Title Image｜AYAYI, the virtual digital human, the official website of Burnmai Technology
On March 25th, the account of Douyin @刘夜秀 updated the fifth episode of the short drama “Earthly Branches Mystery”. In the background of sci-fi + fantasy, virtual figures and real human beings compete with each other, cold technology and warmth Humanity burst into flames. In the comment area, in addition to the praise “You can actually watch movies on Douyin”, there is also the curiosity of “ask the author, how much does this video cost”, and the expectation of “is it time to update”.
In the past six months, the “Liu Yexi” account actually only released 11 works, except for a few promotional videos of about 30 seconds, only the initial debut video and 5 feature films of about 4 minutes. As we all know, the Internet does not have long-term memory, and such output is far from satisfactory for short video creators based on Douyin.
In fact, for domestic users, the past half year has been the period of time when the virtual digital human has the highest sense of existence since its birth.
On the one hand, the technology giants represented by Meta (Facebook) once again throw out the concept of the Metaverse, which has aroused the market’s pursuit. The “virtual digital human” is regarded as the medium for human beings to enter the Metaverse in the future, and then they stand in the air of public opinion; on the other hand , the Beijing Winter Olympics has become a showcase for the current virtual digital human technology, whether it is a virtual avatar of a sports entertainment star (such as Gu Ailing’s virtual avatar Meet Gu) or a public service-oriented virtual anchor (such as CCTV AI sign language anchor), all Let the public have a more intuitive and in-depth understanding of virtual digital people.
Meet Gu 3D Avatar Stylized Digital Intelligence / Image: Web
The data can also reflect the popularity of the domestic virtual digital human industry. According to data from Qixinbao, the number of new virtual digital human-related enterprises in China in 2020 will be 36,080, and in 2021, the number will reach 66,293, a year-on-year increase of 80%. According to Tianyancha data, there are about 300,000 “virtual digital human” related enterprises in my country.
The number of virtual digital human projects currently incubating is temporarily unknown, but according to the ubiquitous parallel model of one team and one technical link supporting multiple virtual digital human IP incubation at the same time, this number should be in the tens of thousands or even more. .
If the history of the virtual digital human in the pure concept stage is blurred, then “Hatsune Miku” launched by Crypton Future Media (a Japanese sound production and music software company) in 2007 can be regarded as the first widely known virtual digital human, which we have previously known. The virtual digital people I see are mostly virtual idols and Vtubers with two-dimensional images. In this round of upsurge, the virtual digital human has gradually opened up diversified development: at the external look and feel level, relying on a new generation of engines and technologies, the hyper-realistic digital human and virtual avatar are already indistinguishable from the real human beings; at the functional application level, it has Virtual employees and virtual hosts with more service attributes have begun to show a trend of large-scale application.
In the “2022 China Virtual Human Industry Development Research Report” released by iiMedia Research, the development process of the domestic virtual human industry is divided into four stages: the embryonic stage (1980-1999), the exploration stage (2000-2014), the primary stage (2015-2019), development stage (2020-present), and the development stage is characterized by “breakthroughs in technology and demand, and virtual idols attract young groups with their unique personalities.”
Development History of China’s Virtual Human Industry / Photo: iiMedia Research
At the current stage of development, the entire industry has entered a period of rapid expansion, but it is still far from maturity. In the near term, there is still no authoritative definition and standardized classification of virtual digital human beings that have been widely recognized; in the long run, the concept of the Metaverse is still early, and virtual digital human development lacks environmental support.
So from now on, the digital people that have hatched and will be hatched may face the same problem – how to live? Just like many new things in the Internet era, digital people will also experience the test of “mushrooms after a rain – big waves wash the sand – survival of the fittest”. Therefore, technological iteration, long-term operation, and commercialization path may be the industry’s most recent issues that need to be concerned.
Note: This article adopts the classification method in the “China Virtual Digital Human Influence Index Report” by Communication University of China. Technically, virtual digital human can be divided into two categories: intelligent drive and real person (people in the middle). In terms of application, it includes three categories: identity type (such as real virtual avatar), service type (such as virtual employee), and performance type (such as virtual idol). “
Technology one step ahead?
