Bullet screen comment data-based movie and television clip editing method and equipment thereof
Through the film and television clip editing method based on barrage comment data, the problems in the prior art are solved, which are difficult to crawl, high computing resources, and manual participation in selection and adjustment, and the automation and efficiency of the editing process are realized, and system cost and energy consumption are reduced.
Patent Information
- Application Number
- CN202411641231.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-18
- Publication Date
- 2025-02-28
AI Technical Summary
When handling large-scale video editing and editing, the existing technology faces problems such as high difficulty in data crawling, high computing resources consumption, manual participation in selection and adjustment, inability to fully automate, and high cost and energy consumption.
The film and television clip editing method based on barrage comment data is adopted, and the video comment data is obtained through the data acquisition module, and feature extraction and word frequency statistics are used using Spark MLlib, and editing and production are combined with a nonlinear editing system and a high-performance processor.
It realizes the flexibility and efficiency of data crawling, reduces the consumption of computing resources, realizes the automation of the editing process, reduces system costs and energy consumption, and improves the efficiency and quality of video editing.
Smart Images
Figure CN119545102A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of multimedia information processing, and in particular to a method and device for editing film and television clips based on bullet screen comment data. Background Art
[0002] In the production of large-scale video projects such as movies and TV series, more professional workstations and servers may be used to analyze bullet screen data and edit video clips. With the development of cloud computing technology, more and more video editing and editing work can be completed through cloud services. Bullet screen data can be analyzed in the cloud, and editing results can also be stored in the cloud, which is convenient for multi-person collaboration and remote access.
[0003] Although it is possible to break through the access restrictions of websites by setting proxy IP in the prior art, some websites may adopt more complex anti-crawler mechanisms, such as dynamic content loading, verification codes, cookies checking, JavaScript rendering, etc., which will increase the difficulty of data crawling. Frequent crawling requests may put pressure on the server, which may be identified as malicious behavior by the server and cause the IP to be blocked. Storing a large amount of data directly in the database may cause slower query and processing speeds, especially when processing large data sets. When using Spark MLlib for feature extraction and word frequency statistics, if the data volume is very large or the feature dimension is very high, a large amount of computing resources may be consumed and the processing time may increase. TF-IDF is a simple word frequency counting method, which ignores the semantic relationship and contextual information between words, and may cause some semantically similar words to be assigned different weights. For text data in certain specific fields, TF-IDF may not be able to effectively extract key information. Although a variety of editing techniques and methods are provided, most of them still require manual participation for selection and adjustment, and it is difficult to achieve a fully automated editing process. For complex video materials, such as special effects, sound effects, subtitles, etc., the existing editing system may not be able to fully meet the needs. Real-time processing of high-quality video materials requires high-performance processors and graphics processors, which may increase the cost and energy consumption of the system. For some small businesses or individual users, purchasing and maintaining high-performance hardware may be a burden. To this end, we propose a film and television clip editing method based on barrage comment data and its device to solve the above problems. Summary of the invention
[0004] The purpose of the present invention is to provide a method and device for editing film and television clips based on barrage comment data to solve the problems raised in the above-mentioned background technology.
[0005] The technical solution of the present invention is: a method and device for editing film and television clips based on barrage comment data, including a data acquisition module, the acquisition module is provided with an API interface, the API interface data is connected to a data crawling module, the data crawling module data is connected to a video ranking module, the data acquisition module signal is connected to a display, the video ranking module is provided with a developer mode, the data acquisition module signal is connected to a music section, the video ranking module signal is connected to a crawled video comment content module, and the crawled video comment content module data is connected to a database.
[0006] Preferably, the module for crawling video comments consists of defining video sections and traversing sections.
[0007] Preferably, the music section is connected to the video chart section data, and the data crawling section is connected to the music section data.
[0008] Preferably, the data summary module is data-connected to the data acquisition module.
[0009] Preferably, the data acquisition module is connected to database data.
