not have similar techniques appearance also now.

The content of the invention

To achieve the above object, the present invention provides a kind of dynamic self-adapting code rate selection side based on MPEG-DASH agreements Method, realizes and improves video quality as far as possible on the premise of smooth playing, greatly enhances user's viewing experience.

The technical solution adopted in the present invention is that the dynamic self-adapting code rate selection method based on MPEG-DASH agreements has Body is followed the steps below：

Step 1, the HTTP connections of server and client are set up；

Step 2, user end to server sends HTTP Get requests, and request is downloaded MPD file and stored；

Step 3, server is received carries out response transmission MPD file to client after the request of client, client is to MPD File is downloaded and the MPD file of download is parsed；In order that video can be commenced play out as early as possible, according to parsing MPD Range of code rates corresponding to the URL of video slicing obtained by file first downloads a part of video segment, service with minimum code rate Device, which is received, gives client transmission video segment file after request, client is downloaded to section file；

Step 4, network bandwidth when being cut into slices to step 3 foradownloaded video is calculated, and is counted and is recorded and uses Kalman Filter forecasting method is predicted to network bandwidth when downloading next video slicing；

Step 5, according to the prediction bandwidth obtained by step 4 and with reference to the remaining situation of current cache area video segment to downloading The code check of next video segment makes a choice；

Step 6, step 4-5 is jumped to if video segment is without completion is all downloaded, continues executing with prediction and downloaded Journey, otherwise exits cyclic process, and method terminates.

Further, the detailed process of the step 1 is：

Step 11：Create a URL object；

Step 12：Create a HTTP link；

Step 13：Data are obtained using IO streams.

Further, the detailed process of the step 2 is：

Step 21：Judge that MPD file whether there is, 0 is otherwise returned if returning to 1 in the presence of if；

Step 22：File is downloaded when file is not present；

Step 23：Stored to downloading the file completed.

Further, the detailed process of the step 3 is：

Step 31：Obtain the specified attribute of Representation elements in MPD file, it is only necessary to obtain id, width, Tri- attributes of height；

Step 32：Obtain the URL of each grade video segment in MPD file；

Step 33：The video segment that minimum code rate is chosen in the URL addresses according to obtained by parsing MPD file is downloaded.

Further, the detailed process of the step 4 is：

Step 41：Obtain the network bandwidth b of downloading slicing_n：

Wherein：t_n(e) represent that the time completed is downloaded in n-th of section；Tn (s) represent n-th section start download when Between；R (n) represents the selected code check of n-th of section；θ represents that duration is played in each section, and its value is 2 seconds；b_nIt is n-th The network download bandwidth value of individual section；

Step 42：Set up system state equation and measurement equation：

Wherein：b_n+1It is the prediction download bandwidth value of (n+1)th section；Y_n+1It is n+1 section bandwidth measuring value；A is shape State transfer matrix, reflects system from the download bandwidth of n-th of section to the state change of (n+1)th；u_nIt is n-th of section Dominant vector；B is control allocation matrix；q_nAnd r_n+1It is system noise and measurement noise, q respectively_nStatistical property E [q_n]=0, r_n+1Statistical property E [r_n+1]=0；H is observing matrix；

Step 43：By the optimum bandwidth value of n-th of burstRemove the state value of (n+1)th burst of predictionDue to mould The random disturbances q of type_nDo not know in advance, soBy the prediction download bandwidth of next burst with it is upper The optimal download bandwidth of one burst be considered as it is identical, even A=1, dominant vector u_n=0, obtain：

Step 44：By last error covariance p_nWith system noise q_nCovariance Q predict new error covariance I.e.

Wherein Q is system noise q_nCovariance, type belongs to white Gaussian noise；For (n+1)th slicer error association side The predicted value of difference, i.e., new error covariance；

Step 45：Calculate kalman gainHere measured value is Calculated value is so H=1 is

Wherein R is measurement noise r_n+1Covariance；

Step 46：It is rightCorrection updates,

NowFor the optimum bandwidth of (n+1)th burst；Y_n+1For the bandwidth measuring value of (n+1)th section；

Step 47：Continue step 48 if completion is not yet all downloaded in section, otherwise exit whole prediction process；

Step 48：Estimate that the iteration of the optimum bandwidth value of the n-th+2 sections is updated operation for next step, that is, update p_n+1Value：

P_n+1It is the optimum bandwidth deviation of (n+1)th section；

Step 49, step 43 is jumped to, repeats to step 47 to terminate always.