The production of virtual digital human is a relatively complicated process. Taking Xiaozhen, a virtual digital astronaut launched by Xinhua News Agency and NExT Studios, a subsidiary of Tencent Interactive Entertainment, as an example, the NExT technical team introduced the production process of the hyper-realistic digital human.
Scanning: The Photogrammetry scanning system is used to scan a specific actor in all directions, including face models, bodies, clothes, etc., and then generate a point cloud model.
Modeling: For body and facial features, the point cloud model is converted into a low-poly model, which is convenient for subsequent binding and production. At the same time, hand-made the places where the scan made a mistake or could not scan, such as the inside of the mouth, eyes, teeth, hair.
Binding: Based on FACS (Facial Expression Coding System), MotionBuilder (3D character animation software) and Maya (3D modeling and animation software), complete face and body binding and generate simulated virtual images.
Animation: The facial and body motion capture scheme is adopted, combined with manual animation refinement in the later stage, to realize the dynamic display of the virtual image.
Rendering: Through Unreal Engine, complete character rendering and animation output, what you see is what you get.
The production of virtual digital humans, like other industrial processes, not only includes product iteration, but also the innovation of production tools. For example, through the development of a FACS-based expression model (Blendshape) production tool, the number of expression models has been greatly increased to solve the problem of rigid facial expressions of avatars; through automated algorithms, the problem of character wearing molds caused by different body proportions between virtual characters and real actors is solved. .
Since 2017, the NExT Studios team has gradually explored two hyper-realistic virtual human production pipelines, named xFaceBuilder® and xMoCap® respectively, corresponding to the face and body production of virtual digital human. In addition to the field of digital human, the two are also used in Tencent’s PC games and mobile game projects.
Virtual digital astronaut Xiao Zheng/Photo: Xinhua News Agency
At present, most large technology companies have technical teams to test the waters of virtual digital human incubation. For application parties who have actual needs and only lack technical reserves, the Baidu Smart Cloud Xiling Digital Human Platform provides a relatively simple one-stop access solution.
Li Shiyan, head of Baidu Intelligent Cloud AI Human-Computer Interaction Laboratory, introduced that Xiling platform can meet the low-cost and rapid customization of digital people, and provide operational support in terms of human design, business and content. Among them, 3D cartoon digital people can be generated in minutes through face recognition and photo modeling, and 2D flat digital people can be generated in hours by uploading live-action videos. In terms of 3D high-definition digital humans, non-specific digital humans (without real or virtual specific reference images) support minute-level generation and AI-driven; singular digital humans (with specific reference image requirements) cost the highest, requiring 10 to 30 days.
It should be noted that the fast and low-cost generation of digital humans and virtual employees and other service requirements are well suited, and the training can basically be seamlessly integrated into the current application scenario. However, for performance-oriented needs such as virtual idols, generating virtual images is only the first step, and more refined polishing and long-term operational support are essential.
CCTV AI Sign Language Anchor/Photo: CCTV News
Several virtual digital human teams who have communicated with Huxiu all believe that the industrial production of digital human is coming. It can be said that the current virtual digital human technology is developing and iterating rapidly, at least in appearance, it is infinitely close to what you want, and even individual hyper-realistic digital human beings are identified as fake just because they are too perfect in aesthetics.
But at the same time, the current bottleneck of digital human interaction is obvious. Existing AI technology is not enough to support digital human to develop independent behavior, and independent thinking goes further. It is only driven by trend concepts and technologies that the development of digital humans is one step ahead of demand. However, it is precisely because of technological bottlenecks that digital humans have not yet opened up their own survival path. Reflected at the moment, every digital person wants to become a long-term “person”, and more depends on the behind-the-scenes operation team to support.
Ignoring the application of identity-based virtual digital humans for the time being, currently performing virtual digital humans are mainly based on To C, while practical digital humans are mainly based on To B and To G applications. Staying active and paying attention is the sword of Damocles hanging over the heads of every avatar.
