WO2016095241A1

WO2016095241A1 - Method and device for selecting video encoding and decoding hardware platform

Info

Publication number: WO2016095241A1
Application number: PCT/CN2014/094496
Authority: WO
Inventors: 沈案
Original assignee: 深圳Tcl数字技术有限公司
Priority date: 2014-12-17
Filing date: 2014-12-22
Publication date: 2016-06-23
Also published as: CN105791833A; CN105791833B

Abstract

Disclosed is a method for selecting a video encoding and decoding hardware platform, comprising: classifying all the test video sources in a test video source library into different test video source subsets corresponding to different video technical indexes; on a pre-set hardware platform, performing an encoding and decoding performance test evaluation on each video encoding and decoding program for all the test video sources in a test video source subset corresponding to a pre-set video technical index, so as to obtain a typical video encoding and decoding program corresponding to the pre-set video technical index; and on the typical video encoding and decoding program, performing the encoding and decoding performance test evaluation on each hardware platform to be selected for all the test video sources in the test video source subset corresponding to the pre-set video technical index, so as to obtain a hardware platform to be selected with an encoding and decoding performance test comprehensive value satisfying a second pre-set condition. Also disclosed is a device for selecting a video encoding and decoding hardware platform. The present invention can improve the efficiency and accuracy of selecting a video encoding and decoding hardware platform.

Description

Method and device for selecting video codec hardware platform

Technical field

The present invention relates to the field of video coding and decoding, and in particular, to a method and apparatus for selecting a video codec hardware platform.

Background technique

With the rapid development and wide application of computer technology, especially digital multimedia technology, the importance of video codec technology in product design and development is self-evident. Generally, a video codec compresses video original information to solve storage and transmission difficulties. When playing video, the video codec decodes the compressed video data and restores the video information by inverse operation. The video codec needs more and more high-demand hardware platform support for the spatial/time domain sampling, macroblock motion compensation, block transform/subband decomposition, or entropy coding in the codec process, and With the improvement of video index resolution, codec delay, video stream control and other technical indicators, the choice of hardware platform will be more important.

At present, the related research in the field of video coding and decoding mainly focuses on the optimization of coding format, codec computation and compatibility, and basically belongs to the research of video codec software. The mainstream video codecs are: DivX, WMV, RealVideo. , Cinepak, etc.; video coding and decoding hardware research is mainly aimed at specific hardware platforms, specific technical indicators to optimize, resulting in a lack of a common and flexible and efficient hardware platform evaluation program in the project development process; In addition, the current hardware The platform evaluation scheme usually needs to test all the test chip sources for each video codec in each video codec on each candidate hardware platform, so that the evaluation workload is large, resulting in an excessive evaluation period; Coupled with changes in project requirements, the assessment will be highly repetitive and seriously affect the progress of the project. Therefore, it is necessary to provide a method for improving the efficiency of video codec hardware platform selection.

Summary of the invention

The main object of the present invention is to provide a method and apparatus for selecting a video codec hardware platform, which aims to improve the efficiency of video codec hardware platform selection.

To achieve the above objective, the present invention provides a method for selecting a video codec hardware platform, including the following steps:

A. classify all test film sources in the test source library according to preset classification rules, and divide into different test film source subsets corresponding to different video technical indicators;

B. On the preset hardware platform, all test film sources in the test chip source subset corresponding to the preset video technical indicators are subjected to codec performance test evaluation for each video codec in the video codec library. Obtaining a typical video codec program that meets the first preset condition corresponding to the preset video technical indicator;

C. On the typical video codec program, all the test chip sources in the test chip source subset corresponding to the preset video technical indicators are coded and tested for each candidate hardware platform in the candidate hardware platform. Evaluate to obtain a candidate hardware platform whose codec performance test comprehensive value satisfies the second preset condition.

Preferably, the process of the codec performance test evaluation is: playing the corresponding test film source on the corresponding video codec on the corresponding hardware platform, subjective visual quality, objective fidelity and the image obtained by the play. The quality of the three aspects of color is comprehensively evaluated to obtain the codec performance evaluation value of the corresponding video codec or the corresponding candidate hardware platform.

Preferably, the preset video technical indicator is one or more, when the preset video technical indicator is two or more;

The step B includes: sequentially, on the preset hardware platform, all the test chip sources in the test chip source subset corresponding to each preset video technical indicator to each video codec in the video codec library. Performing a codec performance test evaluation to obtain a typical video codec corresponding to the first preset condition corresponding to each preset video technical indicator;

The step C includes:

C1, in turn, in the typical video codec corresponding to each preset video technical indicator, all the test film sources in the test chip source subset corresponding to the corresponding preset video technical indicators are for each of all the selected hardware platforms. Selecting a hardware platform for performing a codec performance test evaluation to obtain a codec performance test value corresponding to a corresponding preset video technical indicator of each candidate hardware platform;

C2. Set weight values for two or more codec performance test values corresponding to two or more preset video technical indicators of each candidate hardware platform according to preset requirements, and obtain each candidate hardware in a weighted average manner. The integrated value of the codec performance test corresponding to the two or more preset video technical indicators of the platform, thereby obtaining the candidate hardware platform whose codec performance test comprehensive value is greater than the preset comprehensive value.

Preferably, the step B specifically includes:

B1. All test source sources in the test slice source subset corresponding to the preset video technical indicators are classified according to a preset hierarchical rule, and are divided into different levels of secondary test piece source subsets corresponding to the preset hierarchical rules;

B2. On the preset hardware platform, all test chip sources in the secondary test chip source subset of the preset level are subjected to codec performance test evaluation for each video codec in the video codec library, and The codec performance test evaluation value is used as the initial clustering center of the K-means algorithm;

B3. On the preset hardware platform, all the test film sources except the secondary test piece source subset of the preset level in the test chip source subset corresponding to the preset video technical indicators are used for each of the video coding and decoding programs. A video codec performs a codec performance test evaluation, and the obtained codec performance test evaluation value is used as a clustering basis of the K-means algorithm;

B4. The K-means algorithm performs processing to sort all the video codecs in the video codec library according to the performance and the inferiority, and selects the video codec in the front row to be matched with the preset video technical indicators. Typical video codec.

Preferably, before the step A, the method further includes:

D. Retrieve the newly released test film source, update the newly released test film source to the test film source library, and retrieve the newly released video codec and update the newly released video codec to the video. Codec library.

