TWI610559B - Method and device for optimizing video transcoding - Google Patents

Description

Method and device for optimizing video transcoding

The invention relates to a method and a device for optimizing video transcoding, in particular to a method for simultaneously utilizing a transcoding format and a transcoder, estimating a computing resource required for transcoding, a transcoder load information, and further assigning a transcoding. The task is to the most suitable transcoder to achieve a method and apparatus for optimizing transcoding performance.

Video transcoding technology continues to improve with the continuous development of video compression standards. When the number of converted videos is small, a single transcoder can generally be used to perform transcoding tasks. When the number of video files is large, multiple conversions are used. The code is parallel, or is extended by using a cloud device. In either case, when the transcoding server executes multiple transcoding tasks at the same time, the transcoding task is too complicated and too complicated, which may cause server overload, performance degradation, and server. Unable to respond to other issues, resulting in uncontrollable transcoding time and communication and operation with other devices (such as database, operation interface, NAS network storage device, etc.). Conversely, if too many transcoding servers are used, the server will be idle and waste resources.

In the prior art, a method for assigning a video transcoding work to different transcoding nodes is characterized in that a node manager is introduced, and the node manager determines that the format of the transcoded movie, such as the size of the movie, is redistributed to the transcoding node, however, This method only considers the characteristics of the film itself, but does not consider the ability of the transcoding node itself to transcode and the load at that time, so it cannot be positive. The correct allocation of transcoding work can not achieve the maximum efficiency of the transcoding node.

In the prior art, a method for balancing transcoding rates of different transcoding jobs in the same transcoding server is characterized by observing the rate at which the transcoding work outputs a movie, when the first transcoding work output is higher than the second transcoding. When the code works, the first transcoding work rate is lowered, and the second transcoding work rate is increased. However, this method only balances the transcoding work rate, does not improve the efficiency of transcoding work, and does not consider the characteristics of the film itself to effectively utilize the transcoding server performance.

In view of the shortcomings derived from the above-mentioned conventional methods, the inventor of the present invention has improved and innovated, and after years of painstaking research, he finally succeeded in researching and developing a method for optimizing video transcoding.

The invention provides a method and a device for optimizing video transcoding, which can solve the overload of the transcoding server, the performance degradation, the server can not respond, the transcoding time cannot be controlled, and other devices (such as the database, the operation interface, the NAS network) Road storage devices, etc.) can not properly communicate and operate problems, but also can solve the problem of using too many transcoding servers to cause the server to idle, waste resource costs and so on.

The invention provides a method and a device for optimizing video transcoding. At the same time, using the transcoding format and the characteristics of the transcoder, estimating the computing resources, transcoder load information required for adding the transcoding task, and effectively allocating the transcoding task to The most suitable transcoder allows the most efficient use of computing resources.

The method for optimizing video transcoding of the present invention comprises the steps of: a. The code function function, wherein the function can use the decoding information, the coding information, the filtering information, and predict and calculate the transcoder load information, and the method of determining the function is to decode the information and code in the known transcoding task. The information and filtering information are regarded as inputs, and the corresponding transcoder load information is regarded as an output, and the transcoder performance function coefficient is calculated; step b. after the new transcoding task is added by the transcoder performance function estimation The transcoder load information, and judges whether it is a valid state by using the transcoder load information; and step c. normalizes the transcoder load information that is determined to be in an effective state, according to the normalized transcoder load Information distribution transcoding task. Through this method, the effect of optimizing the allocation transcoding task can be achieved.

The invention provides a method and a device for optimizing video transcoding, which solves the problem that the transcoding server uses different performances due to different encoding standards, film size, quality and other formats, and the transcoding server cannot be estimated in advance. The problem with the computing resources required for the task. The invention first determines the transcoder performance function through the known transcoding state, and then substitutes the new transcoding task into the transcoder performance function to achieve the effect of estimating the transcoder load in advance, thereby optimizing the video transcoding efficiency. .

The transcoder performance function established by the invention provides an automatic evaluation method for the performance of the transcoding server, which can be used for evaluating the expansion of the device, solving the redundant expansion cost, idle operation of the server, and wasting resource costs.

The invention provides a method for estimating the shift time, which allows the user to control and manage the shift time and solve the problem of delay of the shift time caused by the transcoder overload.

Step b. Using the transcoder performance function to estimate the transcoder load information after adding a new transcoding task, and determining whether the transcoder load information is valid, directly determining to join the new transcoding task Whether the post transcoder can be overloaded and solve the transcoding server The load, performance is degraded, the server cannot respond, the transcoding time cannot be controlled, and other devices cannot communicate and operate normally.