At present, the exposure methods of virtual digital people mainly include the following ways: publishing offline videos and pictures (such as Liu Yexi, AYAYI and other hyper-realistic digital people); talent live broadcast, stage performances (such as A-SOUL, Luo Tianyi and other secondary elements) Virtual idols); participate in large-scale events, special scene reports (such as CCTV network Xiao C, Xinhua News Agency Xiaozhen and other media virtual employees). The challenges faced by the above-mentioned digital human IP are also problems that the entire industry must face.
One is cost. On October 31 last year, Liu Yexi became popular with a short video on the theme of “catching monsters”, gaining 3.63 million followers overnight, and his influence ranked first in the “China Virtual Digital Human Influence Index Report” released by the Communication University of China. But for the past five months, Liu Yexi’s official account on Douyin has released a total of 11 videos, and the update frequency is twice a month, and “requests” appear frequently in fan messages.
And AYAYI, ranked fifth in the “Influence Index Report”, has been active in the print media, maintaining a high exposure frequency through hard photos. However, if you look at the news of Xiaohongshu, you can find that the high-frequency photos have not continued to increase the popularity and interaction, but have fallen back from the initial peak. Early photos are exquisite” feedback.
Phenomenal virtual digital person Liu Yexi / Photo: Internet
According to Li Shiyan of Baidu, “It is very expensive to use traditional methods to make CG-level videos and posters. One minute of video production for one month may be considered short, and 2 to 3 months is also normal. At present, the production tools have not been made particularly well. , limiting productivity, which in turn limits the innovation of production relations. In the future, through deep learning of artificial intelligence, the production cycle can be compressed to a certain extent.”
The Tencent NExT Studios team also confirmed the cost bottleneck, “Hyper-realistic digital human is the crown jewel of virtual digital human technology, which is far more difficult than the production of two-dimensional paper human, domestic industry accumulation is limited, and related technical talents are also very scarce. The production capacity of most domestic manufacturers is suppressed at a relatively elementary level. But at the same time, the team only needs to have a long board in one or two links, and it is easy to stand out.
At present, most of the virtual digital human production companies on the market are entrepreneurial companies, which have a strong demand for short-term financing and realization, so they will pursue rapid production or large-scale production, and it is difficult to take into account the fine workmanship. A hyper-realistic virtual digital human image may have to invest at least 1 million to 2 million production costs in order to not be ashamed of seeing people, and even if it is invested, it may not be able to make a satisfactory role. “
The once sensational Huang Renxun virtual human can also be used as a reference. According to NVIDIA’s official blog, the 14-second video of Huang Renxun’s virtual human appearance at the press conference involved 34 3D artists and 15 software engineers, totaling nearly 1,000 man-hours.
The second is content and operation. For virtual idols such as A-SOUL and Luo Tianyi, the development path for direct reference is star artist, so business ability and fan operation ability are two hard indicators. It’s just that the domestic Vtuber industry started late, and it has always been limited to the niche. Professional middle-aged people and creative teams that can customize content for virtual idols are scarce; There are only a handful of operation teams that are familiar with the culture of the fan circle by playing with the fan economy.
Lehua Entertainment, which has many years of experience in the cultivation and management of traffic stars, can be said to represent a higher level of domestic virtual idol operations. In the past year, Lehua has not only cooperated with Fang Wenshan and Xu Song to create new works for A-SOUL, but also made the mistake of not being refined enough in operation – such as “Jiaran Live Playing on Mobile Phones” “event, “suspected exposure to the person in the middle of the night” event.
Virtual idol group A-SOUL/Photo: A-SOUL official Weibo
In addition, there are some cases that also reflect the current barriers to digital human operations. In the middle of last year, Ling_Ling, a realistic digital person of Guofeng, released Xiaohongshu notes, saying that a certain brand of lipstick “moisturizes and does not dry”, which aroused doubts from netizens. “Does AI know how good lipsticks are?” “AI is about the same for things like clothes for characters, but cosmetics and skin care products are not needed.”
Lil Miquela, the originator of foreign digital celebrities, has also been controversial. In 2019, Lil Miquela shared a process of being harassed while sharing a taxi with a stranger through Vlog, but netizens thought it was obvious to use sexual assault as a marketing method. Overspill.
