The following record, from the internal sharing meeting in August 2020, introduced the research and cognition of the Shunwei team in the direction of virtual people, that is, "virtual people are the core interaction mode of the virtual world", and the secondary culture of the target that had been invested at that time.
At that time, Metaverse was not yet a hot word, let alone a hot direction, and we also had to use "Ready Player One" to refer to the vision of this virtual world. Now it seems that the research at that time is still not outdated, share it with you.
Today, the popularity of the virtual person track continues to grow, more and more people see that the virtual person is a part of the core of the virtual world, we are also under this judgment, continue to look for the target, and in 2021 invested in a very good AI-driven virtual person technology company, initially completed the layout of this track. These two companies also achieved more than expected growth in 2021.
In the following you will learn:
Why did we choose to start focusing on the "digital people" space in 2019?
What are the "definitions and representations" of digital people?
What are the "development stages" of the Digital Human Raceway?
In the eyes of investors, what are the "entrepreneurial entry points" of the digital human race track?
Why did we invest in the virtual human ecology company "Sub-Culture"?
When will the transition from 2D to 3D the digital person arrive?
……
How far are we from Metavance in Ready Player One?
Shunwei Capital Vice President Feng Zheng
First, why start to pay attention to digital people in 2019?
Now that we have seen that there may be huge opportunities in the vr and AR fields, many people may ask, is this technology a little far from our real life, and there is still a long time to really apply this technology in real life?
So, as an investor in the VC industry, in addition to analyzing the future, we also need to focus on the present - what technologies are achievable in the present of the huge opportunities that may arise in the future.
What I'm sharing with you today is the opportunity we're currently observing that can be achieved with XR technology, the Digital Man.
A digital person is a person created in a virtual world using AI and 3D rendering technology, and in simple terms, it is a "dummy who can talk to you." The concept of digital humans may seem new, but it doesn't use emerging technologies, but rather, it's the product of the combination of two very mature technologies.
One of these two technologies is computer image technology.
Simply put, it is to create a very good image on the screen, and this technology has been widely used in movies, television and games. For example, Elsa in the Disney movie "Frozen", minions, and the most classic Mickey Mouse image and so on. In addition, there is alita, the battle angel who is considered to have reached the new peak of film and television 3D rendering technology. In the game, 3D rendering is also a very important technology, such as a Switch game that was very popular during the epidemic this year: "Animal Forest Friends", which designed all the very cute people and small animal images and so on.
Another technology is AI.
Specifically, we usually mention: speech recognition, semantic understanding, speech synthesis, image perception, and so on. These technologies allow computers to understand, understand, and express our language and actions. The application of AI technology in reality has also become more common, such as: Xiaomi's little love classmate, Apple's Siri, Amazon assistant Alexa and so on.
The digital person, which combines computer image technology and AI technology, is a new direction that combines making an image and enabling him to interact with people intelligently. Below, I've listed some of the more representative numbers of people.
2. Representative digital people
The first is the "grandmother contestant" Hatsune Mirai among digital people, who was originally a software that can make electronic music, and later created an image with her voice. The second, Trip Love is the hottest virtual YouTuber in Japan, doing some virtual performances on YouTube and getting the love of many people. Next is the Bilibili hatched out of the iris, the image is more Chinese style, which is an image of benchmarking love. It can be seen that the first three images are more two-dimensional and anime-style.
Then there's Lil Miquela, a company in Los Angeles, an ultra-realistic virtual icon who debuted on Instagram, including pictures, videos, interviews, etc., creating the image of a supermodel. In addition, the hyper-realistic style of virtual idols also has IMMA from Japan, this image will be more in line with the Asian aesthetic, IMMA also once filmed an MV with Dou Jingtong.
The latter three images will be closer to reality, first of all, this one wearing a blue suit is the AI female host "New Xiaowei" launched by Xinhua News Agency and Sogou, who can automatically broadcast news according to the content of the press release, and can directly generate a dynamic video, and the effect is also very close to the real broadcaster. Then there is Gatebox, a digital speaker virtual person with a holographic screen made by a Japanese company, and there is a two-dimensional girl in the speaker. As soon as the user returns home, the girl in the speaker can immediately interact with him and have a conversation, such as broadcasting the weather, daily greetings, and so on. The last person in a white coat was Neon, one of a series of digital people released by Samsung at last year's CES conference, whose main idea is that digital people can gradually replace many interesting roles in people's daily lives.
Next, a very important concept that I want to share with you is that digital people will become apps in the 3D world. This is different from the current concept in many media articles and industry analysis.
Third, digital people will become the app in the 3D world
Looking back at history, we can see that the way humans and machines interact is constantly changing, essentially because the technology we interact with is constantly changing. For example, when you first use the computer, the interaction method is to achieve operation by looking at the website and web page in the browser on the monitor, and through the instructions conveyed by mouse clicks and keyboard keys.