Further, the detailed process of the step 5 is：

Step 51：If video slicing is not downloaded all and completed, i.e., there are still prediction bandwidth, then step 52 is performed, otherwise Downloading process terminates；

Step 52：

In order to ensure buffer area occurs without overflow and underflow, at the same ensure buffering area can effectively cancellation bandwidth fluctuation made Into forecasting inaccuracy really bring broadcasting stagnate or buffer overflow possibility, now to two time threshold values of setting buffers q_minAnd q_max, the code to be chosen is determined according to the threshold values of setting buffers and the burst number existed and prediction bandwidth Rate value.

Further, the detailed process of the step 52 is：

Step 521：

Calculate n-th of burst s_nThe remaining duration of buffering area after the completion of download：

T₍t_n(e))=T (t_n(s))+θ-(t_n(e)-t_n(s)) (8)

Wherein T (t_n(e)) represent that n-th of burst has downloaded rear buffering area duration, while being also to start download (n+1)th to cut The duration of buffering area during piece, T (t_n(s)) duration of buffering area when representing to start to download n-th of burst, while being also to have downloaded the The duration of buffering area during n-1 burst；

Step 522：

As T (t_n+1(s)) ＜ q_minWhen buffer area section have the possibility being finished, in order to ensure buffering area has enough regard Frequency section must assure that T (t_n+1(e))≥q_min, then its code check need meet：

The formula equal sign left side is only calculated obtained value and is designated as ψ by ψ without implication, code check r (n+1) value selected by (n+1)th For max (r (1), ψ), but it is due to that code check value is series of discrete value, so final on the premise of ensureing that video playback is smooth Code check value be apart from max (r (1), ψ) it is nearest take fixed code check value downwards, wherein r (1) be when video segment is generated most Small code check；

Step 523：

As T (t_n+1(s)) ＞ q_maxWhen buffering area have the possibility that overflow goes out be loaded with enough sections, continuation, in order to true Protecting buffering area, overflow does not ensure T (t_n+1(e))≤q_max, then its code check need meet：

ξ is without implication, and the value for only obtaining the formula equal sign left side is designated as ξ；

Situation one：R (p) >=ξ is in order to ensure video pictures small shake as far as possible, and the code check value of selection is ξ, simultaneously because Code check value is series of discrete value, so code check value final on the premise of ensureing that video playback is smooth is nearest apart from ξ Fixed code check value is taken upwards；

Situation two：During r (p) ＜ ξ, buffer area overflow must be occurred by continuing download, in order to avoid such case occurs then Start sleep mechanism, that is, a period of time can be suspended by downloading, and n-th of burst s is recalculated after pause_nBuffered after the completion of download The remaining duration in area, proceeds code check judgement, and wherein r (p) is maximal rate when video segment is generated；

Step 524：

Work as q_min≤T(t_n+1(s))≤q_maxWhen, the burst number of buffering area is in an optimal scope, the value model of code check Enclose for：

In order to reduce code check fluctuation, code check value is identical with upper one section code check value holding：R (n+1)=r (n).

Beneficial effects of the present invention：The present invention is the adaptively selected algorithm of dynamic code rate based on DASH agreements, is related to Bandwidth prediction and buffer queue management, realize video quality and change with the dynamic of network condition.Algorithm employs Kalman's filter Ripple predicts bandwidth, greatly improves precision of prediction, and the bandwidth predicted is combined with buffer area restrictive condition, finally Determine the code check of section to be downloaded and download the section file of the code check.It ensure that and carried as far as possible on the premise of smooth playing High video playing quality, greatly enhances user's viewing experience.

