Shulou(Shulou.com)06/03 Report--

Now it is spring, but under the epidemic, the cold winter of the film and television entertainment industry is even more chilly.

A notice "about COVID-19 stopping filming during the epidemic" added haze to the originally anxious film and television production industry. By the end of last year, a number of giants in the film and television industry had fallen into a state of loss, while the Spring Festival films had almost no revenue off-line. At the same time, a number of theatres across the country have also posted notices to cancel or postpone the performance until the second half of the year. It can be predicted that with the continuation of COVID-19 's epidemic prevention and control work in the first quarter, movies and TV dramas, variety shooting and recording, and even offline performances are likely to be completely shut down in the first quarter of this year.

The shutdown of filming will directly lead to a significant reduction in the production capacity of the entertainment industry this year, while the shutdown of offline performances and entertainment activities will also sharply reduce the income of both front-line actors and ordinary entertainers. As the epidemic has spread around the world, it will be very difficult for international entertainers, including music and drama, to come to China in the future. This "force majeure" will be a very severe challenge to the whole production chain of the traditional entertainment industry and the performing arts market.

How does the film and television entertainment industry save itself? How to find new industrial opportunities in the agony of being jobless? It can be said that the arrival of each crisis is an opportunity to achieve new industrial innovation, such as the Great Depression of the United States in 1929 directly stimulated the prosperity of the film industry and Hollywood. Crisis is the external opportunity of industrial upgrading, and technological innovation is the real driving force of industrial upgrading. The film and television entertainment industry has encountered a fatal crisis of production interruption in this epidemic. However, with the maturity of 5G, AI and digital generation technology, new modes of production of entertainment content have emerged one after another.

We take this opportunity to take stock of the new technology cards in our hands, and what innovative techniques we can use to play a few good cards for the upgrading of the film and television entertainment industry.

Break the bottleneck of entertainment production capacity: hit the "spring" of first-hand AI to generate audio and video content

With the intensification of the epidemic, the disruption of offline content production capacity is a fatal blow to the film and television entertainment industry. Audio and video shooting has a natural gathering characteristic, even if there are only two or three people participating in the interview variety show, it may need nearly ten people to participate in the shooting at the scene. So, at this stage, is there a more mature AI and digital technology that can help the film and television entertainment industry resume production quickly?

So one of the problems we have to solve is to obtain real-life images without the need for real-life photography. In fact, from the frequent "face-changing" videos released last year, it is no longer difficult to use the GANs algorithm to generate live-like video content.

Take the news broadcast as an example, most of the traditional news information is presented in the form of words. With the use of artificial intelligence technology, textual news can be efficiently transformed into the audio and video broadcast of virtual AI anchors. Through the anthropomorphic training of image expression learning, sound habit learning and logical thinking, we can vividly create an AI avatar of well-known anchors, and can broadcast real-time news for people through mobile phones and TV screens.

In February last year, Sogou and Xinhua jointly released a standing AI virtual anchor with a more realistic image of facial expressions, natural body movements and lip movements. Prior to this, two AI synthetic anchors had accumulated more than 3400 and 10000 minutes of news reports on the Xinhua platform. With the improvement of the technical ability of "Sogou split", the customization period of its AI synthetic anchors has been greatly shortened, and the synthesis effect and stability have also been significantly improved, which can achieve the ability to continuously copy different types of AI anchors.

With this AI generation technology, through the ability to quickly generate audio and video through text, film and television production companies can, based on the shooting plan, allow actors to submit basic sound and video materials online in accordance with the requirements of the script, and then give them to the later team to generate virtual actor images to complete the performance in the play, greatly saving offline shooting and production costs. And many media, self-media and other entertainment content production can also use this technology to achieve efficient production capacity. Whether it is funny jokes, game commentary, film and TV drama commentary, or knowledge popularization, you can input text into customized virtual anchors to produce more professional audio and video programs.

Although at present, most of the AI audio-visual generation is only used in the relatively simple form of audio-visual content, but with the continuous evolution of related technology, we can have a broader imagination.

At the CES exhibition just past, Samsung launched a new "digital virtual human Neon" project in a high profile. According to the official video, the temperament, language, movements and expressions of the virtual human on the screen are no different from those imitated by the real person, especially the details of the eyes and lips, which are difficult to detect by computer. And after the learning of the real human image by the behavioral neural network, the virtual human of Neon can generate new expressions, actions and even dialogue, so it is more realistic.

Similarly, many domestic technology manufacturers have also carried out the commercial practice of landing. Take the Force animation of a domestic super-realistic "digital virtual human" development company as an example. The system they developed can generate actors who approach human beings and are difficult to distinguish between real and false through computers without using real people.

Specifically, to make these digital images of virtual characters, we need to build a whole set of simulation system according to human mechanics and biological force, including the latest facial muscle system, whole body muscle system, performance capture system, and automatic facial animation recognition system. Taking the body system as an example, they want to divide the realization of the body movement process into five stages: bone animation, muscle solution, fascia solution, fat solution and skin solution (and skin and expression combination), through a series of reactions, to "restore" the squeezing, stretching, sliding and dithering effects of real muscles. The second is the 3D real-time expression training system. This system can import the muscle data generated by dynamic solution, and the data carved or scanned by human 3D image engineers into the "expression controller" to produce expressive movements efficiently.