In the era of mobile Internet, the terminal equipment we use has become a smart phone, the biggest difference between smart phones and computers is the touch screen, the core technology in the touch screen is called multi-touch, which everyone is already very familiar with. So through such interactive technology, our interaction method is no longer to use keyboards, mice and browsers, but to use an app in the mobile phone. Everyone's interaction has become the touch of the finger, and a series of permutations and combinations have been derived, including the operation of sliding up and down the "vibrato" that everyone is very obsessed with now.
We believe that in the world of 3D technology, the terminal devices used by people will become VR and AR, so there is actually no concept of screens in such a world. When we look at the interaction technology, all the images we can see are rendered in 3D, and in the process of interacting with these images, the technology behind them is the AI we just mentioned.
So we think that in the 3D world, the future of the app will become a digital person, we will interact with the digital person to obtain information, the way of interaction is no longer to press the mouse, hit the keyboard or click the screen, but directly with our most natural language and actions.
Therefore, it can be said that digital people are the most important interactive medium for us in the 3D world, and will become the next generation of human-computer interaction. The super unicorn app we are constantly looking for in the era of mobile Internet will become a digital person in the future. The way to interact with digital people is actually closer to the natural habits of human beings, so it will be a more convenient and comfortable experience.
Fourth, the entrepreneurial entry point of the digital human race track
Next, let's talk about the entrepreneurial entry point of digital people.
First of all, there are already many ideal entrepreneurial teams that choose to use the "one-step" approach to directly create a future world composed of countless digital people, and the hottest entrepreneurial direction in this is digital people socializing. But at the same time, we can see that most of these companies are unsustainable.
Among the better startups are VR Chat, which was founded in 2014, and as of now, there are still some loyal users in this app who create avatars to interact and chat. There's also Altspace, founded in 2013, which has now been acquired by Microsoft, and now Microsoft will choose to host some virtual events on this platform. Finally, there's the Horizon project that Facebook (now renamed "Meta") released last year with the hardware device Oculus, which is now a very important business in Facebook. In addition, there are many startups that choose to try digital people on smartphones, such as Zepeto, an application that caught fire some time ago.
At present, these applications have not been fully validated, and the reason is that we believe that the "one-step" digital human social world needs to have a better experience in VR/AR devices, and the speed of development is limited by the popularity of VR/AR devices. However, the current situation is that the penetration rate of VR/AR devices is far from enough, so we think this direction is not the best entry point for startups.
Instead of doing a lot of digital people in the 3D world, for now, the better entrepreneurial direction is starting from the 2D screen.
Startups can first choose to create one or more digital images that are loved by users in 2D screens, such as lil miquela, the Instagram influencer I just introduced, brud, whose operating company is located in Los Angeles, USA, and has received investment from institutions such as Sequoia in the United States, with the latest round of valuations of $125 million.
The company's main source of income is to jointly brand and promote fashion brands through Lil Miquela's virtual supermodel image, so as to harvest brand advertising sponsorship.
Another company with a more Asian character is Rainbow Society from Japan, which completed a huge financing of 1.9 billion yen in 2020, and its revenue model is completely different from Brud's. Rainbow Society mainly uses the two-dimensional image created by the company to conduct online live broadcasting, and obtains income through live broadcast tipping.
So, from a startup perspective, we think that creating and presenting digital people from the 2D screen of a phone or computer is actually a very effective and grounded way to cut in.
V. The first stage of development:
Virtual idols improve the profit distribution mechanism of the content industry
So, in the digital human track, what stages will be divided?
The first stage: "Virtual Idol" improves the profit distribution mechanism of the content industry.
The process of making a "virtual idol" can be roughly divided into the following steps, first of all, the startup needs to make a 3D model first, and the style can be two-dimensional or realistic. After the virtual idol is ready, it is necessary to find a so-called "nakami" (the word comes from Japan), and the nakahito will play the avatar. Through the camera to capture the movements of the person in the middle, so as to complete the anthropomorphic, virtual production. Therefore, the virtual idol can be understood as an image driven by the middle person.
So why is the popularity of virtual idols continuing to grow now, and startups are keen to create a virtual idol?
First of all, at this stage, virtual idols, although they don't seem so smart, solve a problem that plagues everyone in the current content industry: the profit distribution mechanism.
For example, we all think that many brokerage companies look glamorous on the surface, but they actually have a hard time, because most of the income is taken away by the stars. At the same time, we can also see another type of company led by Disney, whose revenue and financial reports are very good, because Mickey Mouse will never ask disney bosses for salary increases. So the first advantage of virtual idols is that it changes the profit distribution mechanism of the content industry, and the company can fully control this image and own all the use rights and revenue rights of virtual idols.