Brief description of the drawings

In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing There is the accompanying drawing used required in technology description to be briefly described, it should be apparent that, drawings in the following description are only this Some embodiments of invention, for those of ordinary skill in the art, on the premise of not paying creative work, can be with Other accompanying drawings are obtained according to these accompanying drawings.

Fig. 1 is the client of the embodiment of the present invention and the connection interaction figure of server.

Fig. 2 is the MPD structural representations of the embodiment of the present invention.

Fig. 3 is the flow chart of the code check adaptive algorithm of the embodiment of the present invention.

Embodiment

Below in conjunction with the embodiment of the present invention, the technical scheme in the embodiment of the present invention is clearly and completely retouched State, it is clear that described embodiment is only a part of embodiment of the invention, rather than whole embodiments.Based on the present invention In embodiment, the every other implementation that those of ordinary skill in the art are obtained under the premise of creative work is not made Example, belongs to the scope of protection of the invention.

As shown in figure 1, the embodiments of the invention provide a kind of dynamic self-adapting code rate selection based on MPEG-DASH agreements Method, including following functions：Set up server and HTTP connections, download and storage MPD file, the parsing MPD of client File, code check are adaptively selected.

Thinking is as follows：The HTTP connections of server and client are initially set up, client is from clothes after the completion of connection is set up Business device downloads MPD file and it is stored and parsed；Secondly in order to shorten the broadcasting that buffer time carries out video as early as possible, and Initial data is accumulated for follow-up bandwidth prediction, first one is downloaded with minimum code rate according to the range of code rates obtained by parsing MPD file Partial video is cut into slices and network bandwidth when downloading partial video section is calculated, counted and recorded, and uses karr Graceful filter forecasting method is predicted to network bandwidth when downloading next video slicing；Finally according to prediction resulting above Bandwidth and the remaining situation of current cache area video segment make a choice to the code check for downloading next video segment, so that being broadcast Put video quality and dynamically adapt to different networking situations.

Specifically follow the steps below：

Step 1,

Initially set up the HTTP connections of server and client；Because DASH in application layer is to be used as host-host protocol using HTTP , and HTTP is based on Transmission Control Protocol, so the session of client and server needs to set up connection in transport layer.

Its detailed process for setting up the HTTP connections of server and client is：

Step 11：Create a URL object；

Step 12：Create a HTTP link；

Step 13：Data are obtained using IO streams.

Step 2,

Then user end to server sends HTTP Get requests, and request is downloaded MPD file and stored.

Detailed process is：

Step 21：Judge that MPD file whether there is, 0 is otherwise returned if returning to 1 in the presence of if；

Step 22：File is downloaded when file is not present；

Step 23：Stored to downloading the file completed.

Step 3,

Server is received carries out response transmission MPD file to client after the request of client, client is entered to MPD file Row is downloaded and the MPD file of download is parsed；In order that video can be commenced play out as early as possible, according to parsing MPD file institute Range of code rates corresponding to the URL of obtained video slicing first downloads a part of video segment with minimum code rate, and server is received Video segment file is sent to client after request, client is downloaded to section file.The structure of MPD file such as Fig. 2 institutes Show.

Detailed process is：

Step 31：The specified attribute of Representation elements in MPD file is obtained, only needs to obtain id here, Tri- attributes of width, height；

Step 32：Obtain the URL of each grade video segment in MPD file；

Step 33：The video segment that minimum code rate is chosen in the URL addresses according to obtained by parsing MPD file is downloaded.

Step 4,

Network bandwidth when being cut into slices to step 3 foradownloaded video is calculated, and is counted and is recorded and uses Kalman filtering pre- Survey method is predicted to network bandwidth when downloading next video slicing.

Fig. 3 is the flow chart of the adaptively selected algorithm of code check, although after client is downloaded and parsed to MPD file The URL of video file has been obtained, but in order that user obtains more preferable audiovisual experience and can't be directly downloaded, it is necessary to pass through Cross the adaptively selected strategy of code check.

It is comprised the following steps that：

Step 41：Obtain the network bandwidth b of downloading slicing_n：

Wherein：t_n(e) represent that the time completed is downloaded in n-th of section；t_n(s) represent n-th section start download when Between；R (n) represents the selected code check of n-th of section；θ represents that duration is played in each section, here, and each section is played Duration is all identical, and its value of identical time span is all cut into for 2 seconds when carrying out video cutting encapsulation；b_nIt is n-th The network download bandwidth value of section.