It can be seen that users’ trust in digital people mainly focuses on audio-visual perception. Although digital humans are close enough to humans in appearance, the gap between reality and virtuality is still difficult to bridge.
However, in Li Shiyan’s view, time can solve many problems. “Because of the development of computer technology, we have been pursuing the digitization of information in the physical world. If there is enough development cycle, feeling and everything else can be digitized.”
The third is to break the homogeneity. At present, there are a large number of virtual hosts, virtual hosts, Vtuber, etc., which are the scenarios with the most AI-driven applications. 7*24 hours on the job, the characteristics of working anytime and anywhere reflect the advantages of technology. However, as mentioned in the “Influence Index Report” mentioned above, the current virtual anchor technology has more borrowing and less autonomy, and needs attention and improvement. “Homogeneous competition” is an important challenge for virtual anchors. Whether it is a professional media operation, an enterprise or an individual operation, its hosting and live broadcast functions are similar, and stronger content planning and operation capabilities are required to stand out.
The Tencent NExT Studios team said, “Virtual digital human is an emerging product, and it is normal for it to attract a certain amount of attention, heat, and tout in the market. With the increasingly advanced digital human technology, the threshold is lowered, such as Nvidia’s Omniverse Avatar, Epic’s With the launch of tools such as MetaHuman Creator, users can see more and more things of the same level, and their aesthetic taste will also improve, and the demand for excellent works will increase. At that time, users will not only be attracted by the beautiful image of virtual digital people, but will also Take a closer look at the performance, connotation, and even the spirit behind the TA. Then the digital people who only rely on a piece of skin and a picture, but lack the deep operational ability behind the image, will definitely ebb and die on the beach.“
business path unknown
According to the “Virtual Digital Human Deep Industry Report” released by qubit, by 2030, the overall market size of my country’s virtual digital human will reach 270 billion, of which identity-based digital human (including virtual idols, virtual avatars, etc.) will account for about 175 billion. There are about 95 billion digital people, and the market is still in the cultivation stage. As for the application level, virtual anchors that replace live-action services and virtual idols with IP potential are the current market hot spots .
In the data provided by iiMedia Research, the virtual idol industry maintains a steady growth trend. In 2021, the overall market size and core market size driven by virtual idols will be 107.49 billion yuan and 6.22 billion yuan respectively, and it is expected to reach 186.61 billion yuan in 2022. and 12.08 billion yuan.
The Tencent NExT Studios team also believes that virtual idols and Vtuber may be the most promising direction for commercialization at present. “After all, they can directly interact with users, and the imagination space is larger than one-way CG videos.
At a time when the future of virtual digital people is unclear, the virtual idol group A-SOUL has injected a lot of confidence into the industry. On March 8, Lehua Entertainment submitted a prospectus to the Hong Kong Stock Exchange. Through the interpretation of the prospectus by various institutions, it is speculated that A-SOUL’s revenue in 2021 should be no less than 17 million. Although this figure only accounts for less than 3% of Lehua’s total revenue, far from being on a par with Lehua’s traffic stars, it is already the best result achieved by virtual digital people so far.
Domestic virtual idol industry forecast/picture: iiMedia Research Consulting
As for other performing and service-oriented digital people, in addition to the realization of traffic, there is no clear business path yet. Virtual anchors with goods can save the labor cost of merchants, but whether they can replace the evaluation function of real people and whether they are competitive or cooperative with real people remains to be verified. It seems that the mission has been completed, and it has returned to its previous tepid state; and most of the “virtual employees” of the current enterprise are not much different from the previous intelligent customer service, and only a few of them can be remembered using unique personalities to show the enterprise. Brand image, IP for marketing.
Huxiu also learned from Burning Mai Technology (AYAYI Incubation Team), “At present, digital people only account for a quarter of the company’s business segment. When AYAYI leads the digital human industry, her task has actually been completed, and we are now preparing Digital labels, incubate digital artists, hold digital art exhibitions, launch digital collection platforms, etc.”
