LiveVideoStackCon2021 Beijing Station Interview: From Cloud to Innovation, New Technologies and New Scenes of Video Cloud

With the advancement of video technology and the iteration of standards, the video industry has entered the digital age from analog to complete the media conversion from film and television to the Internet, and has derived various innovative forms such as ultra-high-definition, 3D, and AR/VR. Especially in the current post-epidemic situation, we can see many new changes in the audio and video technology field, the collaborative interaction between cloud and end, the deep integration of algorithm innovation and engineering applications, the penetration and promotion of scenarios and requirements, under severe challenges, All walks of life have brought new scenes and vitality.

At the upcoming LiveVideoStackCon 2021 Beijing station, experts from Alibaba Cloud Intelligent Video Cloud will work with many industry partners to explore and discuss the innovative exploration of video cloud technology on the cloud. To this end, we interviewed Alibaba researcher-Ye Yan, and senior technical expert-He Yaming, and the two experts had an in-depth dialogue on the new scene application of codec technology and video cloud.

"Video socialization": Video cloud becomes the new infrastructure

From the rise of online video in 2006 to the era of "video socialization", 5G, cloud, and AI have become social development trends. Video is no longer limited to traditional media fields such as movies, television, and advertising. Video conferencing and interaction New applications such as video and e-commerce live broadcast have gradually melted the industrial boundaries, and the demand and technology of the video industry have been continuously upgraded. With the development of technology and the consolidation of infrastructure, video will become a new form of interaction and information carrying.

(Source: iResearch-2021 China Video Cloud Scene Application Insight White Paper)

For the fiercely competitive and rapidly iterating big video industry, the video cloud has gradually developed into a key infrastructure. As we all know, the current video services consume very high resources such as computing power, storage, and bandwidth. For example, a popular live concert may be watched by millions of people, which not only requires powerful end-to-side real-time video Processing power, relying on a large-scale CDN distribution network to complete smooth distribution tasks, and even some special AR/VR visual effects need to be presented through the computing of edge nodes, so just moving the server to the cloud is far from satisfying. For future scenarios, how to use the advantages of cloud to evolve technical architecture and services has become a common issue facing the industry.

Ye Yan: Promote the implementation of next-generation video standards and release industry productivity

Ye Yan is a researcher at Alibaba and the head of Alibaba Cloud Smart Video Cloud Video Standards and Implementation. She is responsible for the technical development of Video Cloud in ITU-T VCEG, ISO/IEC MPEG, AVS and other international and national video standards organizations, involving the development of advanced technologies such as video codec, AI video quality assessment, and VR/AR. She has participated in the formulation of a number of international standards for video codec and streaming media, including H.266/VVC, H.265/HEVC, SHVC and other standards. She is the author of more than 50 academic papers, the inventor of more than 130 US granted patents and more than 230 US patent applications. She is also a senior member of IEEE. She received her bachelor's and master's degrees from the University of Science and Technology of China, and her Ph.D. from the University of California, San Diego.

Video is inseparable from coding and decoding technology, and coding and decoding are inseparable from the guidance of standards. Video standards have always been the infrastructure for the development of the video industry. Video standards cover a wide range, from the system standard MPEG CMAF to the codec standard H.266/VVC. The continuous update and iteration of the video standard improves the efficiency, cost reduction and new experience of video production. It plays a vital role, and it is also related to the future direction of the entire industry.

As a researcher at Alibaba and the head of the Alibaba Cloud Video Cloud Video Standards and Implementation Team, Ye Yan has always been an in-depth participant and promoter of international video standardization work. One of the best occasions to grasp the pulse of the latest industry is through open technical discussions by industry experts and at the same time fully listening to the needs of the market, so that we can iterate over more efficient standards and continue to promote the progress of the industry."

However, facing the new stage of development, the industry has also put forward different opinions on some video standards organizations. There is a view that standards organizations like MPEG have lost their leading role, and everyone is still racking their brains for a fraction of a few performance gains, and this brings greater computational costs. This self-healing innovation It is more of a sense of existence, and does not bring essential technological advancement or innovation. The industry should find new ideas to solve the video compression problem.

In the face of such noise, Ye Yan expressed his own judgment-"I do not agree with the view that the traditional framework and the new framework are regarded as isolation or even opposites. Although it is more and more difficult to mine performance under the traditional framework, this direction is Based on the familiar framework, it is conducive to the implementation of software and hardware, and ECM also fully demonstrates that this framework can still provide considerable performance gains, so it cannot be easily given up. On the other hand, JVET is also exploring what new frameworks or new Tools can be achieved overnight and get substantial performance gains. At the same time, we are also very concerned about the computational cost of what level of water this new framework needs? To be honest, we are still exploring, so we must rely on two-legged walking to find the most effective Potential and achievable next-generation codec technology."