Preferably, the step D further comprises: deleting a video codec that is too low in the video codec library.

In addition, to achieve the above object, the present invention further provides an apparatus for selecting a video codec hardware platform, including:

The source source classification module is configured to classify all test film sources in the test source library according to a preset classification rule, and divide into different test chip source subsets corresponding to different video technical indicators;

The typical program evaluation module is configured to code and decode each video codec in the video codec library in the test chip source subset corresponding to the preset video technical indicator on the preset hardware platform. Testing and evaluating to obtain a typical video codec corresponding to the first preset condition corresponding to the preset video technical indicator;

The hardware evaluation module is configured to code and decode all the test chip sources in the test chip source subset corresponding to the preset video technical indicators on each of the candidate hardware platforms in the candidate video codec. The test is evaluated to obtain a candidate hardware platform whose performance test comprehensive value satisfies the second preset condition.

The typical program evaluation module is configured to sequentially encode and decode each video source in the video codec library in the test chip source subset corresponding to each preset video technical indicator on a preset hardware platform. The program performs a codec performance test evaluation to obtain a typical video codec program corresponding to the first preset condition corresponding to each preset video technical indicator;

The hardware evaluation module is configured to sequentially, in a typical video codec corresponding to each preset video technical indicator, all test source sources in the test chip source subset corresponding to the corresponding preset video technical indicators to all the selected hardware platforms. Each candidate hardware platform performs a codec performance test evaluation to obtain a codec performance test value corresponding to a corresponding preset video technical indicator of each candidate hardware platform; and is used for each waiting according to a preset requirement Selecting two or more codec performance test values corresponding to two or more preset video technical indicators of the hardware platform to set weight values, and obtaining two or more preset video technologies for each candidate hardware platform in a weighted average manner The integrated value of the codec performance test corresponding to the indicator, thereby obtaining the candidate hardware platform whose codec performance test comprehensive value is greater than the preset comprehensive value.

Preferably, the slice source classification module is further configured to classify all the test piece sources in the test piece source subset corresponding to the preset video technical indicators according to a preset classification rule, and divide into different levels corresponding to the preset classification rules. Secondary test chip source subset;

The typical program evaluation module includes an initial cluster center acquisition unit, a cluster basis acquisition unit, and a sort unit:

The initial clustering center acquiring unit is configured to compile each video codec in the video codec library in the preset level of the secondary test chip source subset on the preset hardware platform. Decoding performance test evaluation, and the obtained codec performance test evaluation value is used as an initial clustering center of the K-means algorithm;

The clustering obtaining unit is configured to: on the preset hardware platform, all test film sources except a subset of the secondary test piece source of the preset level in the test chip source subset corresponding to the preset video technical index Each video codec in the codec performs a codec performance test evaluation, and the obtained codec performance test evaluation value is used as a clustering basis of the K-means algorithm;

a sorting unit for sorting all video codecs in the video codec library by performance calculation by K-means algorithm, and selecting a video codec located in the top of the sorting as a preset video technology A typical video codec corresponding to the indicator.

Preferably, the device for selecting a video codec hardware platform further includes an update module, configured to retrieve a newly released test slice source, and update the newly released test slice source to the test slice source library, and use Retrieve the newly released video codec and update the newly released video codec to the video codec library.

Preferably, the update module is further configured to delete a video codec that is too low in the video codec library.

The method for selecting a video codec hardware platform provided by the present invention classifies all the test chip sources in the test chip source library, so as to perform test evaluation on the codec performance of a certain video technical indicator on different candidate hardware platforms. When you can quickly find the appropriate corresponding test source, improve the timeliness of the evaluation test; also obtain the typical video codec corresponding to the preset video technical indicators through the K-means algorithm, so that it will correspond to the preset video technical indicators. When all the test chip sources in the test chip source subset evaluate the codec performance test for each candidate hardware platform, only the typical video codec corresponding to the preset video technical specification needs to perform the codec performance test evaluation, and There is no need to perform codec performance test evaluation in other atypical video codecs, which can greatly reduce the evaluation test workload of the evaluation process and improve the efficiency of the evaluation test. It also integrates two or more preset video technologies by weighted average method. The codec performance evaluation results of the indicator to arrive at the final target hardware platform, Can improve the accuracy of the evaluation test.

DRAWINGS

1 is a schematic flow chart of an embodiment of a method for selecting a video codec hardware platform according to the present invention;

2 is a schematic flowchart showing the steps of obtaining a typical video codec program corresponding to a preset video technical indicator by the K-means algorithm in FIG. 1;

3 is a schematic diagram of a detailed process of the steps of performing a codec performance test evaluation on each candidate hardware platform to obtain a target hardware platform in a typical video codec program in FIG. 1;

4 is a schematic flowchart diagram of another embodiment of a method for selecting a video codec hardware platform according to the present invention;

5 is a schematic diagram of functional modules of an apparatus for selecting a video codec hardware platform according to an embodiment of the present invention;

6 is a schematic diagram of a refinement function module of the typical program evaluation module of FIG. 5;

FIG. 7 is a schematic diagram of functional modules of another embodiment of a device for selecting a video codec hardware platform according to the present invention.

The implementation, functional features, and advantages of the present invention will be further described in conjunction with the embodiments.

detailed description

It is understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

At present, there are many video codec programs on the market, such as DivX, WMV, RealVideo, Cinepak, etc.; there are various video technical indicators for characterizing video performance, such as resolution, bit rate, frame rate and compression ratio; The video codec program is not unique in the configuration of the hardware platform (such as smart TV, computer, etc.) for playing video; on a hardware platform, when playing the same test source on different video encoders, the viewer will have Different look and feel; on a version of the video encoder, when playing the same test source on different hardware platforms, the viewer will have a different look and feel. In the process of developing a new video codec hardware platform for project requirements, R&D engineers usually need to select a target hardware platform with excellent comprehensive performance among multiple candidate hardware platforms.

The present invention provides a method of selecting a video codec hardware platform.

Referring to FIG. 1, in an embodiment, the method for selecting a video codec hardware platform includes:

In step S10, all the test chip sources in the test source library are classified according to a preset classification rule, and are divided into different test chip source subsets corresponding to different video technical indicators.