Every category, and even every digital person, may be exploring its own way of survival. So what kind of digital people might be close to success for some time after that?
At a relatively recent stage, the Tencent NExT Studios team believes that the mainstream ways to create a virtual digital human IP may be as follows:
1. Stars with their own IP attributes. For example, the digital replicas of real people such as virtual Teresa Teng, Mei Lanfang, Will Smith, etc., the operation idea is actually the reproduction of real stars, which is a form of digital human with the most technical and product challenges.
2. A virtual digital human IP linked with a real person. This kind of IP operation method will have more professional and global requirements for the operator. Generally, traditional artist brokerage companies are involved, and the core demands are to break through the existing user circle and avoid the risk of real people collapsing.
3. Create IP based on mature MCN influencers. In this dimension, due to the considerable technical investment costs, the virtual digital human will inevitably tend to be at a disadvantage.
4. IPs created based on stories and game virtual worlds. This kind of IP creation method is more in line with the traditional IP creation idea, from text to plane, three-dimensional, dynamic… The content form will also be in various dimensions from traditional paper media, multi-screen coverage to new media, which is a form of holding high. The overall investment is huge, and the probability of success is funnel-shaped. Only the enduring content and quality can continue to impress the audience and form a real IP.
5. A virtual digital human IP based on the image of the second element. Although this kind of virtual idol has a relatively low technical threshold, it is difficult for users to reach them.
Domestic virtual human industry chain/picture: iiMedia Research
In the longer-term cycle, Li Shiyan of Baidu believes that the digital human is an interactive carrier based on a computer platform and will develop in paragraphs. “The current era of plane computing is mainly based on service-oriented and performance-oriented digital human applications. The next stage is a larger era, which is temporarily defined as the era of spatial computing. The information presented by the computing platform in the era of spatial computing is no longer a plane. It is real-time three-dimensional. At that time, the behavior of core users will probably be socialized, obtaining goods, information consumption, and obtaining services. A virtual avatar that can provide personalized interaction, emotional value and content value is a must. A service-oriented digital human that can provide humane emotional care and ensure efficiency is another application form. In the future era of spatial computing, there will definitely be a larger market than the current era of plane computing. “
write at the end
Just like the core contradiction thrown by Liu Yexi’s “Earthly Branches” short film – with the further development of virtual digital human technology, issues such as public opinion bubbles and ethical constraints may become potential risks, and human beings will also face AI and human beings. , Digital Man and Humanity.
In this regard, Li Shiyan believes, ” In the current field of virtual digital humans, artificial intelligence is more like a production tool . Human creativity and production tools can produce content more efficiently and at a lower cost.”
The Tencent NExT Studios team also shared their views.
“As a product of science and technology, the “soul” or “god” behind the virtual digital human still needs to be fully expressed by humans or artificial intelligence. With the development of computer graphics technology, motion capture technology and AI technology, the virtual digital human will be in the scene. Great progress has been made in technology, emotional technology, and interactive technology, which will make future digital people closer and closer to real people, or even surpass them.
On the whole, the evolutionary path of the virtual digital human is from anthropomorphic to homohuman, and finally to superhuman, realizing the return to the real and the super-real. In the anthropomorphic stage, the digital virtual human is a highly realistic three-dimensional animated character synthesized by the computer; in the homogenization stage, the appearance, emotion, communication ability, understanding ability and other aspects of the digital virtual human are the same as those of the real person; in the superhumanization stage At this stage, the ability of digital virtual human will surpass that of natural human.
In the future, the development of virtual digital human cannot rely on technology alone. To have sufficient competitiveness, it is necessary to create a new experience and have a deep understanding of the entire application scenario. Meeting user needs, giving users an unprecedented emotional experience, and allowing users to accept virtual humans are the only way for virtual humans to succeed. “
Posted by:CoinYuppie，Reprinted with attribution to:https://coinyuppie.com/virtual-digital-people-are-not-easy-to-live/
Coinyuppie is an open information publishing platform, all information provided is not related to the views and positions of coinyuppie, and does not constitute any investment and financial advice. Users are expected to carefully screen and prevent risks.