CN100428802C - Self adaptive multimedia playing method - Google Patents

Abstract

The invention relates to a method for adaptive multimedia broadcasting. Its characteristics show as follow: it records the arriving time about the received multimedia data, then it pre-detects the preliminary rate for broadcasting according to described time above, then it adapts the predictive value by using the current buffer occupancy rate, and it sets a limit about the predictive value, then it broadcasts the multimedia according the restricted rate. This method effectively reduces the possibility of buffer overflow and may also real-time tracks the changes in network delay to change the rate of broadcasting.

Description

The method that a kind of adaptive multimedia is play

Technical field

The present invention relates to the multimedia adaptive player method, belong to the multimedia communication technology field.

Background technology

Multimedia communication is meant the novel communication mode that multiple media information----sound, image, figure, data, text etc. can be provided simultaneously in the call process.It is the product that the communication technology and computer technology combine.The single medium communication mode traditional with phone, telegram, fax, compunication etc. compares, utilize multimedia communication, be separated by ten thousand li user can not only acoustic image both pictures and texts are excellent ground exchange of information, be distributed in the multimedia messages of different location, can also act in agreement be presented in face of the user as a complete information, and the user has complete mutual control ability to the communication overall process." best effort " that real network provided do not guarantee the delay variation of end-to-end delay, grouping of single grouping and packet loss, needs one section buffering area to come the shake that smoothly brings thus at the receiving terminal of medium.The delay variation of network is the principal element that influences the streaming media playing quality.But less buffering area may exhaust and take place buffering very soon and overflow in relatively poor network environment; Bigger buffering area can increase broadcast time-delay end to end again and be not suitable for real time business.

Adopt fixing speed to play, the time-delay that can not adapt to network changes, and causes overflowing of buffering area simultaneously.Therefore need be according to the adaptive change playback rate of buffer performance.As shown in Figure 1, this method may further comprise the steps:

(1) preestablishes normal playback rate;

(2) change playback rate according to current buffer performance;

Above-mentioned prior art changes playback rate according to current buffer performance, but do not fully take into account the time-varying characteristics of network, can not adjust dynamically along with the variation of network delay, under the more violent situation of network change, can not well eliminate the influence that delay variation brings the streaming media playing quality.For can better tracking network the variation of time-delay, need further consider the time-delay of network.

Summary of the invention

The objective of the invention is for overcoming the weak point of prior art, the method that a kind of adaptive multimedia that proposes is play, can utilize the statistical property of network delay shake, reduce effectively buffering area overflow possible in the variation of tracking network time-delay in real time change broadcasting speed, improve the subjective vision observation effect.

The invention is characterized in that this method realizes according to the following steps successively in PC:

Step (1) is to writing down and deposit in the data buffer zone time of advent of the multi-medium data received;

Step (2), according to the multimedia time that receives as follows to multimedia playback rate

Carry out tentative prediction:

$\frac{1}{E\left[J\left(i\right)\right]+\mathrm{k\&sigma;}\left[J\left(i\right)\right]}<\hat{F}<\frac{n}{\mathrm{nE}\left[J\left(i\right)\right]-\mathrm{k\&sigma;}\left[J\left(i\right)\right]},$

Wherein, J (i)=t (i+1)-t (i) i=1,2 ..., 9, t (i) is the time of reception, J (i) is that the time of reception of two groups of multi-medium datas is poor,

$E\left[J\left(i\right)\right]=\frac{1}{M}\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}J\left(i\right),$ E[J (i)] be the average of J (i), M is the number of the J (i) that calculates in the step (2), $\mathrm{\&sigma;}\left[J\left(i\right)\right]=\sqrt{\frac{1}{M}\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{(J\left(i\right)-E[J\left(i\right)\left]\right)}^2},$ σ [J (i)] is the variance of J (i), k=3, n is a positive integer, makes nE[J (i)]-k σ [J (i)]＞0;

Step (3), the playback rate that as follows step (2) is obtained with the current occupancy in data buffer zone

Revise:

${\hat{F}}^{\&prime;}=\left\{\begin{array}{cc}\frac{O_B-O_s}{1-O_s}\&times;(\frac{n}{\mathrm{nE}\left[J\left(i\right)\right]-\mathrm{k\&sigma;}\left[J\left(i\right)\right]}-\frac{1}{E\left[J\left(i\right)\right]})+\frac{1}{E\left[J\left(i\right)\right]}& \mathrm{if}{O}_B\&GreaterEqual;O_s,\\ \frac{O_B-O_s}{O_s}\&times;(\frac{1}{E\left[J\left(i\right)\right]}-\frac{1}{E\left[J\left(i\right)\right]+\mathrm{k\&sigma;}\left[J\left(i\right)\right]})+\frac{1}{E\left[J\left(i\right)\right]}& \mathrm{if}{O}_B<O_s\end{array}\right.$ Wherein, O _BBe the current occupancy in data buffer zone, by the total B of the data buffer zone number of receiving terminal _MaxNumber B with the current data buffer zone that has taken _cPass through formula $O_B=\frac{B_c}{B_{\max }}$ Obtain O _sBe the occupancy of data buffer zone expectation, be based on reduce the data buffer zone overflow may and propose, establish O _s=0.5;

Step (4), the playback rate that step (3) is obtained by following formula

Limit:

Final playback rate F _FinalFor:

Wherein, F _LaslIt is the last playback rate that uses.

Characteristics of the present invention are to utilize the statistical property of Chebyshev inequality and delay variation, and the playback rate that prediction is suitable utilizes the current occupancy of buffering area to regulate playback rate jointly simultaneously.According to the visual characteristic of human eye playback rate is limited again afterwards.

This method can reduce the possibility that buffering area overflows effectively, and the variation of tracking network time-delay in real time changes broadcasting speed again simultaneously, has reduced the infringement to play quality that causes because of the delay variation that happens suddenly.Because unqualified concrete media business, by different parameter settings, this algorithm is applicable to multiple media application such as voice, videos.

Simulation result: under personal computer platform WindowsXP operating system, realize with the C/C++ programming, translation and compiling environment is Micosoft Visual C++6.0, and relevant parameter is referring to embodiment, and medium are selected video for use, transmitting terminal frame per second F=5 frame/second, network delay meets the even distribution at [50,100] millisecond, and has 5% probability generation to meet [150 than the network delay of length, 300] Hao Miao even distribution, computer artificial result is seen Fig. 3.Can see, method of the present invention guarantee buffering area can keep stable in, guaranteed that also broadcasting speed can be with the buffering dynamic change, the mean change amplitude of speed has reduced, the variation of adjacent speed is also compared even.

Description of drawings

Fig. 1 is the adaptive media player method FB(flow block) of prior art.

The adaptive media player method FB(flow block) that Fig. 2 proposes for the present invention.

Fig. 3 is computer artificial result figure:

(a) buffering effect, wherein solid line is the method for the present invention's proposition, dotted line is the method for prior art,

(b) playback rate effect, wherein * is the method for the present invention's proposition, o is the method for prior art

Embodiment

The method that the adaptive media that the present invention proposes is play reaches embodiment in conjunction with the accompanying drawings in personal computer platform and further specifies as follows:

Method flow of the present invention may further comprise the steps as shown in Figure 2:

(1) time of the multi-medium data that receives is carried out record;

(2) according to the time that receives media play speed is carried out tentative prediction;

(3) with the current occupancy of buffering area the predicted value of playback rate is revised;

(4) (3) being obtained playback rate limits;

(5) medium of playing with the speed after the restriction.

The specific embodiment of each step of said method of the present invention is described in detail as follows respectively:

Said method step (1) is carried out record to system's time of reception of 10 groups of multi-medium datas receiving.Be recorded as t (1)～t (10).Can obtain time of reception difference J (i)=t (i+1)-t (i) i=1 of two multi-medium datas, 2 ..., 9.Deposit data in buffer storage.

The embodiment of said method step (2) is: average and the variance method of calculating J (i) are as follows: $E\left[J\left(i\right)\right]=\frac{1}{M}\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}J\left(i\right),$ $\mathrm{\&sigma;}\left[J\left(i\right)\right]=\sqrt{\frac{1}{M}\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{(J\left(i\right)-E[J\left(i\right)\left]\right)}^2}.$ Wherein M is the number 9 of the J (i) that calculates in the above-mentioned steps (1).Calculate the initial predicted value of media play speed $\hat{F}=\frac{1}{E\left[J\left(i\right)\right]}$ With Scope $\frac{1}{E\left[J\left(i\right)\right]+\mathrm{k\&sigma;}\left[J\left(i\right)\right]}<\hat{F}<\frac{n}{\mathrm{nE}\left[J\left(i\right)\right]-\mathrm{k\&sigma;}\left[J\left(i\right)\right]}.$ Wherein parameter k is desirable 3, makes nE[J (i) by choosing positive integer n]-k σ [J (i)]＞0.