In this embodiment, the test source library is used to collect more authoritative or professional video codec test institutions (such as some mainstream video codec manufacturers or standard organizations such as the International Telecommunication Union) for different video technical indicators. The corresponding test chip source; before testing the codec performance test of the hardware platform to be selected, first classify all test film sources in the test source library, and the classification rule is a test piece that will be used to test the same video technical index. The source points are in the same test source subset, so that when the codec performance test for a certain video technology indicator is evaluated on different codec hardware platforms, the corresponding corresponding test source can be quickly found, and the time limit of the evaluation test can be improved. In addition, when the codec performance test for a certain video technology indicator is performed on different codec hardware platforms, the evaluation of the test film source that does not correspond to the video technical indicator is not performed, and the evaluation test of the evaluation process can be greatly reduced. The workload can improve the efficiency of the evaluation test.

Step S20: Perform, on a preset hardware platform, all the test chip sources in the test chip source subset corresponding to the preset video technical indicators, and perform codec performance test evaluation on each video codec in the video codec library. And obtaining a typical video codec that meets the first preset condition corresponding to the preset video technical indicator.

Since there are many video codec programs that are currently available, for a given video codec, the codec performance may be different for different video technical indicators; in this embodiment, the currently occurring video codec is collected in the video code. In the decoding library, before the evaluation of the codec performance test of the candidate hardware platform, the preset hardware platform (in this embodiment, the preset hardware platform is a mature hardware platform, such as a computer), All the test chip sources in the test chip source subset corresponding to the video technical indicators are subjected to codec performance test evaluation for each video codec program, and all video codecs are sorted according to the performance and performance of the codec performance test evaluation. The video codec in the top of the sorting is selected as a typical video codec corresponding to the preset video technical indicator. In this embodiment, the first preset condition is that the performance sorting is in the forefront, of course, the typical video codec There can be more than one, usually you can choose the performance of the top three video codecs in the top three as the default Specifications frequency corresponding to typical video codecs.

Step S30, on the typical video codec, all the test chip sources in the test chip source subset corresponding to the preset video technical indicators are coded and decoded for each candidate hardware platform in all the selected hardware platforms. The test is evaluated to obtain a candidate hardware platform whose codec performance test comprehensive value satisfies the second preset condition.

In this embodiment, when all test chip sources in the test chip source subset corresponding to the preset video technical indicators are subjected to codec performance test evaluation for each candidate hardware platform, only need to correspond to preset video technical indicators. The typical video codec performs the codec performance test evaluation without the need for codec performance test evaluation in other atypical video codecs, which can greatly reduce the evaluation test workload of the evaluation process and improve the efficiency of the evaluation test; In this embodiment, when the codec performance test for the preset video technical indicator is performed on each candidate hardware platform, the corresponding codec performance test comprehensive value is obtained, and the corresponding codec of the candidate hardware platform is selected. When the performance test comprehensive value is greater than the preset comprehensive value, the candidate hardware platform is the target hardware platform.

In this embodiment, whether the video codec is coded and tested by the test chip source on the preset hardware platform, or the codec source is used to encode and decode the hardware platform in the typical video codec. Performance test evaluation, the process of codec performance test evaluation is similar, specifically, the test chip source is played on the measured video codec program on the preset hardware platform or the test piece is on the candidate hardware platform. The source is played on a typical video codec program, and the quality of the coded video codec or the hardware platform to be selected is evaluated by evaluating the quality of the image obtained by the playback. Preferably, the quality of the image obtained by the playback is optimized. The following three aspects are evaluated: 1. The subjective visual quality of the images obtained by the broadcast is evaluated, and the evaluation sub-items may include, but are not limited to, distortion sensitivity, sharpness, and edge distortion, etc., and the evaluation means may be passed through but not limited to Questionnaires, etc.; second, evaluate the objective fidelity of the images obtained by the broadcast, and the evaluation sub-projects can Including but not limited to brightness performance, contrast performance and structural similarity, etc.; third, the color of the image obtained by the broadcast is evaluated, and the evaluation sub-items may include, but are not limited to, hue, brightness, saturation, etc.; The quality of the image is evaluated in the above three aspects. It can be, but is not limited to, scoring the image obtained for each evaluation sub-item based on the preset scoring criteria, and finally evaluating the score and value or all the evaluations of all the evaluation sub-items. The average score value of the sub-item is used as the codec performance evaluation value of the video codec program or the candidate hardware platform to be tested, and the codec performance evaluation value corresponding to different test video codec programs or different candidate hardware platforms is compared. The pros and cons of the codec performance, the conventional codec decoding program with a large evaluation value of the codec performance or the codec performance of the candidate hardware platform is superior.

Further, in this embodiment, referring to FIG. 2, when the preset video technical indicator is one, step S20 specifically includes:

Step S21: All the test film sources in the test film source subset corresponding to the preset video technical indicators are classified according to a preset hierarchical rule, and are divided into different levels of secondary test film source subsets corresponding to the preset hierarchical rules.

Step S22, on the preset hardware platform, all the test chip sources in the secondary test chip source subset of the preset level are subjected to codec performance test evaluation for each video codec in the video codec library, and The obtained codec performance test evaluation value is used as the initial clustering center of the K-means algorithm;

Step S23, on the preset hardware platform, all the test film sources except the secondary test piece source subset of the preset level in the test chip source subset corresponding to the preset video technical index are used in all the video codec programs. Each video codec performs a codec performance test evaluation, and the obtained codec performance test evaluation value is used as a clustering basis of the K-means algorithm;

Step S24, sorting all the video codecs in the video codec library by performance calculation by the calculation process of the K-means algorithm, and selecting the video codec in the front row to be matched with the preset video technical indicators. Typical video codec.

In this embodiment, a typical video codec corresponding to the preset video technical index is obtained by the K-means algorithm, and the initial cluster center and the clustering basis need to be obtained before the K-means algorithm is processed. For example, the test chip source subset corresponding to the preset video technical indicator is divided into different levels of secondary test piece source subsets corresponding to the preset hierarchical rule, so that only one of the initial cluster centers can be obtained when acquiring the initial cluster center. The test of the secondary test source subset of the level (usually selecting the intermediate level so that the initial clustering center has good convergence) can reduce the test volume of the initial cluster center and improve the evaluation test. Efficiency, in particular, when the similarity of the source of the secondary test piece source subset is high, the initial cluster center can be obtained even by testing only a part of the source of the secondary test piece source subset.