Indeed, formulating a generation of coding standards is a very difficult task and cannot be accomplished overnight. Taking the current industry's latest standard VVC as an example, the pre-research work before it officially started took about 3 years. It is precisely because of this that, less than a year after the VVC standard was finalized, JVET established the ECM software platform in the first half of this year to carry out the technical pre-research and development of the next-generation coding standard. Ye Yan said: "Although the current compression capacity of ECM has exceeded VVC by about 14%, based on previous experience, this pre-research work will take several years to achieve the compression performance gain requirements of the new generation of standards. In the market and business In today's ever-changing world, I expect that the past few years will witness the rise of many 5G video application scenarios."

He Yaming: "Cloud + Terminal + Service" is the future trend of video cloud

He Yaming is a senior technical expert of video cloud in Alibaba Cloud's intelligent business group, and the head of video cloud technology research and development. Prior to joining Alibaba, he worked at Facebook and Microsoft in the United States, worked as the Principal Software Engineer at Microsoft, engaged in the research and development of video coding and video cloud, and was responsible for the research and development of real-time audio, video and live broadcast technology at Facebook. In just a few years, he will integrate Facebook Messenger and Facebook Live. The two products were built from scratch to a star product with 1 billion users.

"Audio and video have natural cloud-native attributes, and'cloud + terminal + service' is the general trend of future audio and video development." This is the judgment made by He Yaming, senior technical expert of Alibaba Cloud Intelligent Video Cloud and head of video cloud technology research and development.

In He Yaming’s view, the development of audio and video has always been the best practice of cloud native: cloud infrastructure-including central nodes, edge nodes, and CDN networks are the basis for ensuring large-scale audio and video distribution and transmission; cloud computing capabilities and Random flexibility can bring unlimited computing power to audio and video services while effectively controlling costs and deriving more new scenarios. In addition, as audio and video end-side devices are becoming more and more abundant today, the coordination of “cloud” and “end” becomes more and more important. In 2020, Alibaba Cloud proposed the “cloud integration” strategy. In this context, Its path advantages are becoming more and more prominent-relying on the powerful cloud computing power of Alibaba Cloud, it can make the terminal smarter, lighter, and more flexible, allowing developers to create innovative applications with thousands of people, and its development efficiency, operation The maintenance cost and ductility have been greatly optimized. On the road of "integration of cloud, integration of cloud and edge, integration of software and hardware", He Yaming especially emphasized the important role of AI in it-"We especially emphasize the application of AI, from intelligent video coding, image enhancement to super-resolution From smart beauty, virtual background, beautifying voice to video cartoonization, it can be said that we are using the AI power of the entire group to promote the audio and video scenes to a broader space."

(Alibaba Cloud Smart Video Cloud participates in the National Key R&D Project of Cloud Broadcasting Platform for the Winter Olympics)

"At this summit, the special theme of Alibaba Cloud Video Cloud is'From Cloud to Innovation, New Technologies and New Scenes of Video Cloud". Here I want to emphasize the word'innovation'. Shanghai Cloud is already in the video industry. We have basically completed the process of cloud nativeization. The real problem we face is how to complete the next stage of innovation on the cloud. Manufacturers should transform from providing resources and tools to providing services and ecology as a breakthrough." He Yaming Said so.

At present, most of the leading cloud vendors in China have strong technical service capabilities and a complete content consumption ecology, making video products service-oriented, through API, PaaS services, PaaS+, SaaS tools, SDKs, low-code platforms and other means to reduce The access threshold of video technology can better serve developers, and ultimately better serve video production and consumers.

Today, in the face of fierce competition from domestic leading cloud vendors in the field of video cloud, He Yaming sees more opportunities: "This is a trend we are very willing to see, and it is also the result of our continuous advancement of the industry. Alibaba Cloud also hopes that more and more people with lofty ideals will join the video cloud team, and together bring audiovisual into a new era."

Technology and Scenarios: Future Innovation and Challenges of Video Cloud

At the Alibaba Cloud Smart Cloud Summit held in Beijing in May 2021, Zhang Jianfeng, President of Alibaba Cloud’s Smart Business Group, announced that Alibaba Cloud will add “good services” on the basis of “deepening the foundation, strengthening the middle platform, and strengthening the ecology”. "As an important strategy. Video cloud technology is a field where cloud computing, artificial intelligence, network and other technologies are very closely integrated with industry scenarios. Alibaba Cloud has always adhered to the deep cultivation of the underlying technology, the application of middle-stage technology and the innovation of service scenarios.

