** Introduction: ** Video cloud drives intelligence media innovation
On May 27th, CCBN (the 28th China International Radio and Television Information Network Exhibition) was held in Beijing. In the selection of the “CCBN Annual Innovation Award”, Ali Cloud video cloud won the CCBN product Innovation Outstanding award by virtue of AI editorial department’s media industry professional solutions.
With the development of technology, the innovation and application of 5G, 4K, AI and other new technologies has become the consensus for the transformation of the media industry. In the face of 986 million mobile Internet users, it has become a development trend of the media industry to give full play to the advantages of 5G network technology and build an all-media distribution system integrating small screen of mobile phone and large screen of TV. In the new form of communication, users are more inclined to personalized audio-visual needs, which brings new requirements and challenges to the media industry content production efficiency and production output. How to use artificial intelligence and data technology to improve the production efficiency and production output of video has become one of the key work of AI editorial department’s media fusion transformation.
Production is the core of the entire production and broadcasting process, which connects collection, broadcasting and storage. Only after the production process is fully cloud, can the efficiency of converged media be truly improved. In the whole process of fusion media production and broadcasting, the newly upgraded AI Editorial Department 2.0 integrates short video production and high-bit-rate precision editing, and realizes two-way interaction and fusion production of large and small screens through extensive use of AI technology. At the same time, the integration of cloud and end collaborative production mode, accelerate the efficiency of the whole content production.
In terms of practical operation, the AI editorial department achieves this through three subdivisions: cloud production of true 4K, 4K video enhancement and short video production of fusion media on cloud:
First of all, in terms of the infrastructure of 4K cloud production, the major innovation made by AI editorial department is the introduction of SCC high performance computing and CPFS parallel file system. To meet the high performance requirements of cloud production, low-latency interconnection between compute nodes and storage nodes is implemented to provide better production experience, avoid virtualization loss, and further improve performance. To meet high storage capacity and bandwidth requirements, CPFS provides linear throughput expansion along with storage nodes, up to 300GB/s for a single cluster. The AI editorial department provides rich resources suitable for media business, including various personalized hardware devices such as FPGA and GPU, to meet different business demands. In addition, the AI editorial department has carried out a lot of optimization for GPU, including GPU virtualization, sharing and decentralization, which can greatly improve the utilization rate of GPU and reduce the cost of cloud production.
Secondly, in terms of 4K video enhancement, the AI editorial department mainly improves the picture quality of 1080p and lower-quality videos through the three core technologies of video supersegmentation, video frame insertion and color enhancement, and meets the requirements of ULTRA HD video in frame rate, resolution and color. The 4K enhanced capability provided by the AI editorial Department not only meets the professional standards, that is, the standard definition of the GENERAL Administration in the implementation Guide of 4K Ultra HD TV Program Production Technology, but also meets the requirements of professional technical review and supervision in the industry. It not only meets the objective parameter standards, but also significantly improves subjective visual experience. It includes detail restoration, noise suppression, picture smoothness, video brightness, color contrast and light and dark detail preservation. In terms of AI video super-segmentation, the AI editorial department adopts deep learning to conduct lossless amplification of the picture, which can achieve a 4-fold increase in resolution. While repairing and restoring details such as portrait and scene, it can suppress block noise, compression noise and other image quality damage to make the picture sharper. AI editorial department innovatively proposed the enhancement of key parts to achieve the most effect at the minimum cost, such as focusing on the optimization of portrait face and realizing better restoration of human skin texture and hair details.
Finally, in terms of short video production of cloud fusion media, the AI editorial department has the following five core advantages based on the current short video hot spots and user preferences:
1. AI editorial department provides powerful prediction and evaluation capabilities to help editors decide what kind of short videos to make and make targeted productions.
2. The AI editorial department provides powerful multi-mode tag capability to redefine the content of videos. It makes comprehensive use of image, voice, text and other AI capabilities to systematically generate tags.
3. In view of the pain points of retrieval, AI editorial department provides powerful multi-mode retrieval, including text search, picture search, video search and support for natural speech understanding, knowledge map and other full-process multi-mode retrieval capabilities.
4. Provide massive and multi-category video templates to produce a large number of professional short videos in a short time.
5. The AI editorial department provides rich and professional intelligent auxiliary tools, which can meet the needs of rapid short video production in a scene.
Since the release of AI Editorial Department 2.0, the Technology Department of People’s Daily has cooperated with Aliyun Video cloud to apply this ability to the content production of the TWO sessions. The AI technology of Aliyun is adopted, and after the massive data training of People’s Daily, the AI application innovation is carried out in combination with the business scenes of People’s Daily. AI editorial department “can see, listen, understand and look up”, can realize the automation, mass and intelligent production of all kinds of new media products, achieve multi-mode retrieval, intelligent clearance, intelligent video production and other functions, greatly improve the efficiency of news production, and comprehensively help People’s Daily TWO sessions report. After continuous iteration and improvement, on the eve of the two sessions in 2021, Ali Cloud video cloud for the People’s Daily AI intelligence editorial department, comprehensive new upgrade of the function, integration of five new AI capabilities, to help improve the quality and efficiency of all media reports on the two sessions.
In the past, AI editorial department radio and Television used professional equipment locally to produce HD video.
Today, AI newsrooms use IP, intensification, and virtualization to decouple specialized equipment from production software.
In the future, through the AI editorial department, Aliyun Video cloud will cooperate with more partners, use big data and artificial intelligence technology to provide remote, interactive and intelligent cloud production capabilities, complete more imaginable business scenarios, and help the digital transformation and upgrading of the media industry.
The original link
This article is the original content of Aliyun and shall not be reproduced without permission.