The following preset video technical indicators take the resolution as an example to illustrate how to obtain a typical video codec corresponding to the resolution:

First, all test chip sources in the subset of test chip sources corresponding to the resolution are classified according to their own resolution size approximation. For example, for all test chip sources in the current resolution test chip source subset, according to 176× 144, 352×288, 1280×720, 1920×1080 perform four levels of division; then, when acquiring the corresponding initial cluster center, the resolution can be 1280×720 (intermediate level, of course, here too) The test slice source in the secondary test chip source subset of the selected resolution 352×288 is subjected to codec performance test evaluation for each video codec in all video codecs on a preset hardware platform; When obtaining the corresponding clustering basis, all the test chip sources in the remaining level of the secondary test chip source subset are coded and decoded on each of the video codec programs on the preset hardware platform. Test evaluation; finally, all video codecs are sorted by performance according to the calculation process of K-means algorithm, so as to select the video editing in the top of the sorting. A typical procedure for the resolution of video codecs corresponds.

It should be emphasized that, for different video technical indicators, the corresponding video codec is usually different from each other. Therefore, when the preset video technical indicators are two or more, step S21 needs to be performed separately for different video technical indicators. Step S24, to obtain a corresponding typical video codec corresponding to the corresponding video technical indicator. That is, when the preset video technical indicators are two or more, step S20 is: sequentially, on the preset hardware platform, all the test film source pairs in the test chip source subset corresponding to each preset video technical indicator. Each video codec in the codec library performs a codec performance test evaluation to obtain a typical video codec corresponding to the first preset condition corresponding to each preset video technical indicator.

Further, in this embodiment, referring to FIG. 3, when the preset video technical indicators are two or more, step S30 includes:

Step S31: sequentially, in the typical video codec corresponding to each preset video technical indicator, all test source sources in the test slice source subset corresponding to the corresponding preset video technical indicators to each of all the selected hardware platforms. The candidate hardware platform performs the codec performance test evaluation to obtain the codec performance test value corresponding to the corresponding preset video technical indicator of each candidate hardware platform;

Step S32, synthesizing two or more codec performance test values corresponding to two or more preset video technical indicators of each candidate hardware platform, to obtain two or more preset videos of each candidate hardware platform. The comprehensive value of the codec performance test corresponding to the technical indicator, thereby obtaining the candidate hardware platform whose integrated value of the codec performance test satisfies the second preset condition.

In this embodiment, when the preset video technical indicators are two or more, since the typical video codec programs corresponding to each preset video technical indicator are usually different, it is required to sequentially correspond to each preset video technical indicator. The typical video codec program will perform all the test chip sources in the test chip source subset corresponding to the corresponding preset video technical indicators, and perform codec performance test evaluation and evaluation test on each candidate hardware platform in all the selected hardware platforms. The codec performance test value corresponding to the corresponding preset video technical indicator of each candidate hardware platform is obtained. Obviously, the codec performance test value of each candidate hardware platform is also two or more, and the two or more are passed. The codec performance test value is comprehensively compiled to obtain the integrated value of the codec performance test, and is compared with the preset integrated value, and the candidate hardware platform whose codec performance test value is greater than the preset comprehensive value is the target hardware platform.

In this embodiment, specifically, step S32 is specifically, according to the preset requirement (required for the research and development project), two or more codec performance test values corresponding to two or more preset video technical indicators of each candidate hardware platform are respectively The weight value is set, and the coded performance test comprehensive value corresponding to the two or more preset video technical indicators of each candidate hardware platform is obtained in a weighted average manner, so that the integrated value of the codec performance test is greater than the preset comprehensive value. Optional hardware platform. For example, when the R&D project requires the target hardware platform to have better codec performance for no resolution and code rate, the resolution is especially important. Therefore, the codec performance test value corresponding to the resolution can be set to be greater than 0.5, and The codec performance test value corresponding to the code rate is set to be less than 0.5, and the coded performance test comprehensive value of each candidate hardware platform is obtained in a weighted average manner, and then compared with the preset integrated value. In this embodiment, by combining the codec performance evaluation results of two or more preset video technical indicators to obtain a final target hardware platform, the accuracy of the evaluation test can be improved.

The method for selecting a video codec hardware platform provided by this embodiment classifies all test chip sources in the test chip source library, so as to perform codec performance test on a certain candidate hardware platform for a certain video technical indicator. During the evaluation, it can quickly find the appropriate corresponding test source and improve the timeliness of the evaluation test. The K-means algorithm also obtains the typical video codec corresponding to the preset video technical indicators, so that it will be compared with the preset video technical indicators. When all the test chip sources in the corresponding test chip source subset perform the codec performance test evaluation on each candidate hardware platform, only the typical video codec program corresponding to the preset video technical indicator needs to perform the codec performance test evaluation. There is no need to perform codec performance test evaluation in other atypical video codecs, which can greatly reduce the evaluation test workload of the evaluation process and improve the efficiency of the evaluation test. It also integrates more than two preset videos by weighted average method. Codec performance evaluation results of technical indicators to arrive at the final target hardware platform Can improve the accuracy of the evaluation test.

Referring to FIG. 4, based on the foregoing embodiment of the method for selecting a video codec hardware platform of the present invention, before the step S10 is performed, the method for selecting a video codec hardware platform further includes:

Step S40, retrieving the newly released test piece source, updating the newly released test piece source to the test piece source library, and retrieving the newly released video codec and updating the newly released video codec to Video codec library.

Due to the rapid development of digital multimedia technology, the development of codec technology is getting faster and faster. The specific performance is that the video codec program is updated frequently, and the corresponding test source is also updated frequently. If so, if the hardware platform is to be edited. Before the decoding performance test evaluation, the test source library and the video codec library are not updated, and it is difficult to ensure that the test film source and the video codec program for the codec performance evaluation test of the hardware platform to be selected can cover important codec technology. And its development trend. In this embodiment, the actions of updating the test fragment source library and the video codec library include acquiring a new test film source, a new video codec, and a video encoding and decoding program that is too low in version and no longer used. The video codec is deleted, and the corresponding test film source is deleted at the same time. In this embodiment, before all the test piece sources in the test piece source library are classified according to the preset classification rule, the new test sample source is retrieved, and the newly released test piece source is updated to the test piece. In the source library, simultaneously retrieve to obtain the newly released video codec, and update the newly released video codec to the video codec library to ensure that the codec performance evaluation test can be used for the hardware platform to be selected. Cover important codec technology and its development trends.