[0010] Preferably, step S1: the data crawling interface can be obtained through the open API interface of the video, open the video ranking module, select any section to enter the developer mode, and directly crawl the corresponding data. To crawl the music section data, just set rid=3 in the following URL to obtain all JSON format video data under the music section; Step S2: By setting the proxy IP (Internet Protocol), you can break through the restrictions on visitors imposed by the website and pass it as a parameter into the requests to achieve data crawling; Step S3: by defining the video section and the traversal section, a request is made to the website to obtain data and store the data in the database; Step S4: Use the database to implement comment classification as follows: words=jieba.cut(strings,cut_all=False); Step S5: SparkMLlib uses the hashing function to implement word frequency statistics for the word features. However, different word features may get the same hash value and cause hash conflicts. The conflicts can be avoided by increasing the number of buckets in the hash table. Step S6: Call the transform of HasdingTf to hash the terms into feature vectors. Assume that the number of buckets in the hash table is 2000. The steps are as follows: hashingTF=HashingTF(inputCol="words",outputCol="rawFeatures",numFeatures=2000)featurizeData=hashingTF.transform(wordsData); Step S7: Call IDF to modify the word frequency feature vector to improve the ability of different words to distinguish texts. The steps are as follows: idf=IDF(inputCol="rawFeatures",outputCol="features")idfModel=idf.fit(featurizeData); Step S8: Call transform to get the TFIDF metric value corresponding to each word. The steps are as follows: rescaledData=idfModel.transform(featurizeData); Step S9: Review similarity is achieved by normalizing the TF-IDF value to construct a review similarity weight vector. The steps are as follows: normalizer=Normalizer(inputCol="features",outputCol="norm")data=normalizer.transform(rescaledData); Step S10: Using a non-linear editing system equipped with a high-performance processor and a graphics processor to process high-quality video materials in real time, and completing efficient editing and production work by using the non-linear editing system in conjunction with video editing software; Step S11: Action sequential editing is performed by switching the camera while the character is still moving. The editing point can be switched according to simple camera settings such as the direction of the action or the character turning around; Step S12: The off-cut editing first cuts to the insert shot through the screen, and then inserts it back to the main shot. The insert shot can be a shot in the same space as the protagonist, used to assist the plot development or express the inner world of the character; Step S13: Cross-cutting is a process of switching between two different scenes to establish the interaction between characters or to show the inner emotional entanglement of the characters; Step S14: Jump cut editing is performed by editing the same shot, usually used to express the passage of time or to increase the urgency of the shot; Step S15: Jump cut editing is performed by editing the same shot, which is usually used to express the passage of time or to increase the urgency of the shot; Step S16: Matching the two linked shots usually has the same action or similar composition, which is used to achieve a smooth transition effect; Step S17: Fade-in / Fade-out clips enter or fade out from a completely black screen through lens blur, which is often used for transition effects; Step S18: The superimposed clip is superimposed from one shot to another shot to form a mixed visual effect; Step S19: Jump editing is a sudden editing method used to break the mood or atmosphere of the previous scene and switch from a quiet scene to an intense scene; Step S20: Hiding the editing points by using shadows or camera movements to achieve seamless editing; Step S21: Split and cut a shot into several segments and use them in different places.
[0011] The present invention provides a method and device for editing film and television clips based on bullet screen comment data through improvement, which has the following improvements and advantages compared with the prior art: First: The present invention uses visualization to effectively express data that people cannot understand through graphical means, which can accurately, efficiently, concisely and comprehensively convey certain information, help people analyze and reason about data, explore the value behind the data, and make complex data easier to understand and use.