Step c. normalizes the transcoder load information that is determined to be in an active state, and distributes the transcoding task according to the normalized transcoder load information, and provides a use of different transcoding capabilities and different transcoding conditions. The method of directly comparing the transcoder load between multiple transcoders in the state solves the problem that the load cannot be evaluated when a plurality of transcoders are used at the same time.

The present invention provides a method and apparatus for optimizing video transcoding. Compared with the prior art, the present invention can simultaneously use the transcoding format and the characteristics of the transcoder, and estimate the computing resources required for the transcoding task. The code load information effectively allocates the transcoding task to the most suitable transcoder, so that the computing resources can be used most effectively.

The detailed description of the preferred embodiments of the present invention is intended to be limited to the scope of the invention, and is not intended to limit the scope of the invention. The patent scope of this case.

In summary, this case is not only innovative in terms of space type, but also can enhance the above-mentioned multiple functions compared with the customary items. It should fully meet the statutory invention patent requirements of novelty and progressiveness, and apply for it according to law. This invention patent application, in order to invent invention, to the sense of virtue.

500‧‧‧ front-end operation page module

501‧‧‧ Transcoder efficiency function decision module

502‧‧‧ Transcoder Load Information Judgment Module

503‧‧‧Transcoding Task Assignment Module

504‧‧‧ Transcoding server

S100~S102‧‧‧Step process

S1001~S1005‧‧‧Step procedure

S1011~S1015‧‧‧Step process

S1021~S1022‧‧‧Step process

FIG. 1 is a schematic flow chart of a method for optimizing video transcoding according to the present invention.

2 is a schematic flow chart of a method for optimizing video transcoding according to the present invention.

3 is a schematic flow chart of a method for optimizing video transcoding according to the present invention.

4 is a schematic flow chart of a method for optimizing video transcoding according to the present invention.

FIG. 5 is a schematic diagram of an implementation of an apparatus for optimizing video transcoding according to the present invention.

The technical features, contents, and advantages of the present invention, as well as the advantages thereof, can be understood by the reviewing committee, and the present invention will be described in detail with reference to the accompanying drawings. The subject matter is only for the purpose of illustration and description. It is not intended to be a true proportion and precise configuration after the implementation of the present invention. Therefore, the scope and configuration relationship of the attached drawings should not be interpreted or limited. First described.

Please refer to FIG. 1 , wherein the method for optimizing video transcoding in the present invention, wherein step a. determines a transcoder performance function, corresponding to step S100 in the figure, decoding information, encoding information, filtering information in a known transcoding task. As input, the corresponding transcoder load information is regarded as output, and the transcoder performance function coefficient is calculated. The function of the transcoder performance function is to decode information, coding information, filtering information and transcoding. Correlate and convert between load information. When decoding information, encoding information, and filtering information are known, the information can be calculated by translating the information into the transcoder function. In order to estimate and calculate the transcoder load information, the transcoder efficiency function must be determined first. The method of determining the transcoder performance function is equivalent to finding the transcoder performance function coefficient. The calculation method is as shown in step a. Said. Step b. using the transcoder performance function It is estimated that the transcoder load information is added after the new transcoding task, and the transcoder load information is used to judge whether it is valid. Corresponding to S101 in Fig. 1, after determining the transcoder performance function coefficient, a new transfer is performed. The code task information is substituted into the transcoder performance function, and the transcoder load information can be calculated, and the transcoder load information is determined to be a valid state or an overload state. Step c. normalize the transcoder payload information that is determined to be in an active state, and assign a transcoding task according to the normalized transcoder load information, corresponding to S102 in the figure, to achieve the effect of optimizing video transcoding.

The transcoder load information in step a includes an arithmetic unit (CPU) load (cpuLd), an input/output (IO) wait rate (ioWa), a hardware interrupt request (IRQ) rate (hwInt), and a software interrupt request (IRQ) rate. (swInt), cache memory remaining amount (mem), decoding information (dec) includes source movie resolution (srcRes), source video bit rate (srcBitr), source video encoding format (srcCod), encoding information (enc) contains Target video resolution (tarRes), destination video bit rate (tarBitr), destination video encoding format (tarCod), filtering information (filt) including source movie resolution (srcRes), destination movie resolution (tarRes), filtering format (filtType) ).