Step 42：Set up system state equation and measurement equation：

Wherein：b_n+1It is the prediction download bandwidth value of (n+1)th section；Y_n+1It is n+1 section bandwidth measuring value；A is shape State transfer matrix, reflects system from the download bandwidth of n-th of section to the state change of (n+1)th；u_nIt is n-th of section Dominant vector；B is control allocation matrix；q_nAnd r_n+1It is system noise and measurement noise, q respectively_nStatistical property E [q_n]=0, r_n+1Statistical property E [r_n+1]=0；H is observing matrix.

Step 43：By the optimum bandwidth value of n-th of burstRemove the state value of (n+1)th burst of predictionDue to mould The random disturbances q of type_nDo not know in advance, soUnder usual condition due to bandwidth excursion not It is very big, it is possible to regarding the optimal download bandwidth of the prediction download bandwidth of next burst and a upper burst as , even A=1, dominant vector u_n=0, obtain：

Step 44：By last error covariance p_nWith system noise q_nCovariance Q predict new error covariance I.e.

Wherein Q is system noise q_nCovariance, its type belongs to white Gaussian noise；For (n+1)th slicer error association The predicted value of variance, i.e., new error covariance.

Step 45：Calculate kalman gainHere measured value is Calculated value is so H=1 is

Wherein R is measurement noise r_n+1Covariance；

Step 46：It is rightCorrection updates,

NowFor the optimum bandwidth of (n+1)th burst；Y_n+1For the bandwidth measuring value of (n+1)th section.

Step 47：Continue step 48 if completion is not yet all downloaded in section, otherwise exit whole prediction process.

Step 48：Estimate that the iteration of the optimum bandwidth value of the n-th+2 sections is updated operation for next step, that is, update p_n+1Value：

P_n+1It is the optimum bandwidth deviation of (n+1)th section；

Step 49, step 43 is jumped to, repeats to step 47 to terminate always.

Step 5：According to the prediction bandwidth obtained by step 4 and with reference to the remaining situation of current cache area video segment to downloading The code check of next video segment makes a choice, and detailed process is：

Step 51：If video slicing is not downloaded all and completed, i.e., there are still prediction bandwidth, then step 52 is performed, otherwise Downloading process terminates.

Step 52：

In order to ensure buffer area occurs without overflow and underflow, at the same ensure buffering area can effectively cancellation bandwidth fluctuation made Into forecasting inaccuracy really bring broadcasting stagnate or buffer overflow possibility, now to two time threshold values of setting buffers q_minAnd q_max, the code to be chosen is determined according to the threshold values of setting buffers and the burst number existed and prediction bandwidth Rate value.

Wherein, step 52 is specially：

Step 521：

Calculate n-th of burst s_nThe remaining duration of buffering area after the completion of download：

T₍t_n(e))=T (t_n(s))+θ-(t_n(e)-t_n(s)) (8)

Wherein T (t_n(e)) represent that n-th of burst has downloaded rear buffering area duration (while being also to start download (n+1)th to cut The duration of buffering area during piece), T (t_n(s)) when representing to start to download n-th of burst buffering area duration (while being also to have downloaded the The duration of buffering area during n-1 burst).

Step 522：

As T (t_n+1(s)) ＜ q_minWhen buffer area section have the possibility being finished, in order to ensure buffering area has enough regard Frequency section must assure that T (t_n+1(e))≥q_min, then its code check need meet：

ψ has no implication, and the formula equal sign left side is only calculated to obtained value and is designated as ψ.Code check r (n+1) selected by (n+1)th Value should be max (r (1), ψ), but be due to that code check value is series of discrete value, so ensureing the smooth premise of video playback Down final code check value be apart from max (r (1), ψ) it is nearest take fixed code check value downwards.Wherein r (1) generates for video segment When minimum code rate.