[0012] Second: After the present invention cuts the footage into multiple segments, the segments can be freely recombined, adjusted and arranged as needed during the editing process to create different rhythms, emotions and visual effects. Cutting can help editors accurately control the rhythm of the video. By inserting or deleting certain segments at different time points, the rhythm of the video can be accelerated or slowed down to adapt to specific narrative needs or the emotions of the audience. Through cutting, important footage segments can be placed in a more prominent position, or a certain scene or information point can be emphasized by repeatedly playing a certain segment. If a shot is too long or contains unnecessary parts, these redundant parts can be removed through cutting to make the video more compact and efficient. By combining different footage segments together, new visual effects and narrative methods can be created. This innovative editing technique can attract the audience's attention and improve the video viewing experience. The screen size, resolution and playback speed of different platforms and devices may be different. By cutting the footage into multiple segments and adjusting and optimizing them according to the characteristics of different platforms, you can ensure that the video can present the best effect on different platforms. If the video needs to be modified or adjusted later, the cut footage segments can be more easily replaced, added or deleted, thus saving time and cost. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] The present invention will be further explained below in conjunction with the accompanying drawings and embodiments: Figure 1 It is a schematic diagram of a specific implementation process of the present invention; Figure 2 It is a schematic diagram of a scatter plot of video comments of the present invention. DETAILED DESCRIPTION
[0014] The present invention is described in detail below, and the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0015] The present invention provides a method and device for editing film and television clips based on bullet screen comment data through improvement. The technical solution of the present invention is: like Figure 1 - Figure 2As shown, a method for editing film and television clips based on barrage comment data and a device thereof include a data acquisition module, the acquisition module is provided with an API interface, the API interface data is connected to a data crawling module, the data crawling module data is connected to a video ranking module, the data acquisition module signal is connected to a display, the video ranking module is provided with a developer mode, the data acquisition module signal is connected to a music section, the video ranking module signal is connected to a crawling video comment content module, the crawling video comment content module data is connected to a database, through the API interface, the data acquisition module can efficiently obtain data from multiple sources, such as websites, API services, etc., the data crawling module further enhances the flexibility and ability of data acquisition, especially for those websites that do not provide API or have more API restrictions, the video ranking module can update the most popular videos in real time, and display them to users through the display to provide timely information, and the developer mode allows developers to more flexibly adjust and optimize the video ranking algorithm and display method. In addition to the video overall list, the system is also connected to the music section, which can provide users with a variety of content choices. This diversity can attract users with different interests and improve user satisfaction and activity. The module for crawling video comment content can collect and analyze users' evaluation and feedback on videos, and provide data support for content recommendation and improvement. This user feedback is crucial for optimizing content and improving user experience. The database is used to store crawled video comment content and other important data to ensure data security and persistence. The database also supports efficient data query and analysis, providing a foundation for subsequent data mining and machine learning. The modular design makes the system easier to expand and maintain. When you need to add new features or modify existing features, you only need to focus on specific modules. This design can also improve the stability and reliability of the system. By analyzing user feedback and behavior data in the database, the system can build an intelligent recommendation algorithm to recommend content that is more in line with users' interests.
[0016] Furthermore, the module for crawling video comments consists of defining a video section and a traversal section. The modular design makes the code clearer, easier to understand and maintain. Different functions are encapsulated in independent modules, reducing the coupling of the code. Once the video section is defined, the code of the section can be used multiple times for different crawling tasks. Only the relevant parameters of the traversal section need to be modified. This improves the efficiency of the code and reduces duplication of work. The traversal section can be flexibly adjusted as needed, such as changing the order of traversal, adding filtering conditions, etc. The code for defining the video section is usually fixed, while the traversal section can be customized according to specific needs. When defining the video section, unified error handling can be performed for possible errors, such as network request failure, page structure changes, etc., which reduces the time required to handle these errors in the traversal section. In order to reduce the complexity of errors, a cache mechanism can be added to the traversal section to avoid repeated crawling of data that has already been obtained. The traversal strategy can also be optimized, such as using multi-threading, asynchronous IO and other technologies to increase the crawling speed. Independent modules are easier to unit test or integration test. You can test separately whether the code that defines the video section correctly parses the page structure, or test whether the traversal section traverses all video comments as expected. When you need to crawl a new video platform or a new video type, you only need to add new code to define the video section without modifying the code of the traversal section. This makes the system easier to expand and maintain. When the page structure changes, you only need to modify the code that defines the video section without modifying the code of the entire crawling module, which reduces maintenance costs and improves system stability.