A step of said performance function transcoder trans (), comprising a transcoder function performance _{coefficients: d ir, d ib, d} io, e ir, e ib, e io, f is, f it, f io, The information can be obtained between the decoded information, the encoded information, the filtered information and the transcoder load information to obtain the required information. The multivariate linear equation can be expressed and calculated as follows:

Where n ₀ =-1, k = n _{i -1} +1 to k = n _i use the i-th source video encoding format, the destination video encoding format or the filtering format, md is the total number of source video encoding formats, me is the destination video The total number of encoding formats, mf is the total number of filtering formats, d _ir , d _ib , d _io , e _ir , e _ib , e _io , f _is , f _it , f _io are the coefficients of the transcoder performance function, where the The coefficient of the coder function is a real number.

The transcoder load information described in step a, when the input/output waiting rate, the hardware interrupt request rate, and the software interrupt request rate in the transcoder load information are less than the corresponding threshold value, the cache memory remaining amount is greater than the phase When the corresponding threshold value is used, the transcoder load information is regarded as the operator load ( cpuLd ), and the operator load ( cpuLd ) is expressed and calculated by a multivariate linear equation as follows:

x=[d _1r d _1b d _1o …e _1r e _1b e _1o …f _1s f _1t f _1o …fn _mf s fn _mf t fn _mf o] ^T b=[cpuLd ₁…cpuLd _j …cpuLd _J ] ^T ，其中A係為解碼資訊、編碼資訊、濾波資訊，矩陣中的元素為來源影片解析度(srcRes)、來源影片碼率(srcBitr)、來源影片編碼格 式(srcCod)、目的影片解析度(tarRes)、目的影片碼率(tarBitr)、目的影片編碼格式(tarCod)、濾波格式(filtType)，b為運算器負載(cpuLd)，J為已知的轉碼任務的個數。 The transcoder efficiency function coefficient (x) described in step a can be solved by matrix operation or by numerical analysis method (such as Gauss-Jordan), wherein the matrix pattern is expressed as follows: Ax = b

x = [ d _{1 r} d _{1 b} d _{1 o} ... e _{1 r} e _{1 b} e _{1 o} ... f _{1 s} f _{1 t} f _{1 o} ... fn _mf s fn _mf t fn _mf o ] ^T b = [ cpuLd ₁ ... cpuLd _j ... cpuLd _J ] ^T , where A is the decoding information, coding information, filtering information, the elements in the matrix are the source film resolution (srcRes), the source film rate (srcBitr), the source film encoding format (srcCod), Destination video resolution (tarRes), destination video bit rate (tarBitr), destination video encoding format (tarCod), filtering format (filtType), b is the operator load (cpuLd), and J is the number of known transcoding tasks. .

Referring to FIG. 2, step (a) of the method for optimizing video transcoding in the present invention a. determining the transcoder performance function further includes step a1, corresponding to step S1001 in the figure, recording the transcoder load information when performing transcoding. And a plurality of decoding information, a plurality of coding information, and a plurality of filtering information in the transcoder, wherein any one of the input and output waiting rate, the hardware interrupt request rate, and the software interrupt request rate is greater than a corresponding threshold value, or a cache When the remaining amount of the memory is less than the corresponding threshold value, it is recorded as an overload state, otherwise it is an active state, step a2, corresponding to the step S1002 in the figure, comparing the same source video encoding format, the target video encoding format, the filtering format but different movie resolutions. Or the code rate of the transcoder load information, the minimum operator load in the overload state is set to the source video encoding format, the destination video encoding format, the filter format of the transcoder operator load upper limit, step a3, corresponding map Step S1003, accumulating the same source video encoding format, the destination video encoding format, the filtering format but different movie resolution or video bit rate The number of valid states of the encoder, step a4, corresponding to step S1004 in the figure, comparing the cumulative number of valid states of the transcoder with the number of coefficients of the transcoder function, when the cumulative number of valid states of the transcoder is equal to the transcoding When the number of coefficients of the function function is calculated, the transcoder performance coefficient S1005 is calculated, otherwise, the process returns to step S1001.

Referring to FIG. 3, step b of the method for optimizing video transcoding of the present invention b. predicting the transcoder load information after adding a new transcoding task by using the transcoder performance function, and loading the transcoder with the transcoder Whether the information is judged to be in a valid state, and further comprising: step b1, corresponding to step S1011 in the figure, substituting decoding information, coding information, and filtering information of the new transcoding task into the transcoder The function function, solve the transcoder load information, treat the transcoder load information as the operator load, and add the transcoder operator load at that time to obtain the estimated operator load, step b2, corresponding steps in the figure S1012, comparing the estimated operator load with a transcoder operator load upper limit having the same source video encoding format, a destination video encoding format, and a filtering format, and determining that the estimated computing device load is less than the upper limit S1013 S1014, on the other hand, is determined to be the overload state S1015.