Below I will introduce several main lines of development in this industry.
The first is the most basic hardware and software vendors to provide low-level technical support: for example, Unity and Unreal, which do 3D rendering in software; and Noiton and Vicon in terms of hardware, we can see the devices worn by actors when watching film and television drama tidbits, which come from these hardware companies.
Further up are technology service providers who are able to integrate these foundational technologies to help other companies create and drive their own digital human avatars.
Representative companies at this level include: Facerig, which uses the lowest threshold and most down-to-earth technology solutions, and there are many online anchors who are using it, and we can download this software for dozens of yuan on the Steam platform. Facerig mainly supports users to use the CAMERA on the PC side to capture their own face movements, thereby driving some of the built-in digital images of Facerig and making a half-body virtual digital person performance. The threshold for achieving this effect is relatively low and belongs to the entry level.
In addition, there is the "super dimension" from Guangzhou, the company will provide users with software, so that users can buy their own motion capture equipment, import their own models, so that users can support users to make more professional virtual idols driven by motion capture devices.
XMOV from Shanghai, also known as Mofa Technology, the company's technical strength will be stronger, can support users to make some hyper-realistic, higher-precision virtual idols. Why is high accuracy an important metric? The higher the degree of anthropomorphism required by virtual idols, the higher the requirements for accuracy. Representative companies in this field also include Evenbang and Facegood.
Above the technical service providers are the operators. The first image in the following figure is the original and most influential virtual idol "MOMO sauce" on Douyin, the original meaning here is: MOMO sauce was created in accordance with the way of virtual idols on the first day, and the operator behind her is a project incubated by a company that does video big data analysis, and currently has millions of fans on Douyin, and the daily content is mainly presented by video and live broadcast forms. The second is "a zen monk", who is actually the head of an MCN organization and is a very influential IP. Because his body is an IP made in accordance with the way of film and television animation, he belongs to the virtual idol of "halfway out of the house". At present, a small Zen monk will cooperate with some of the technical service providers mentioned above to achieve interactive live broadcasts and performances. Le Element is a very well-known game company that has produced their own virtual idol variety shows and web dramas. Shanghai Henian is also a relatively old virtual idol production company.
Virtual idols are made, where do they go to perform?
At the top level, there is the display platform. We can see that in the current stage of using 2D screens as the medium, there are many display platforms: Bilibili, Douyin, iQiyi, Kuaishou... These are all well-known companies. They are also currently working on virtual idol-related projects, such as Bilibili, which has set up a company to produce and operate virtual idols; iQiyi has also produced a virtual idol band. These platforms have also made great investment in the field of virtual idols.
To sum up, there are currently four main lines in the virtual idol industry: basic hardware and software suppliers, technical service providers, operators and display platforms.
From an investment point of view, we will grasp the two main lines in the middle:
The first is the production level around the virtual idol, that is, 3D model generation and driving technology, the core of which is to achieve faster and lower cost production and drive 3D models;
The second is the main line around the operator, successful IP creation and operation, that is, how to successfully create a favorite IP image, so it can also be summarized as an "artistic" main line.
Therefore, the investment idea of virtual idols can be simply summarized as: two main lines of technology and art. In the case of Disney, for example, Disney's technology is constantly iterating, but its core ability is to constantly create images that can be loved by audiences. Therefore, it can be seen that technology and art are parallel to each other and support each other.
Sixth, the second stage of development:
Digital people replace simple human-to-human interactions
Let's talk about the second stage of "digital people": digital people replace simple human-to-human interaction, and the core is to improve efficiency.
For the sake of being more intuitive, I made an analogy here: smart manufacturing is currently a very hot field, and the main reason is that we think that "robots" are essentially an alternative to simple interactive work between humans and machines.
Then the analogy is in the field of "digital people", that is, digital people can replace simple human-to-human interaction. For example, digital people can replace simple customer service work, sales consulting work, medical simple consultation, and so on.
Sogou, its news AI anchor, can automatically generate an unlimited amount of program content as long as it has a text; Sogou has also cooperated with other companies to make some digital human customer service, such as video interviewers and so on. Mofa Technology, in the bank lobby of the digital receptionist, this is also an entry point we have observed in this industry, the demand gradually increased.
So what's the technology behind these scenarios? That's what we just said:
Speech recognition, semantic understanding: let digital people hear and understand the needs of users;
Speech synthesis: let digital people express themselves in the form of language;
Image recognition: let digital people know what users are doing;
3D drive: Help digital people to make some body language and expression movements in addition to making sounds, so that digital people look more real.