Step 523：

As T (t_n+1(s)) ＞ q_maxWhen buffering area have the possibility that overflow goes out be loaded with enough sections, continuation, in order to true Protecting buffering area, overflow not must assure that T (t_n+1(e))≤q_max, then its code check need meet：

ξ similarly has no concrete meaning, and the value for only obtaining the formula equal sign left side is designated as ξ；

Situation one：R (p) >=ξ is in order to ensure video pictures small shake as far as possible, and the code check value of selection should be ξ, simultaneously Because code check value is series of discrete value, thus code check value final on the premise of ensureing that video playback is smooth be apart from ξ most Near takes fixed code check value upwards.

Situation two：During r (p) ＜ ξ, buffer area overflow must be occurred by continuing download, in order to avoid such case occurs then Start sleep mechanism, that is, a period of time can be suspended by downloading, and n-th of burst s is recalculated after pause_nBuffered after the completion of download The remaining duration in area, proceeds code check judgement.Wherein r (p) is maximal rate when video segment is generated.

Step 524：

Work as q_min≤T(t_n+1(s))≤q_maxWhen, the burst number of buffering area is in an optimal scope, the value model of code check Enclose for：

In order to reduce code check fluctuation, code check value is identical with upper one section code check value holding：R (n+1)=r (n).

Step 6：Step 4-5 is jumped to if video segment is without completion is all downloaded, prediction is continued executing with and downloaded Journey, otherwise exits cyclic process, and method terminates.

The method that the present invention employs Kalman prediction in bandwidth prediction, because the principle of this method defers to pre- The iteration process of survey-amendment, so ensure that its bandwidth prediction precision；Hold secondly as mobile client has storage Not enough actual conditions are measured, such case most probably causes buffer area to occur " spilling " phenomenon, when lower overflow occurs for buffer area, Video playback can be stagnated, when being overflowed in generation, bandwidth resources can not be made full use of to improve video playing quality again. Buffer area control strategy, and the side by the way that bandwidth prediction is combined with buffer area control are added again in order to solve this problem Method predicts the code check of video segment to be downloaded, can so cause the selected slice code check can to meet the actual feelings of client-cache Condition, the problem of reduction is brought by buffer overflow, preferentially ensure that video smooth playing, and selected code check can be made to meet network again Bandwidth actual change situation, it is ensured that video quality, greatly enhances user's viewing experience.

The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the scope of the present invention.It is all Any modification, equivalent substitution and improvements made within the spirit and principles in the present invention etc., are all contained in protection scope of the present invention It is interior.

Claims (7)

1. the dynamic self-adapting code rate selection method based on MPEG-DASH agreements, it is characterised in that specifically enter according to following steps OK：

Step 1, the HTTP connections of server and client are set up；

Step 2, user end to server sends HTTP Get requests, and request is downloaded MPD file and stored；

Step 3, server is received carries out response transmission MPD file to client after the request of client, client is to MPD file It is downloaded and the MPD file of download is parsed；In order that video can be commenced play out as early as possible, according to parsing MPD file Range of code rates corresponding to the URL of resulting video slicing first downloads a part of video segment with minimum code rate, and server is received Video segment file is sent to client after to request, client is downloaded to section file；

Step 4, network bandwidth when being cut into slices to step 3 foradownloaded video is calculated, and is counted and is recorded and uses Kalman filtering Predicted method is predicted to network bandwidth when downloading next video slicing；

Step 5, it is according to the prediction bandwidth obtained by step 4 and next to downloading with reference to the remaining situation of current cache area video segment The code check of individual video segment makes a choice；

Step 6, step 4-5 is jumped to if video segment is without completion is all downloaded, continues executing with prediction and downloading process, it is no Cyclic process is then exited, method terminates.

2. the dynamic self-adapting code rate selection method according to claim 1 based on MPEG-DASH agreements, its feature exists In the detailed process of the step 1 is：

Step 11：Create a URL object；

Step 12：Create a HTTP link；

Step 13：Data are obtained using IO streams.

3. the dynamic self-adapting code rate selection method according to claim 1 based on MPEG-DASH agreements, its feature exists In the detailed process of the step 2 is：

Step 21：Judge that MPD file whether there is, 0 is otherwise returned if returning to 1 in the presence of if；

Step 22：File is downloaded when file is not present；

Step 23：Stored to downloading the file completed.