It should be emphasized that, in this embodiment, the test source library and the video codec library are independent of each other, and the test source in the test source library is classified according to a preset rule and from the video codec library. The selection of typical video codecs in all video codecs is also independent of each other, thus improving the flexibility of the entire process of selecting a video codec hardware platform.

The present invention further provides an apparatus for selecting a video codec hardware platform.

Referring to FIG. 5, in an embodiment, the apparatus for selecting a video codec hardware platform includes:

The source classification module 10 is configured to classify all test film sources in the test source library according to preset classification rules, and divide into different test chip source subsets corresponding to different video technical indicators;

The typical program evaluation module 20 is configured to code and decode each video codec in the video codec library in the test chip source subset corresponding to the preset video technical indicator on the preset hardware platform. Performing a performance test evaluation to obtain a typical video codec corresponding to the first preset condition corresponding to the preset video technical indicator;

The hardware evaluation module 30 is configured to code and decode all the test chip sources in the test slice source subset corresponding to the preset video technical indicators on each of the candidate hardware platforms in the candidate video codec. The performance test is evaluated to obtain a candidate hardware platform whose performance test comprehensive value satisfies the second preset condition.

In this embodiment, the slice source classification module 10 includes a test slice source library for collecting more authoritative or professional video codec test institutions (such as some mainstream video codec manufacturers or the International Telecommunication Union). The standard test organization source for the different video technical indicators is released; before the coded performance test evaluation of the selected hardware platform, the slice source classification module 10 first tests all the test source sources in the test source library. To classify, the classification rule is to divide the test piece sources used to test the same video technical indicators into the same test chip source subset, so as to perform the evaluation of the codec performance test for a certain video technical indicator on different codec hardware platforms. It can quickly find the appropriate corresponding test source and improve the timeliness of the evaluation test.

Since there are many video codec programs that are currently available, for a given video codec, the codec performance may be different for different video specifications. In this embodiment, the typical program evaluation module 20 includes a video codec library. The video codec library is used to collect the currently occurring video codec; before the candidate hardware platform performs the codec performance test evaluation, the typical program evaluation module 20 is first on the preset hardware platform (in this embodiment, The hardware platform is a mature hardware platform, such as a computer. All the test chip sources in the test chip source subset corresponding to the preset video technical indicators are subjected to codec performance test evaluation for each video codec program, and all will be evaluated. The video codec is sorted according to the performance of the codec performance test evaluation, and the video codec in the top of the sort is selected as a typical video codec corresponding to the preset video technical indicator; in this embodiment, A preset condition is that the performance ordering is in the forefront. Of course, there are multiple typical video codecs. Generally, the three video codecs ranked in the top three can be selected as the typical video codec corresponding to the preset video technical indicators.

In this embodiment, the hardware evaluation module 30 only needs to be in the preset with all the test chip sources in the test chip source subset corresponding to the preset video technical indicators for performing the codec performance test evaluation on each candidate hardware platform. The typical video codec corresponding to the video technical indicators evaluates the codec performance test without the need for codec performance test evaluation in other atypical video codecs, which can greatly reduce the evaluation test workload of the evaluation process and improve the evaluation. The efficiency of the test; in this embodiment, when the hardware evaluation module 30 performs the test and evaluation of the codec performance test for the preset video technical indicator for each candidate hardware platform, the corresponding codec performance test comprehensive value is obtained. When the corresponding codec performance test value of the selected hardware platform is greater than the preset integrated value, the candidate hardware platform is the target hardware platform.

Further, in this embodiment, the slice source classification module 10 is further configured to classify all test slice sources in the test slice source subset corresponding to the preset video technical indicators according to a preset hierarchical rule, and divide into and preset preset hierarchical rules. Corresponding different levels of secondary test chip source subsets;

Referring to FIG. 6, the typical program evaluation module 20 specifically includes an initial cluster center obtaining unit 21, a clustering basis obtaining unit 22, and a sorting unit 23:

The initial clustering center acquiring unit 21 is configured to perform, on a preset hardware platform, all the test chip sources in the secondary test chip source subset of the preset level to each video codec in the video codec library. Codec performance test evaluation, the obtained codec performance test evaluation value as the initial clustering center of the K-means algorithm;

The clustering obtaining unit 22 is configured to: on the preset hardware platform, all the test chip source pairs except the secondary test piece source subset of the preset level in the test chip source subset corresponding to the preset video technical index Each video codec in the video codec performs a codec performance test evaluation, and the obtained codec performance test evaluation value is used as a clustering basis of the K-means algorithm;

The sorting unit 23 is configured to sort all the video codecs in the video codec library by performance calculation by the K-means algorithm, and select the video codec in the top of the sorting as the preset video. A typical video codec corresponding to the technical specifications.

In this embodiment, the typical program evaluation module 20 obtains a typical video codec corresponding to the preset video technical specification by using the K-means algorithm; specifically, the initial clustering center needs to be obtained before the K-means algorithm is processed. And the clustering basis; in this embodiment, the source of the test piece corresponding to the preset video technical indicator is divided into a subset of the secondary test piece source corresponding to the preset hierarchical rule by the slice source classification module 10, so as to facilitate the initial The cluster center acquisition unit 21 can test only a subset of the secondary test piece source when only one of the levels (usually selecting the middle level thereof so that the initial cluster center has good convergence) when acquiring the initial cluster center. That is, thereby reducing the amount of testing to obtain the initial clustering center, and improving the efficiency of the evaluation test. In particular, when the similarity of the source of the secondary test chip source subset is high, it is even possible to pass only the secondary test piece. The initial clustering center is obtained by testing a part of the source in the source subset.