In this track, the startups we are optimistic about need to be able to make good digital people apps. The first is the technical level: the company needs to develop an efficient technical solution that allows digital people to be "smart" enough to understand and solve my problems. The second is the artistic level, where digital people need to have an attractive image.
7. Secondary culture
Next, I would like to introduce a virtual idol operator that is shunned to invest in the early stage: The Second Generation Culture. (Shunwei Capital led the Second World Culture Pre-A round of financing in 2020 and continued to support it in subsequent rounds of financing.) ）
The core competency of the second-generation culture is to design and operate an IP avatar and use the star to expand its influence, while also helping the star to undertake more business cooperation. The first star virtual person produced by the second generation culture is the avatar "Di Li LengBa" created for Di Li Reba.
Initially, the expectations of the second generation culture for this project were mainly to increase the visibility of the company's case and obtain supplies in the flow of funds. However, in the process, it was gradually found that the market for celebrity avatars was larger than expected, and there were demands from celebrities, fans and brands:
For celebrities, avatars can jump out of the limitations of artist identity and activity time, and show more ideas. For fans, you can see the different aspects of the star from the avatar, and the virtual image products launched in cooperation with 52TOYS, Twelve Buildings, etc., also meet the consumer demands of derivatives; for brands, compared with real people, the virtual image of the star can cooperate with marketing activities with high frequency and convenience, and innovate at the visual and content levels, and the restrictions are smaller.
At present, the celebrity avatar is already the most mature business line of the second generation culture, and it also has a relatively clear change scene. From the early stage of concept design, image production, to the later stage of operation planning, the second generation culture has formed a relatively stable set of processes.
The cooperation between the second generation and the stars has also developed a model of co-creation of IP from the early simple provision of solutions.
After Di Lieba, the second generation of culture also created his virtual brother "Taosman" with Huang Zitao. In the newly released cooperation with artists, the second world culture no longer simply designs the virtual image of the artist, but cooperates with Ouyang Nana's personal characteristics to create a virtual band. The copyright proceeds from these co-creation projects will be shared between Theatrical Culture and the Star Team.
This year, the second world of culture and Mofa Technology cooperation, launched the national style hyper-realistic avatar "Ling-Ling", she was set as a Chinese Peking Opera Meipai enthusiast, the appearance of the balance of the East and West for the aesthetic preferences of Chinese appearance, is a Dan phoenix-eyed, tall and thin young girl. At present, in the domestic market of 18-30 years old, there is a relative lack of young cultural inheritors with distinct national style temperament and international texture, and Ling is aiming at the blank of this market segment, and the target is the virtual supermodel Net red Lil Miquela on foreign Instagram.
In addition, The Second World Culture has also launched a virtual DJ "Purple", which specializes in electronic music, trend brands and future technologies, and she has reached cooperation with Owhat, Wonderland New Vision Magazine, etc., and participated in iQIYI's virtual talent show "Transdimensional Rising Star", which will appear more on the stage of offline music festivals and livehouses in the future.
To sum up, the core competitiveness of the second-generation culture is to create some interesting, multi-level, and good-looking images, and expand its influence through variety shows, videos and other forms.
VIII. Third Stage of Development:
Everyone has their own digital image
Next, I'll share with you the third stage of "Digital People": Everyone has their own digital image. The reason why this stage is placed behind the case sharing is because at present, this stage is still a very beautiful and very distant vision of the future.
The very popular movie "Ready Player One" in the past few years is a specific interpretation of digital people and the future scenarios of VR/AR technology. In the world of the future, it will be a 3D world; everyone will have a digital virtualized social avatar; we will interact with many intelligent digital people, or NPCs (Non Player Character), and finally, the combination of gaming and social.
Stills from Ready Player One
One of the very hot concepts in the current discussion of 5G is cloud gaming, the core of which is to drive 3D rendering in the cloud to generate graphics. A lot of the games we're seeing so far are where many players play together in the cloud, but from a digital person's perspective, cloud gaming may evolve to interact with digital people in the cloud. So we think that the concept of cloud gaming is only half right, and the real application in the future may be cloud digital people.
You can imagine that in the future, when we order food in a restaurant, it may be a digital person driven by the cloud, coming to your table and asking you what help you need and so on.
Stills from the movie "Out of Control Player"
Finally, to sum up, digital people have a very wide range of applications in 2D screens, and in the future 3D world, digital people will also become a very important role.
The most interesting thing is that the technology and experience gradually accumulated by digital human technology in the current development of 2D media can be completely migrated to the 3D world. When VR/AR devices gradually replace smartphones in the future, digital people will also complete the transition from 2D to 3D. So I think that the digital person is a "snowball" track, that is: there are applications in 2D media, there is a future in the 3D world, and its technology and image can be migrated.
Looking forward to the future of 3D worlds.