4. the dynamic self-adapting code rate selection method according to claim 1 based on MPEG-DASH agreements, its feature exists In the detailed process of the step 3 is：

Step 31：Obtain the specified attribute of Representation elements in MPD file, it is only necessary to obtain id, width, Tri- attributes of height；

Step 32：Obtain the URL of each grade video segment in MPD file；

Step 33：The video segment that minimum code rate is chosen in the URL addresses according to obtained by parsing MPD file is downloaded.

5. the dynamic self-adapting code rate selection method according to claim 1 based on MPEG-DASH agreements, its feature exists In the detailed process of the step 4 is：

Step 41：Obtain the network bandwidth b of downloading slicing_n：

$b_n=\frac{r\left(n\right)}{t_n\left(e\right)-t_n\left(s\right)}\mathrm{\θ}---\left(1\right)$

Wherein：t_n(e) represent that the time completed is downloaded in n-th of section；t_n(s) represent that n-th of section starts the time downloaded；r (n) the selected code check of n-th of section is represented；θ represents that duration is played in each section, and its value is 2 seconds；b_nIt is to cut for n-th The network download bandwidth value of piece；

Step 42：Set up system state equation and measurement equation：

$\left\{\begin{array}{c}{b}_{n+1}={\mathrm{Ab}}_n+{\mathrm{Bu}}_n+q_n\\ Y_{n+1}=H({\mathrm{Ab}}_n+{\mathrm{Bu}}_n+q_n)+r_{n+1}\end{array}\right.---\left(2\right)$

Wherein：b_n+1It is the prediction download bandwidth value of (n+1)th section；Y_n+1It is n+1 section bandwidth measuring value；A is that state turns Matrix is moved, system is reflected from the download bandwidth of n-th of section to the state change of (n+1)th；u_nIt is the control of n-th of section Vector；B is control allocation matrix；q_nAnd r_n+1It is system noise and measurement noise, q respectively_nStatistical property E [q_n]=0, r_n+1 Statistical property E [r_n+1]=0；H is observing matrix；

Step 43：By the optimum bandwidth value of n-th of burstRemove the state value of (n+1)th burst of predictionDue to model Random disturbances q_nDo not know in advance, soBy the prediction download bandwidth of next burst and upper one The optimal download bandwidth of individual burst be considered as it is identical, even A=1, dominant vector u_n=0, obtain：

$b_{n+1}={\hat{b}}_{n+1}^{-}={\hat{b}}_n---\left(3\right)$

Step 44：By last error covariance p_nNew error covariance is predicted with system noise qn covariance QI.e.

$p_{n+1}^{-}=p_n+Q---\left(4\right)$

Wherein Q is system noise q_nCovariance, type belongs to white Gaussian noise；For (n+1)th slicer error covariance Predicted value, i.e., new error covariance；

Step 45：Calculate kalman gainHere measured value is meter Calculation value is so H=1 is

$K_{n+1}=\frac{p_{n+1}^{-}}{p_{n+1}^{-}+R}---\left(5\right)$

Wherein R is measurement noise r_n+1Covariance；

Step 46：It is rightCorrection updates,

${\hat{b}}_{n+1}={\hat{b}}_{n+1}^{-}+K_{n+1}(Y_{n+1}-{\hat{b}}_{n+1}^{-})---\left(6\right)$

NowFor the optimum bandwidth of (n+1)th burst；Y_n+1For the bandwidth measuring value of (n+1)th section；

Step 47：Continue step 48 if completion is not yet all downloaded in section, otherwise exit whole prediction process；

Step 48：Estimate that the iteration of the optimum bandwidth value of the n-th+2 sections is updated operation for next step, that is, update p_n+1Value：

$p_{n+1}=(1-K_{n+1})p_{n+1}^{-}---\left(7\right)$

P_n+1It is the optimum bandwidth deviation of (n+1)th section；

Step 49, step 43 is jumped to, repeats to step 47 to terminate always.