The embodiment of said method step (3) is: use formula ${\hat{F}}^{\&prime;}=\left\{\begin{array}{cc}\frac{O_B-O_s}{1-O_s}\&times;(\frac{n}{\mathrm{nE}\left[J\left(i\right)\right]-\mathrm{k\&sigma;}\left[J\left(i\right)\right]}-\frac{1}{E\left[J\left(i\right)\right]})+\frac{1}{E\left[J\left(i\right)\right]}& \mathrm{if}{O}_B\&GreaterEqual;O_s,\\ \frac{O_B-O_s}{O_s}\&times;(\frac{1}{E\left[J\left(i\right)\right]}-\frac{1}{E\left[J\left(i\right)\right]+\mathrm{k\&sigma;}\left[J\left(i\right)\right]})+\frac{1}{E\left[J\left(i\right)\right]}& \mathrm{if}{O}_B<O_s\end{array}\right.$ Predicted value to playback rate Revise.O wherein _BExpression buffering area current occupancy is by the total B of the buffering area number of receiving terminal _MaxWith the current buffering area number B that has taken _cPass through formula $O_B=\frac{B_c}{B_{\max }}.$ O _sOccupancy for the buffering area expectation in order to reduce the possibility of buffer overflow and underflow simultaneously, generally can be made as 0.5.

The embodiment of said method step (4) is: (3) are obtained playback rate, use formula

The playback rate that obtains is limited.F wherein _LaslBe last playback rate, F _FinalIt is this playback rate that finally obtains.

Said method step (5) the final playback rate F that obtains _FinalCarry out media play.

Claims (1)

1, a kind of method of adaptive multimedia broadcast is characterized in that, this method realizes in PC according to the following steps successively:

Step (1) is to writing down and deposit in the data buffer zone time of advent of the multi-medium data received;

Step (2), according to the multimedia time that receives as follows to multimedia playback rate

Carry out tentative prediction:

$\frac{1}{E\left[J\left(i\right)\right]+\mathrm{k\&sigma;}\left[J\left(i\right)\right]}<\hat{F}<\frac{n}{\mathrm{nE}\left[J\left(i\right)\right]-\mathrm{k\&sigma;}\left[J\left(i\right)\right]},$

Wherein, J (i)=t (i+1)-t (i) i=1,2 ..., 9, t (i) is the time of reception, J (i) is that the time of reception of two groups of multi-medium datas is poor,

$E\left[J\left(i\right)\right]=\frac{1}{M}\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}J\left(i\right),$ E[J (i)] be the average of J (i), M is the number of the J (i) that calculates in the step (2),

$\mathrm{\&sigma;}\left[J\left(i\right)\right]=\sqrt{\frac{1}{M}\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{(J\left(i\right)-E[J\left(i\right)\left]\right)}^2},$ σ [J (i)] is the variance of J (i), k=3,

N is a positive integer, makes nE[J (i)]-k σ [J (i)]＞0;

Step (3), the playback rate that as follows step (2) is obtained with the current occupancy in data buffer zone

Revise:

${\hat{F}}^{\&prime;}=\left\{\begin{array}{cc}\frac{O_B-O_s}{1-O_s}\&times;(\frac{n}{\mathrm{nE}\left[J\left(i\right)\right]-\mathrm{k\&sigma;}\left[J\left(i\right)\right]}-\frac{1}{E\left[J\left(i\right)\right]})+\frac{1}{E\left[J\left(i\right)\right]}& \mathrm{if}{O}_B\&GreaterEqual;O_s\\ \frac{O_B-O_s}{O_s}\&times;(\frac{1}{E\left[J\left(i\right)\right]}-\frac{1}{E\left[J\left(i\right)\right]+\mathrm{k\&sigma;}\left[J\left(i\right)\right]})+\frac{1}{E\left[J\left(i\right)\right]}& \mathrm{if}{O}_B<O_s\end{array}\right.,,$

Wherein, O _BBe the current occupancy in data buffer zone, by the total B of the data buffer zone number of receiving terminal _MaxNumber B with the current data buffer zone that has taken _cPass through formula $O_B=\frac{B_c}{B_{\max }}$ Obtain,

O _sBe the occupancy of data buffer zone expectation, be based on reduce the data buffer zone overflow may and propose, establish O _s=0.5;

Step (4), the playback rate that step (3) is obtained by following formula

Limit:

Final playback rate F _FinalFor:

Wherein, F _LastIt is the last playback rate that uses.

Images