The calculation method of estimating the operator load in step S1011 may also be performed by substituting the decoding information, coding information, and filtering information of the new transcoding task and the original transcoding task of the transcoder into the transcoder performance function to solve the transcoding. Load information and treat the transcoder load information as the estimated operator load.

Referring to FIG. 4, step c of the method for optimizing video transcoding according to the present invention c normalizes the transcoder load information determined to be in an active state, and distributes the transcoding task according to the normalized transcoder load information. And further comprising: step c1, corresponding to step S1021 in the figure, dividing the estimated operator load (cpuLd _effective ) determined to be in the valid state by the corresponding upper limit of the transcoder operator load (cpuLd _max ), and obtaining the regular The estimated operator load (cpuLd _normalized ) is calculated as follows: cpuLd _normalized = cpuLd _effective / cpuLd _max step c2, corresponding to step S1022 in the figure, comparing the normalized predictor load of all transcoders, will be transferred The code task is assigned to the transcoder with the lowest load on the normalized predictor.

FIG. 5 is a schematic diagram of an apparatus for optimizing video transcoding according to the present invention. The front-end operation page module 500 provides user setting and monitoring conversion tasks, input source files and retrieval destination files; and the transcoder performance function is determined. The module 501 performs the content of step a, determines a transcoder performance function, and inputs decoding information, coding information, and filtering in a known transcoding task. Information, corresponding transcoder load information, calculate and output the coefficient of the transcoder performance function, wherein the transcoder performance function can use the decoding information, the encoding information, the filtering information to estimate and calculate the transcoder load The transcoder load information judging module 502 executes the content of step b, inputs new transcoding task information, and adds one transcoder load after the new transcoding task is estimated by the transcoder performance function. The information is further determined by the transcoder load information, and the transcoder load information is output; the transcoding task assignment module 503 executes the content of step c, inputs the transcoder load information, and normalizes Normally determining the transcoder load information that is determined to be in an active state, and assigning a transcoding task according to the transcoder payload information; and transcoding the server 504, accepting the transcoding task to perform transcoding, and achieving optimized video conversion The efficacy of the code.

To sum up, this case is not only innovative in terms of technical thinking, but also has many of the above-mentioned functions that are not in the traditional methods of the past. It has fully complied with the statutory invention patent requirements of novelty and progressiveness, and applied for it according to law. Approved this invention patent application, in order to invent invention, to the sense of virtue.

S100~S102‧‧‧Step process

Claims (10)