6. the dynamic self-adapting code rate selection method according to claim 1 based on MPEG-DASH agreements, its feature exists In the detailed process of the step 5 is：

Step 51：If video slicing is not downloaded all and completed, i.e., there are still prediction bandwidth, then step 52 is performed, is otherwise downloaded Process terminates；

Step 52：

In order to ensure buffer area occurs without overflow and underflow, at the same ensure buffering area can effectively cancellation bandwidth fluctuation caused by Broadcasting stagnation or the possibility of buffer overflow that forecasting inaccuracy is really brought, now to two time threshold values q of setting buffers_minWith q_max, the code check value to be chosen is determined according to the threshold values of setting buffers and the burst number existed and prediction bandwidth.

7. the dynamic self-adapting code rate selection method according to claim 6 based on MPEG-DASH agreements, its feature exists In the detailed process of the step 52 is：

Step 521：

Calculate n-th of burst s_nThe remaining duration of buffering area after the completion of download：

T(t_n(e))=T (t_n(s))+θ-(t_n(e)-t_n(s)) (8)

Wherein T (t_n(e)) represent that n-th of burst has downloaded rear buffering area duration, while being also when starting to download (n+1)th section The duration of buffering area, T (t_n(s) duration of buffering area when) representing to start to download n-th of burst, while being also to have downloaded (n-1)th The duration of buffering area during individual burst；

Step 522：

As T (t_n+1(s)) ＜ q_minWhen buffer area section have the possibility being finished, in order to ensure buffering area has enough videos to cut Piece must assure that T (t_n+1(e))≥q_min, then its code check need meet：

$r(n+1)\≤b_{(n+1)}+\frac{b_{(n+1)}}{\mathrm{\θ}}\left(T\right(t_{n+1}\left(s\right))-q_{min})=\mathrm{\ψ}---\left(9\right)$

The formula equal sign left side is only calculated obtained value and is designated as ψ, code check r (n+1) value selected by (n+1)th is by ψ without implication Max (r (1), ψ), but be due to that code check value is series of discrete value, so final on the premise of ensureing that video playback is smooth Code check value be apart from max (r (1), ψ) it is nearest take fixed code check value downwards, wherein r (1) is minimum when video segment is generated Code check；

Step 523：

As T (t_n+1(s)) ＞ q_maxWhen buffering area have the possibility that overflow goes out be loaded with enough sections, continuation, in order to ensure Overflow does not ensure T (t to buffering area_n+1(e))≤q_max, then its code check need meet：

$r(n+1)\≥b_{(n+1)}+\frac{b_{(n+1)}}{\mathrm{\θ}}\left(T\right(t_{n+1}\left(s\right))-q_{max})=\mathrm{\ξ}---\left(10\right)$

ξ is without implication, and the value for only obtaining the formula equal sign left side is designated as ξ；

Situation one：R (p) >=ξ is in order to ensure video pictures small shake as far as possible, and the code check value of selection is ξ, simultaneously because code check Value is series of discrete value, so code check value final on the premise of ensureing that video playback is smooth is apart from nearest upward of ξ Take fixed code check value；

Situation two：During r (p) ＜ ξ, buffer area overflow must be occurred by continuing download, in order to avoid such case occurs then to start Sleep mechanism, that is, a period of time can be suspended by downloading, and n-th of burst s is recalculated after pause_nBuffering area is remained after the completion of download Remaining duration, proceeds code check judgement, and wherein r (p) is maximal rate when video segment is generated；

Step 524：

Work as q_min≤T(t_n+1(s))≤q_maxWhen, the burst number of buffering area is in an optimal scope, the span of code check For：

$b_{(n+1)}+\frac{b_{(n+1)}}{\mathrm{\&theta;}}\left(T\right(t_{n+1}\left(s\right))-q_{max})<r(n+1)<b_{(n+1)}+\frac{b_{(n+1)}}{\mathrm{\&theta;}}\left(T\right(t_{n+1}\left(s\right))-q_{\min })---\left(11\right)$

In order to reduce code check fluctuation, code check value is identical with upper one section code check value holding：R (n+1)=r (n).