[0017] Furthermore, the music section is connected with the video chart section data, and the data crawling section is connected with the music section data. After the music section and the video chart section are connected, the system can recommend music and video content that is more in line with the user's preferences based on the music usage in the video and user feedback. Through comprehensive analysis of video and music data, the system can predict trends and recommend popular content to users in advance. When a user shows interest in specific content (such as a singer or style of music) in one section (such as the music section), the system can recommend videos related to the content in another section (such as the video chart section), thereby increasing the user's browsing time and engagement. This cross-recommendation helps to increase user stickiness and improve the overall usage rate of the platform. Data connection allows the platform to conduct in-depth analysis of popular trends and user behaviors in music and videos, and provide data support for content production, marketing strategies, etc. Obtaining external data through the data crawling section can further enrich the platform's data set and enhance the accuracy and reliability of data analysis. Based on the connected data, the platform can better Accurately understand user needs and preferences, thereby guiding content creators to produce content that is more in line with market demand. At the same time, data connection can also help the platform optimize content distribution strategies to ensure that content can reach the users who are most likely to be interested. Through accurate content recommendations and personalized interface displays, the platform can provide users with more intimate and personalized services, thereby improving user experience. When users browse and interact on the platform, they can feel the platform's in-depth understanding and rapid response to their needs, which will enhance users' trust and loyalty to the platform. Data connection can help the platform monitor and identify potential risks and violations in real time, such as copyright issues, malicious content, etc. Through continuous monitoring and analysis of data, the platform can take timely measures to prevent risks from occurring and maintain the health and safety of the platform. Data connection allows users to switch and interact more naturally between music and video content, enhancing users' social experience. Users can share related video links in the music comment area, or discuss background music in the video comment area, thereby increasing interaction and communication between users.
[0018] Furthermore, the data summary module is connected to the data acquisition module. The data acquisition module can automatically collect, extract and store data from various data sources, including databases, APIs, Web pages, etc. This automated method can greatly improve the efficiency of data acquisition and reduce the waste of human resources. Through the connection between the data acquisition module and the data summary module, real-time updating and synchronization of data can be achieved to ensure that the data displayed in the data summary module is always the latest. In the process of collecting data, the data acquisition module can perform data cleaning and correction, remove duplicate data, delete invalid data, fill in missing values, etc., thereby improving the quality and accuracy of the data. This high-quality data can provide more accurate and reliable data support for the data summary module, enabling it to more accurately display the trends and changes of various data. The data connection between the data summary module and the data acquisition module can make the data analysis process smoother and more efficient. The data summary module can directly obtain the data collected by the collection module without the need for additional data import and sorting. At the same time, since the data collection module can obtain a large amount of data, these data can provide richer materials and basis for data analysis, making the data analysis results more comprehensive and in-depth. The data summary module can provide strong support for the decision-making of enterprises or organizations by displaying the trends and changes of various data. The data connection between the data collection module and the data summary module can ensure that the data required for decision-making can be collected and displayed in the data summary module in a timely and accurate manner. This timely and accurate data support can help enterprises or organizations better understand market trends, customer needs, competitive situations and other information, so as to make more informed and effective decisions. The large amount of data obtained through the data collection module can help enterprises or organizations discover new business opportunities and innovations. For example, by analyzing user needs and behavioral characteristics, products or services that are more in line with market demand can be developed. At the same time, the data summary module can intuitively display the performance of these innovative achievements in the market, thereby providing strong support for the innovative development of enterprises or organizations.
[0019] Furthermore, the data acquisition module is connected to the database data. The data acquisition module is directly connected to the database, avoiding the middle-layer conversion and reducing the network transmission time and resource consumption. Therefore, the data transmission and processing speed is faster. By writing SQL statements, the data acquisition module can freely control the data content and format to be obtained, which greatly improves the flexibility of data processing. If a new data source needs to be added, it only needs to modify the SQL statement without making major changes to the entire system. It has good scalability and can obtain more original and real data by directly connecting to the database, avoiding possible errors in the middle layer and ensuring the accuracy of the data. The data acquisition module can perform data cleaning during the data extraction process, such as removing duplicate data, deleting invalid data, and filling missing values. Etc., to improve the quality and accuracy of data. The data acquisition module can be connected to a variety of different data sources, including databases, APIs, Web pages, files, etc., providing rich data support for business applications. The data acquisition module can realize automatic data collection and updating. By setting scheduled tasks or event trigger mechanisms, data can be automatically collected regularly or in real time, saving manpower and time costs and improving data acquisition efficiency. Through the connection between the data acquisition module and the database, enterprises can obtain and analyze business data in real time, provide data support for decision-making, and promote business development and innovation. Although there may be security risks in directly connecting to the database, the risk of data leakage can be effectively reduced by implementing strict security management measures, such as restricting access rights and encrypting transmission.