A method for optimizing video transcoding, the steps of which are as follows: a. Determine a transcoder performance function, and regard the decoding information, coding information, and filtering information in the known transcoding task as input, corresponding to one turn The code load information is regarded as an output, and the coefficient of the transcoder performance function is calculated, wherein the transcoder performance function can estimate and calculate the transcoder load information by using decoding information, coding information, and filtering information; b. using the transcoder performance function to estimate the transcoder load information after adding a new transcoding task to obtain a predictor load, and using the transcoder load information and the predictor The load is compared to determine whether the transcoder load information is in an active state; and c. normalizes the transcoder load information determined to be in a valid state, and assigns a transcoding task according to the transcoder load information. The method for optimizing video transcoding as described in claim 1, wherein the transcoder load information includes: an operator (CPU) load (cpuLd), an input/output (IO) wait rate (ioWa), and a hard Volume interrupt request (IRQ) rate (hwInt), software interrupt request (IRQ) rate (swInt), cache memory remaining amount (mem), decoding information (dec) includes: source movie resolution (srcRes), source video code Rate (srcBitr), source video encoding format (srcCod), encoding information (enc) includes: destination video resolution (tarRes), destination video bit rate (tarBitr), destination video encoding format (tarCod), filtering information (filt) contains : source movie resolution (srcRes), destination movie resolution (tarRes), and filter format (filtType). The method for optimizing video transcoding as described in claim 1, wherein the transcoder performance function is a multi-variable linear equation, and the representation and calculation are as follows: Where trans ( cpuLd, ioWa, hwInt, swInt, mem ) is the transcoder performance function, n ₀ =-1, k = n _{i -1} +1 to k = n _i using the i-th source video encoding format , the destination video encoding format or filtering format, md is the total number of source video encoding formats, me is the total number of the target video encoding formats, mf is the total number of filtering formats, d _ir , d _ib , d _io , e _ir , e _ib , e _io , f _Is , f _it , f _io are the coefficients of the transcoder performance function, wherein the coefficients of the transcoder performance function are real numbers. The method for optimizing video transcoding according to claim 1, wherein the input/output waiting rate, the hardware interrupt request rate, and the software interrupt request rate in the transcoder load information are smaller than a corresponding threshold value. When the remaining amount of the cache memory is greater than the corresponding threshold value, the transcoder load information is regarded as the operator load (cpuLd), and the operator load (cpuLd) is represented by a multivariate linear equation, and the calculation and calculation are as follows: d _ir , d _ib , d _io , e _ir , e _ib , e _io , f _is , f _it , f _io are the coefficients of the transcoder performance function, wherein the coefficients of the transcoder performance function are real numbers . The method for optimizing video transcoding as described in claim 1, wherein the coefficient of the transcoder performance function is solved by matrix operation or by numerical analysis, and the matrix pattern is expressed as follows: Ax = b b =[ cpuLd ₁ ... cpuLd _j ... cpuLd _J ] ^T , where x is the coefficient of the transcoder performance function, A is the decoding information, coding information, filtering information, and the elements in the matrix are the source film resolution (srcRes) , source video bit rate (srcBitr), source video encoding format (srcCod), destination movie resolution (tarRes), destination video bit rate (tarBitr), destination video encoding format (tarCod), filtering format (filtType), b is the operation Load (cpuLd), J is the number of known transcoding tasks. The method for optimizing video transcoding according to claim 1, wherein the step a further comprises: a1. recording the transcoder load information, decoding information, coding information, and filtering information when performing transcoding, wherein If the input/output waiting rate, the hardware interrupt request rate, and the software interrupt request rate are greater than the corresponding threshold value, or the cache memory remaining amount is less than the corresponding threshold value, the recording is an overload state, otherwise it is an active state; A2. Comparing the transcoder load information with the same source video encoding format, the destination video encoding format, the filtering format but different movie resolution or video bit rate, the minimum operator load in the overload state is set to the source video encoding format. , the target video encoding format, the filter format of the transcoder operator load upper limit; A3. accumulating the number of transcoder valid states with the same source video encoding format, destination video encoding format, filtering format but different movie resolution or video bit rate; and a4. when the accumulated transcoder valid state times is equal to the turn When the number of coefficients of the code function is calculated, the transcoder efficiency coefficient is calculated, and vice versa, returning to step a1. The method for optimizing video transcoding according to claim 1, wherein the step b further comprises: b1. Substituting decoding information, coding information, and filtering information of the new transcoding task into the transcoder performance function. Solving the transcoder load information, treating the transcoder load information as an operator load and adding the transcoder operator load at that time to obtain the estimated operator load; and b2. comparing the estimated operator load with The upper limit of the transcoder operator load with the same source video encoding format, the destination video encoding format, and the filtering format. If the estimated computing device load is less than the upper limit, it is determined to be a valid state, and vice versa, an overload state. For example, the method for optimizing the video transcoding described in claim 7 of the patent scope, the calculation method of the estimated operator load, and the decoding information of the original transcoding task of the new transcoding task and the transcoder, Encoding information, filtering information, and substituting the transcoder function to solve the transcoder load information, and treating the transcoder load information as the estimated operator load. The method for optimizing video transcoding according to claim 1, wherein the step c further comprises: c1. dividing the estimated operator load (cpuLd _effective ) determined to be in a valid state by the corresponding one. The upper limit of the encoder operation load (cpuLd _max ), the normalized estimated operator load (cpuLd _normalized ) is obtained, which is expressed as follows: cpuLd _normalized = cpuLd _effective / cpuLd _max ; and c2. Compare the normalization of all transcoders The operator load is estimated, and the transcoding task is assigned to the transcoder with the lowest load of the normalized predictor. A device for optimizing video transcoding, comprising: a transcoder performance function determining unit, determining a transcoder performance function, decoding information and encoding in a known transcoding task obtained from a transcoder Information, filtering information, corresponding transcoder load information, calculating and outputting coefficients of the transcoder performance function, wherein the transcoder performance function can use the decoding information, the encoding information, the filtering information to estimate and calculate the conversion Transcoder load information; a transcoder load information judging unit, inputting a new transcoding task information, and using the transcoder performance function to estimate the transcoder load information after adding a new transcoding task to obtain An estimated operator load, and comparing the transcoder load information with the estimated operator load to determine whether the transcoder load information is valid, and outputting the transcoder load information; and a transcoding The task assignment unit inputs the transcoder load information, normalizes the transcoder load information determined to be in an active state, and assigns a transcoding task according to the transcoder load information.