First, the slice source classification module 10 classifies all test chip sources in the subset of test chip sources corresponding to the resolution according to their own resolution size approximation, for example, for all tests in the current resolution test chip source subset. The slice source is divided into four levels according to 176×144, 352×288, 1280×720, and 1920×1080; then, the initial cluster center acquiring unit 21 can obtain a resolution of 1280 when acquiring the corresponding initial cluster center. The test chip source of the secondary test chip source subset of the ×720 level (intermediate level, of course, the resolution can also be selected as 352×288) is on each of the video coding and decoding programs on the preset hardware platform. The video codec performs a codec performance test evaluation; and the clustering obtaining unit 22 obtains the corresponding clustering source of the remaining level of the second test piece source subset on the preset hardware platform. A codec performance test evaluation is performed for each video codec in all video codecs; finally, the sorting unit 23 compiles all the videos by the calculation process of the K-means algorithm. The decoding program is sorted according to the performance and the inferiority, so that the video codec located in the top of the sorting is selected as a typical video codec corresponding to the resolution.

It should be emphasized that, for different video technical indicators, the corresponding typical video codec is usually different. Therefore, when the preset video technical indicators are more than two, the K-means algorithm needs to be adopted for different video technical indicators. Obtain the corresponding typical video codec corresponding to the corresponding video technical indicators. That is, when the preset video technical indicators are two or more, the typical program evaluation module 20 is specifically configured to sequentially perform all the tests in the test chip source subset corresponding to each preset video technical indicator on the preset hardware platform. The film source performs a codec performance test evaluation on each video codec in the video codec library to obtain a typical video codec corresponding to the first preset condition corresponding to each preset video technical indicator.

Further, in this embodiment, when the preset video technical indicators are two or more, the hardware evaluation module 30 is specifically configured to sequentially and corresponding presets on a typical video codec corresponding to each preset video technical indicator. All test chip sources in the test chip source subset corresponding to the video technical indicators perform codec performance test evaluation on each candidate hardware platform of all the selected hardware platforms to obtain the corresponding preset video of each candidate hardware platform. The codec performance test value corresponding to the technical indicator; and the two or more codec performance test values corresponding to the two or more preset video technical indicators of each candidate hardware platform are comprehensively sorted to obtain each candidate to be selected The integrated value of the codec performance test corresponding to the two or more preset video technical indicators of the hardware platform, thereby obtaining the candidate hardware platform whose comprehensive value of the codec performance test meets the second preset condition.

In this embodiment, when the preset video technical indicators are more than two, since the typical video codec corresponding to each preset video technical indicator is usually different, the hardware evaluation module 30 is required to be in each preset. The typical video codec corresponding to the video technical indicator performs codec performance test on each of the candidate hardware platforms in all the candidate hardware platforms in the test chip source subset corresponding to the corresponding preset video technical indicators. The evaluation and evaluation test will obtain the codec performance test value corresponding to the corresponding preset video technical indicator of each candidate hardware platform. Obviously, the codec performance test value of each candidate hardware platform is also two or more. The two or more codec performance test values are comprehensively combined to obtain a comprehensive value of the codec performance test, which is compared with the preset integrated value, and the codec performance test value is greater than the preset integrated value of the candidate hardware platform as the target hardware platform. .

In this embodiment, specifically, the hardware evaluation module 30 respectively determines, for each candidate hardware platform, two or more codec performance test values corresponding to two or more preset video technical indicators according to preset requirements (requirements of the research and development project). The weight value is set, and the coded performance test comprehensive value corresponding to the two or more preset video technical indicators of each candidate hardware platform is obtained in a weighted average manner, so that the integrated value of the codec performance test is greater than the preset comprehensive value. Optional hardware platform. For example, when the R&D project requires the target hardware platform to have better codec performance for no resolution and code rate, where resolution is particularly important, the hardware evaluation module 30 can set the codec performance test value corresponding to the resolution to If the codec performance test value corresponding to the code rate is set to be less than 0.5, the coded performance test integrated value of each candidate hardware platform is obtained in a weighted average manner, and then compared with the preset integrated value. In this embodiment, the hardware evaluation module 30 integrates the codec performance evaluation results of two or more preset video technical indicators to obtain a final target hardware platform, which can improve the accuracy of the evaluation test.

The device for selecting a video codec hardware platform provided by this embodiment classifies all the test chip sources in the test chip source library by the slice source classification module 10, so as to perform a certain video technical index on different candidate hardware platforms. When the codec performance test is evaluated, the corresponding corresponding test chip source can be quickly found to improve the timeliness of the evaluation test; the typical program evaluation module 20 also obtains a typical video code corresponding to the preset video technical index through the K-means algorithm. The decoding process is such that when all the test chip sources in the test chip source subset corresponding to the preset video technical indicators are subjected to the codec performance test evaluation for each candidate hardware platform, only the typical corresponding to the preset video technical indicators is required. The video codec performs the codec performance test evaluation without the need for codec performance test evaluation in other atypical video codecs, which can greatly reduce the evaluation test workload of the evaluation process and improve the efficiency of the evaluation test; The evaluation module 30 also integrates two or more preset video technology fingers by means of weighted averaging The target codec performance evaluation results to arrive at the final target hardware platform can improve the accuracy of the evaluation test.

It should be emphasized that, in this embodiment, the slice source classification module 10, the typical program evaluation module 20, and the hardware evaluation module 30 have high independence, and the coupling degree between the modules is low, which is improved. The flexibility of the device that selects the video codec hardware platform also improves the ease of maintenance of the device that selects the video codec hardware platform.

Referring to FIG. 7, in the foregoing embodiment of the apparatus for selecting a video codec hardware platform of the present invention, the apparatus for selecting a video codec hardware platform further includes:

The update module 40 is configured to retrieve a newly released test piece source, update the newly released test piece source to the test piece source library, and retrieve the newly released video codec and retrieve the newly released video. The codec is updated to the video codec library.

Due to the rapid development of digital multimedia technology, the development of codec technology is getting faster and faster. The specific performance is that the video codec program is updated frequently, and the corresponding test source is also updated frequently. If so, if the hardware platform is to be edited. Before the performance evaluation of the decoding test, the update module 40 does not update the test source library and the video codec library, and it is difficult to ensure that the test film source and the video codec can be covered for the codec performance evaluation test of the hardware platform to be selected. Important codec technology and its development trend. In this embodiment, the update module 40 updates the test fragment source library and the video codec library, including acquiring a new test slice source, a new video codec, and a version of the video codec that is too low. The video codec that is no longer used deletes, deletes the corresponding test source, and so on. In this embodiment, before the slice source classification module 10 classifies all the test slice sources in the test slice source library according to the preset classification rule, the update module 40 first searches to obtain the newly released test slice source, and the new release is performed. The test sample source is updated into the test source library, and retrieved to obtain the newly released video codec, and the newly released video codec is updated to the video codec library to ensure that the hardware platform is to be selected. It can cover important codec technology and its development trend when performing codec performance evaluation test.