[0020] Further, step S1: the data crawling interface can be obtained through the open API interface of the video. Open the video ranking module, select any section to enter the developer mode, and you can directly crawl the corresponding data. To crawl the music section data, just set rid=3 in the following URL to obtain all the JSON format video data under the music section; Step S2: By setting the proxy IP (Internet Protocol), you can break through the restrictions on visitors imposed by the website and pass it as a parameter into the requests to achieve data crawling; Step S3: by defining the video section and the traversal section, a request is made to the website to obtain data and store the data in the database; Step S4: Use the database to implement comment classification as follows: words=jieba.cut(strings,cut_all=False); Step S5: SparkMLlib uses the hashing function to implement word frequency statistics for the word features. However, different word features may get the same hash value and cause hash conflicts. The conflicts can be avoided by increasing the number of buckets in the hash table. Step S6: Call the transform of HasdingTf to hash the terms into feature vectors. Assume that the number of buckets in the hash table is 2000. The steps are as follows: hashingTF=HashingTF(inputCol="words",outputCol="rawFeatures",numFeatures=2000)featurizeData=hashingTF.transform(wordsData); Step S7: Call IDF to modify the word frequency feature vector to improve the ability of different words to distinguish texts. The steps are as follows: idf=IDF(inputCol="rawFeatures",outputCol="features")idfModel=idf.fit(featurizeData); Step S8: Call transform to get the TFIDF metric value corresponding to each word. The steps are as follows: rescaledData=idfModel.transform(featurizeData); Step S9: Review similarity is achieved by normalizing the TF-IDF value to construct a review similarity weight vector. The steps are as follows: normalizer=Normalizer(inputCol="features",outputCol="norm")data=normalizer.transform(rescaledData); Step S10: Using a non-linear editing system equipped with a high-performance processor and a graphics processor to process high-quality video materials in real time, and completing efficient editing and production work by using the non-linear editing system in conjunction with video editing software; Step S11: Action sequential editing is performed by switching the camera while the character is still moving. The editing point can be switched according to simple camera settings such as the direction of the action or the character turning around; Step S12: The off-cut editing first cuts to the insert shot through the screen, and then inserts it back to the main shot. The insert shot can be a shot in the same space as the protagonist, used to assist the plot development or express the inner world of the character; Step S13: Cross-cutting is a process of switching between two different scenes to establish the interaction between characters or to show the inner emotional entanglement of the characters; Step S14: Jump cut editing is performed by editing the same shot, usually used to express the passage of time or to increase the urgency of the shot; Step S15: Jump cut editing is performed by editing the same shot, which is usually used to express the passage of time or to increase the urgency of the shot; Step S16: Matching the two linked shots usually has the same action or similar composition, which is used to achieve a smooth transition effect; Step S17: Fade-in / Fade-out clips enter or fade out from a completely black screen through lens blur, which is often used for transition effects; Step S18: The superimposed clip is superimposed from one shot to another shot to form a mixed visual effect; Step S19: Jump editing is a sudden editing method used to break the mood or atmosphere of the previous scene and switch from a quiet scene to an intense scene; Step S20: Hiding the editing points by using shadows or camera movements to achieve seamless editing; Step S21: Cut a shot into several segments and use them in different places. Draw a scatter plot of video comments using the June data as an example. Figure 2 As shown, from Figure 2 It can be seen that, on the one hand, the similarity of most comments is low. Although the content that users pay attention to is similar, young people do not blindly follow the trend in commenting, but they do not use similar language as comments, which reflects the high degree of differentiation of user language expression. On the other hand, young people are easy to accept new things. Anyone can access other people’s contributed videos online for secondary processing, and express themselves through language output of opinions. They have a strong individualistic color, which makes the society full of expectations for them.