Video coding and decoding is a technical field in which Alibaba has always had an advantageous position in the industry, and it is also the specific action of the group to adhere to the research of basic audio and video technologies. The Alibaba Cloud Video Standards team has just finished the intense technical development of the new generation of international video codec standard H.266/VVC in mid-2020, and immediately put in manpower and began to vigorously promote the development of codecs based on H.266/VVC Work. Soon afterwards, Alibaba Cloud released the real-time high-definition codec Ali266, which strongly promoted the implementation of H.266/VVC standard applications and truly opened the way for H.266/VVC to be commercialized.

When talking about the difficulties in the development of Ali266, Ye Yan said: "A mature commercial encoder must pass the deep optimization of the algorithm to meet the requirements of real-time encoding speed. In order to get the powerful compression performance provided by H.266/VVC, it must For the input video content, the most reasonable coding tool can be selected quickly and accurately from the many coding tools provided by VVC. Therefore, we developed Ali266 along this trajectory and went deep into the VVC coding tool set, through the qualitative and quantitative analysis of each coding tool To help us choose coding tools. At the same time, we also pay special attention to subjective quality in the process of algorithm optimization. When encountering conflicts with objective quality indicators, we will be more inclined to ensure higher subjective quality. That is to ensure the ultimate user experience. Ali266 can reach the encoding speed of real-time HD and real-time full HD at the first time, and at the same time, it has a sufficient gap with the encoding performance of HEVC. It is directly related to our adoption of such a development strategy. Now it is emerging VR/MR requires a higher resolution video format as a technical base support, so the bandwidth saving ability provided by VVC is also more valuable. Therefore, we will continue to invest in the development of Ali266, so that it can run faster and faster, and reach it in the near future. Real-time encoding capabilities of ultra-high-definition 4K or even 8K. It will also provide a good landing scene for more efficient encoding and decoding standards."

Not only the deep cultivation in the field of audio and video technology, but with the in-depth integration of Alibaba Cloud's video cloud business with the overall business of Alibaba Group and the deep cultivation practice of industry customers, Alibaba Cloud Video Cloud and People's Daily New Media, Taobao Live, LAZADA, Youku, etc. The scene cooperation with external customers has also become more abundant. In 2018, Alibaba Cloud and Olympic Broadcasting Service Company joined forces to create the Olympic broadcast cloud OBS Cloud. This year, the Olympic Broadcast Cloud was put into use for the first time at the Tokyo Olympics to provide broadcast support on the cloud for global broadcasters. This is the first time in the history of the Olympic Games that cloud computing has been used to support global video broadcast, allowing global audiences to break through the epidemic barrier on the cloud.

(For the 2020 Tokyo Olympics, Alibaba Cloud cooperated with the International Olympic Committee to achieve the entire "Olympic Cloud")

In the face of the continuing global epidemic, He Yaming predicts that the demand for video technology will continue to grow in live broadcasting, conferences, e-commerce, entertainment, and collaboration-"With the development of 5G, AR, and VR technologies and the development of infrastructure Perfect, lower latency (<100ms), higher definition (8K+), and more immersive (3D holographic, surround sound) interaction methods will change many industries. In addition to people, audio and video will also make people and things. More connections are established between things, and the way of human interaction will be upgraded again. Remember a popular saying in the media: the beginning is the end. It means that human beings first receive information and experience the world by sight. From the voice to the text to the picture and then to the video, it finally returned to its original form. I think this judgment is not entirely correct. The interactive form of video is still evolving. The movie Matrix and the number one player, including the recently popular Metaverse, have already given We paint a brain hole in the future form of communication."

From cloud to innovation, new technologies and new scenarios of video cloud

Topic

⏰ Activity time: 2021/10/30 14:00-18:00
🚀 How to participate: Coordinate Beijing, offline participation (free)

Scan the QR code in the picture or click to read the original text
Learn more about special events
↓↓↓

Scan the QR code to join the group
Learn more about LVS conference and video cloud information
↓↓↓

"Video Cloud Technology" Your most noteworthy audio and video technology public account, pushes practical technical articles from the front line of Alibaba Cloud every week, and exchanges and exchanges with first-class engineers in the audio and video field. The official account backstage reply [Technology] You can join the Alibaba Cloud Video Cloud Product Technology Exchange Group, discuss audio and video technologies with industry leaders, and get more industry latest information.