The above are only the preferred embodiments of the present invention, and are not intended to limit the scope of the invention, and the equivalent structure or equivalent process transformations made by the description of the present invention and the drawings are directly or indirectly applied to other related technical fields. The same is included in the scope of patent protection of the present invention.

Claims

A method for selecting a video codec hardware platform, wherein the method for selecting a video codec hardware platform comprises the following steps:

A. classify all test film sources in the test source library according to preset classification rules, and divide into different test film source subsets corresponding to different video technical indicators;

B. On the preset hardware platform, all test film sources in the test chip source subset corresponding to the preset video technical indicators are subjected to codec performance test evaluation for each video codec in the video codec library. Obtaining a typical video codec program that meets the first preset condition corresponding to the preset video technical indicator;

C. On the typical video codec program, all the test chip sources in the test chip source subset corresponding to the preset video technical indicators are coded and tested for each candidate hardware platform in the candidate hardware platform. Evaluate to obtain a candidate hardware platform whose codec performance test comprehensive value satisfies the second preset condition.
The method for selecting a video codec hardware platform according to claim 1, wherein the process of evaluating the codec performance test is: placing a corresponding test piece source on a corresponding video codec on a corresponding hardware platform. Play, comprehensively evaluate the quality of the subjective visual quality, objective fidelity and color of the played image to obtain the codec performance evaluation value of the corresponding video codec or the corresponding candidate hardware platform.
The method for selecting a video codec hardware platform according to claim 1, wherein the preset video technical indicator is one or more, when the preset video technical indicator is two or more;

The step B includes: sequentially, on the preset hardware platform, all the test chip sources in the test chip source subset corresponding to each preset video technical indicator to each video codec in the video codec library. Performing a codec performance test evaluation to obtain a typical video codec corresponding to the first preset condition corresponding to each preset video technical indicator;

The step C includes:

C1, in turn, in the typical video codec corresponding to each preset video technical indicator, all the test film sources in the test chip source subset corresponding to the corresponding preset video technical indicators are for each of all the selected hardware platforms. Selecting a hardware platform for performing a codec performance test evaluation to obtain a codec performance test value corresponding to a corresponding preset video technical indicator of each candidate hardware platform;

C2. Set weight values for two or more codec performance test values corresponding to two or more preset video technical indicators of each candidate hardware platform according to preset requirements, and obtain each candidate hardware in a weighted average manner. The integrated value of the codec performance test corresponding to the two or more preset video technical indicators of the platform, thereby obtaining the candidate hardware platform whose codec performance test comprehensive value is greater than the preset comprehensive value.
The method for selecting a video codec hardware platform according to claim 2, wherein the preset video technical indicator is one or more, when the preset video technical indicator is two or more;

The step B includes: sequentially, on the preset hardware platform, all the test chip sources in the test chip source subset corresponding to each preset video technical indicator to each video codec in the video codec library. Performing a codec performance test evaluation to obtain a typical video codec corresponding to the first preset condition corresponding to each preset video technical indicator;

The step C includes:

C1, in turn, in the typical video codec corresponding to each preset video technical indicator, all the test film sources in the test chip source subset corresponding to the corresponding preset video technical indicators are for each of all the selected hardware platforms. Selecting a hardware platform for performing a codec performance test evaluation to obtain a codec performance test value corresponding to a corresponding preset video technical indicator of each candidate hardware platform;

C2. Set weight values for two or more codec performance test values corresponding to two or more preset video technical indicators of each candidate hardware platform according to preset requirements, and obtain each candidate hardware in a weighted average manner. The integrated value of the codec performance test corresponding to the two or more preset video technical indicators of the platform, thereby obtaining the candidate hardware platform whose codec performance test comprehensive value is greater than the preset comprehensive value.
The method of selecting a video codec hardware platform according to claim 1, wherein the step B specifically includes:

B1. All test source sources in the test slice source subset corresponding to the preset video technical indicators are classified according to a preset hierarchical rule, and are divided into different levels of secondary test piece source subsets corresponding to the preset hierarchical rules;

B2. On the preset hardware platform, all test chip sources in the secondary test chip source subset of the preset level are subjected to codec performance test evaluation for each video codec in the video codec library, and The codec performance test evaluation value is used as the initial clustering center of the K-means algorithm;

B3. On the preset hardware platform, all the test film sources except the secondary test piece source subset of the preset level in the test chip source subset corresponding to the preset video technical indicators are used for each of the video coding and decoding programs. A video codec performs a codec performance test evaluation, and the obtained codec performance test evaluation value is used as a clustering basis of the K-means algorithm;

B4. The K-means algorithm performs processing to sort all the video codecs in the video codec library according to the performance and the inferiority, and selects the video codec in the front row to be matched with the preset video technical indicators. Typical video codec.
The method of selecting a video codec hardware platform according to claim 2, wherein the step B specifically includes:

B1. All test source sources in the test slice source subset corresponding to the preset video technical indicators are classified according to a preset hierarchical rule, and are divided into different levels of secondary test piece source subsets corresponding to the preset hierarchical rules;

B2. On the preset hardware platform, all test chip sources in the secondary test chip source subset of the preset level are subjected to codec performance test evaluation for each video codec in the video codec library, and The codec performance test evaluation value is used as the initial clustering center of the K-means algorithm;

B3. On the preset hardware platform, all the test film sources except the secondary test piece source subset of the preset level in the test chip source subset corresponding to the preset video technical indicators are used for each of the video coding and decoding programs. A video codec performs a codec performance test evaluation, and the obtained codec performance test evaluation value is used as a clustering basis of the K-means algorithm;

B4. The K-means algorithm performs processing to sort all the video codecs in the video codec library according to the performance and the inferiority, and selects the video codec in the front row to be matched with the preset video technical indicators. Typical video codec.
The method of selecting a video codec hardware platform according to claim 1, wherein the step A further comprises:

D. Retrieve the newly released test film source, update the newly released test film source to the test film source library, and retrieve the newly released video codec and update the newly released video codec to the video. Codec library.
The method for selecting a video codec hardware platform according to claim 7, wherein the step D further comprises: deleting a video codec that is too low in the video codec library.
An apparatus for selecting a video codec hardware platform, wherein the apparatus for selecting a video codec hardware platform comprises:

The source source classification module is configured to classify all test film sources in the test source library according to a preset classification rule, and divide into different test chip source subsets corresponding to different video technical indicators;

The typical program evaluation module is configured to code and decode each video codec in the video codec library in the test chip source subset corresponding to the preset video technical indicator on the preset hardware platform. Testing and evaluating to obtain a typical video codec corresponding to the first preset condition corresponding to the preset video technical indicator;

The hardware evaluation module is configured to code and decode all the test chip sources in the test chip source subset corresponding to the preset video technical indicators on each of the candidate hardware platforms in the candidate video codec. The test is evaluated to obtain a candidate hardware platform whose performance test comprehensive value satisfies the second preset condition.
The apparatus for selecting a video codec hardware platform according to claim 9, wherein the process of evaluating the codec performance test is: placing a corresponding test piece source on a corresponding video codec on a corresponding hardware platform. Play, comprehensively evaluate the quality of the subjective visual quality, objective fidelity and color of the played image to obtain the codec performance evaluation value of the corresponding video codec or the corresponding candidate hardware platform.
The device for selecting a video codec hardware platform according to claim 9, wherein the preset video technical indicator is one or more, when the preset video technical indicator is two or more;

The typical program evaluation module is configured to sequentially encode and decode each video source in the video codec library in the test chip source subset corresponding to each preset video technical indicator on a preset hardware platform. The program performs a codec performance test evaluation to obtain a typical video codec program corresponding to the first preset condition corresponding to each preset video technical indicator;

The hardware evaluation module is configured to sequentially, in a typical video codec corresponding to each preset video technical indicator, all test source sources in the test chip source subset corresponding to the corresponding preset video technical indicators to all the selected hardware platforms. Each candidate hardware platform performs a codec performance test evaluation to obtain a codec performance test value corresponding to the corresponding preset video technical indicator of each candidate hardware platform;

The hardware evaluation module is further configured to set a weight value for each of the two or more codec performance test values corresponding to the two or more preset video technical indicators of each candidate hardware platform according to a preset requirement, and perform weighted average manner. Obtaining a comprehensive codec performance test value corresponding to two or more preset video technical indicators of each candidate hardware platform, thereby obtaining a candidate hardware platform whose codec performance test comprehensive value is greater than a preset comprehensive value.
The device for selecting a video codec hardware platform according to claim 10, wherein the preset video technical indicator is one or more, when the preset video technical indicator is two or more;

The typical program evaluation module is configured to sequentially encode and decode each video source in the video codec library in the test chip source subset corresponding to each preset video technical indicator on a preset hardware platform. The program performs a codec performance test evaluation to obtain a typical video codec program corresponding to the first preset condition corresponding to each preset video technical indicator;

The hardware evaluation module is configured to sequentially, in a typical video codec corresponding to each preset video technical indicator, all test source sources in the test chip source subset corresponding to the corresponding preset video technical indicators to all the selected hardware platforms. Each candidate hardware platform performs a codec performance test evaluation to obtain a codec performance test value corresponding to the corresponding preset video technical indicator of each candidate hardware platform;

The hardware evaluation module is further configured to set a weight value for each of the two or more codec performance test values corresponding to the two or more preset video technical indicators of each candidate hardware platform according to a preset requirement, and perform weighted average manner. Obtaining a comprehensive codec performance test value corresponding to two or more preset video technical indicators of each candidate hardware platform, thereby obtaining a candidate hardware platform whose codec performance test comprehensive value is greater than a preset comprehensive value.
The apparatus for selecting a video codec hardware platform according to claim 9, wherein:

The slice source classification module is further configured to classify all test film sources in the test slice source subset corresponding to the preset video technical indicators according to a preset hierarchical rule, and divide into two different tests corresponding to the preset hierarchical rules. Source subset

The typical program evaluation module includes an initial cluster center acquisition unit, a cluster basis acquisition unit, and a sort unit:

The initial clustering center acquiring unit is configured to compile each video codec in the video codec library in the preset level of the secondary test chip source subset on the preset hardware platform. Decoding performance test evaluation, and the obtained codec performance test evaluation value is used as an initial clustering center of the K-means algorithm;

The clustering obtaining unit is configured to: on the preset hardware platform, all test film sources except a subset of the secondary test piece source of the preset level in the test chip source subset corresponding to the preset video technical index Each video codec in the codec performs a codec performance test evaluation, and the obtained codec performance test evaluation value is used as a clustering basis of the K-means algorithm;

a sorting unit for sorting all video codecs in the video codec library by performance calculation by K-means algorithm, and selecting a video codec located in the top of the sorting as a preset video technology A typical video codec corresponding to the indicator.
The apparatus for selecting a video codec hardware platform according to claim 10, wherein:

The slice source classification module is further configured to classify all test film sources in the test slice source subset corresponding to the preset video technical indicators according to a preset hierarchical rule, and divide into two different tests corresponding to the preset hierarchical rules. Source subset

The typical program evaluation module includes an initial cluster center acquisition unit, a cluster basis acquisition unit, and a sort unit:

The initial clustering center acquiring unit is configured to compile each video codec in the video codec library in the preset level of the secondary test chip source subset on the preset hardware platform. Decoding performance test evaluation, and the obtained codec performance test evaluation value is used as an initial clustering center of the K-means algorithm;

The clustering obtaining unit is configured to: on the preset hardware platform, all test film sources except a subset of the secondary test piece source of the preset level in the test chip source subset corresponding to the preset video technical index Each video codec in the codec performs a codec performance test evaluation, and the obtained codec performance test evaluation value is used as a clustering basis of the K-means algorithm;

a sorting unit for sorting all video codecs in the video codec library by performance calculation by K-means algorithm, and selecting a video codec located in the top of the sorting as a preset video technology A typical video codec corresponding to the indicator.
The device for selecting a video codec hardware platform according to claim 9, wherein the device for selecting a video codec hardware platform further comprises an update module, wherein the update module is configured to retrieve and obtain a newly released test film source, and The newly released test piece source is updated into the test piece source library, and used to retrieve the newly released video codec and update the newly released video codec to the video codec library.
The apparatus for selecting a video codec hardware platform according to claim 15, wherein the update module is further configured to delete a video codec that is too low in the video codec library.