[0021] The above description enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method and device for editing film and television clips based on bullet screen comment data, comprising a data acquisition module, characterized in that: The acquisition module is provided with an API interface, the API interface data is connected to a data crawling module, the data crawling module data is connected to a video ranking module, the data acquisition module signal is connected to a display, the video ranking module is provided with a developer mode, the data acquisition module signal is connected to a music section, the video ranking module signal is connected to a crawling video comment content module, and the crawling video comment content module data is connected to a database.
2. A method and device for editing film and television clips based on bullet screen comment data according to claim 1, characterized in that: The module for crawling video comments consists of defining a video section and a traversal section.
3. A method and device for editing film and television clips based on bullet screen comment data according to claim 1, characterized in that: The music section is connected to the video chart section data, and the data crawling section is connected to the music section data.
4. A method and device for editing film and television clips based on bullet screen comment data according to claim 1, characterized in that: The data summary module is data-connected to the data acquisition module.
5. A method and device for editing film and television clips based on bullet screen comment data according to claim 1, characterized in that: The data acquisition module is connected with the database data.
6. A method and device for editing film and television clips based on bullet screen comment data according to claim 1, characterized in that: Step S1: The data crawling interface can be obtained through the open API interface of the video. Open the video ranking module, select any section to enter the developer mode, and then directly crawl the corresponding data. To crawl the music section data, just set rid=3 in the following URL to obtain all the JSON format video data under the music section; Step S2: By setting the proxy IP (Internet Protocol), you can break through the restrictions on visitors imposed by the website and pass it as a parameter into the requests to achieve data crawling; Step S3: by defining the video section and the traversal section, a request is made to the website to obtain data and store the data in the database; Step S4: Use the database to implement comment classification as follows: words=jieba.cut(strings,cut_all=False); Step S5: SparkMLlib uses the hashing function to implement word frequency statistics for the word features. However, different word features may get the same hash value and cause hash conflicts. The conflicts can be avoided by increasing the number of buckets in the hash table. Step S6: Call the transform of HasdingTf to hash the terms into feature vectors. Assume that the number of buckets in the hash table is 2000. The steps are as follows: hashingTF=HashingTF(inputCol="words",outputCol="rawFeatures",numFeatures=2000)featurizeData=hashingTF.transform(wordsData); Step S7: Call IDF to modify the word frequency feature vector to improve the ability of different words to distinguish texts. The steps are as follows: idf=IDF(inputCol="rawFeatures",outputCol="features")idfModel=idf.fit(featurizeData); Step S8: Call transform to get the TFIDF metric value corresponding to each word. The steps are as follows: rescaledData=idfModel.transform(featurizeData); Step S9: Review similarity is achieved by normalizing the TF-IDF value to construct a review similarity weight vector. The steps are as follows: normalizer=Normalizer(inputCol="features",outputCol="norm")data=normalizer.transform(rescaledData); Step S10: Using a non-linear editing system equipped with a high-performance processor and a graphics processor to process high-quality video materials in real time, and completing efficient editing and production work by using the non-linear editing system in conjunction with video editing software; Step S11: Action sequential editing is performed by switching the camera while the character is still moving. The editing point can be switched according to simple camera settings such as the direction of the action or the character turning around; Step S12: The off-cut editing first cuts to the insert shot through the screen, and then inserts it back to the main shot. The insert shot can be a shot in the same space as the protagonist, used to assist the plot development or express the inner world of the character; Step S13: Cross-cutting is a process of switching between two different scenes to establish the interaction between characters or to show the inner emotional entanglement of the characters; Step S14: Jump cut editing is performed by editing the same shot, usually used to express the passage of time or to increase the urgency of the shot; Step S15: Jump cut editing is performed by editing the same shot, which is usually used to express the passage of time or to increase the urgency of the shot; Step S16: Matching the two linked shots usually has the same action or similar composition, which is used to achieve a smooth transition effect; Step S17: Fade-in / Fade-out clips enter or fade out from a completely black screen through lens blur, which is often used for transition effects; Step S18: The superimposed clip is superimposed from one shot to another shot to form a mixed visual effect; Step S19: Jump editing is a sudden editing method used to break the mood or atmosphere of the previous scene and switch from a quiet scene to an intense scene; Step S20: Hiding the editing points by using shadows or camera movements to achieve seamless editing; Step S21: Split and cut a shot into several segments and use them in different places.