Under such a simulation system, the expression and action of the virtual human will be infinitely close to the real person. With the introduction of machine learning and deep learning of AI, the production efficiency of these virtual characters has been improved rapidly, and the possibility of commercialization will be greatly enhanced. Although artificial virtual humans can not reach the natural level of real people in performance, they can fill in the gaps of some non-important roles to a certain extent and change the gathering attribute of personnel in the production of audio-visual content. Perhaps we can look forward to the future recording of a MV, dancers can be synthesized through virtual human in the later stage, and busy streets and crowds can also be processed in the later stage when shooting movies.

With the ability of AI to automatically generate audio and video technology, as long as there is excellent content creation ability, any platform can play the "spring" card to win the future.

Reality shines into the virtual: the "Wang Fang combination" of AI knowledge graph + fan economy

Under the epidemic, filming stopped and actors stopped working, which is a lose-lose situation for both film and television production companies and actor agencies. Is there no solution to this game?

No. As long as another way, film and television production companies and brokerage companies can also find new ways to cash.

Throughout the current fan economy, the trend of "everything can be a meal circle" is becoming more and more obvious. Apart from the traffic stars, pandas in the zoo, anime characters and even characters in the novel can have their own "fan circle" and have a group of fans who play the list, support and fight against the triad. From this point of view, what fans consume is not only the works of "artists", but also the "people" of the fans.

This kind of person setting is not without the possibility of commercialization. Last year, an accounting software was all the rage, which features the combination of chatbot and accounting function, and users can set the chatbot as their idol. In the process of bookkeeping, the chatbot will communicate simply with the user in the artist's "tone". This mode of "chatting" with idols has brought rapid traffic growth to this software.

With the support of NLP technology, chatbots can imitate their tone, language habits and so on vividly by learning the social media content of their idols. The ability to define entity concepts through knowledge graph enables chatbots to better reflect their personality-idol's social relationships, daily preferences, habits and so on. The addition of these details can add to the reality of the interaction.

Perhaps by working with hardware manufacturers of smart speakers (screens) and conversational software such as intelligent assistants and intelligent customer service, artists' personal IP can be combined with software and hardware to become a new product that can interact with and accompany fans. At the same time, through the dynamic knowledge graph, a unique common memory knowledge system between virtual idols and users can be established to increase the "sense of nurturance".

Perhaps at a time when the content output of the entertainment industry is difficult, the brokerage company opens the license and cooperates with intelligent speaker manufacturers to launch the "personality package" and "voice package" of an artist IP, which can not only benefit from cash, but also promote users' enthusiasm for voice interaction with intelligent hardware.

With the continuous evolution of AI technology, domestic film and television, brokerage companies may also play a "AI technology + fan ecology" Wang Fang combination.

Variety show breaks through time and space: 5G distributed recording plays "flush"

At present, affected by the epidemic crisis, many variety shows and performances across the country have to be postponed or even cancelled. The mode of performance on the stage and offline recording and performance of the audience in person ushered in the realistic need of change for the first time because of the sudden crisis.

With the blessing of 5G network technology and AI capability, distributed real-time recording is becoming a reality. As we all know, the ultra-high speed, ultra-low latency and a large number of terminal networking support of 5G network make simultaneous live broadcast across regions and even across borders possible. On December 20 last year, on the occasion of the 20th anniversary of Macao's return to the motherland, students from Peking University and Macao University of Science and Technology sang "Song of Seven sons" with the help of 5G technology and 5G mobile phones, synchronizing time and space 2000 kilometers apart. to produce a wonderful experience of technical aesthetics and human emotion. Previously, the 5G multi-place ensemble concert jointly held by domestic operators and Huawei also successfully proved the technical feasibility of multi-location simultaneous performance.

At the same time, in the "host contest" of CCTV in 2019, the evaluation system of "professional + public" was creatively adopted. On the other hand, the judges who participate in the online public review create their own 3D virtual images through AI face generation technology. Each virtual judge with different images and vivid expressions can be seen on the big screen. The excellent performance of this technology can provide an excellent solution for many variety shows recorded in the shed.

Although there are no real cases of 5G distributed recorded variety shows to share. But we can simulate the recording of debate programs in an epidemic environment. Due to the impact of the epidemic, it is impossible for judges and viewers to be present, and debaters have to debate at home, so how does the host organize this debate contest that everyone participates in? Obviously, the host needs a camera and two huge screens, one showing live footage of several judges and contestants, and the other showing the avatars and scoring data of the virtual audience. Others just need a camera and a TV screen. With the help of the 5G live broadcast and the host's voice control of the debate process, everyone can synchronously complete the online debate and even compete.

I believe that if you record this variety show, you will find the picture a little funny. However, under such harsh conditions, this real-time distributed live recording is already a feasible solution. With the maturity of holographic projection technology, it will be more natural for us to talk and debate face to face with people thousands of miles away. Sitting at home, with the blessing of 5G and AI, will give you a flush of "divine travel thousands of miles".

Historically, military warfare and entertainment games have been the earliest users of cutting-edge technology. The former means the self-destruction of civilization, while the latter is the creation of civilization. This epidemic war is also a difficult test for us to deal with together, but for all industries, especially the film and television entertainment industry discussed in this article, it is also an excellent opportunity for technological breakup and industrial innovation. Of course, opportunities always belong to a small number of forerunners.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.