The shape of virtual digital human is changeable, the classification is diverse, and the five modules form a general framework. Virtual digital human refers to the comprehensive product that exists in the non physical world, is created and used by computer means, and has multiple human characteristics (appearance characteristics, human performance ability, interaction ability, etc.). Virtual digital human can be divided according to personality symbol and graphic dimension, and can also be divided according to character graphic dimension. Character image, voice generation module, animation generation module, audio and video synthesis display module and interaction module constitute the general system framework of virtual digital human.
The virtual digital human market has gradually entered a mature period, and the commercialization process has accelerated. In 1982, the world’s first virtual singer, Lin Mingmei, was born. The virtual digital human industry has experienced four stages: germination, exploration, primary and growth. With the breakthrough of technology year by year, the manufacturing process has been effectively simplified, from the earliest manual drawing to the current CG and artificial intelligence synthesis. In the virtual digital human industry chain, the basic layer provides software and hardware support; The platform layer provides technology for production and development; The application layer provides solutions for application fields and enables the development of various fields. The entry threshold of virtual digital human industry is high, the differentiation is significant, and the audience group continues to expand. Therefore, the value of virtual digital human is prominent, the application field continues to expand, and it is expected to accelerate the commercialization process in the future.
The three core technologies of modeling, driving and rendering are the underlying architecture, and many technologies have practical applications. The mainstream technology of character modeling is still static scanning. Compared with static reconstruction technology, dynamic light field 3D reconstruction technology with high visual fidelity has become a key development direction in the future. In terms of driving technology, the intelligent synthesis of mouth action has been successfully applied to 2D and 3D virtual digital human. Motion capture scheme is the core technology of 3D digital human motion generation, which can achieve millimeter error at most. In terms of character rendering, the progress of PBR rendering technology and the emergence of new rendering technologies such as heavy lighting make the skin texture of virtual digital human real and break through the horror valley effect.
Virtual digital human has rich application scenarios and is the main entertainment and service industry. With the application and development of artificial intelligence, virtual digital human has expanded from early pan entertainment to finance, culture and tourism, medical treatment, retail and other fields, and is developing towards intelligence, portability, refinement and diversification. And realize the identity function of real human virtual avatar, and gradually realize commercial realization in the field of AI virtual digital human. It is recommended to focus on the sectors with strong synergy such as games, social networking, marketing and virtual idols: 1) RPG games are the focus of the game market, which has high requirements for the setting of open world and player roles, and leading manufacturers mostly grasp the mature virtual pinch technology; Engine manufacturers provide high-precision virtual human production platform for “water sellers” in the virtual digital human market. 2) In the next era, social networking focuses on creating a sense of immersion, and the demand for users’ virtual avatars has increased. 3) The virtual digital human KOL has widely appeared on the head platform, and its benefits (market, image and reputation) have been highlighted. 4) In 2020, the scale of China’s virtual idol core industry will be 3.46 billion yuan, yoy + 70.3%. It is expected to reach 6.22 billion yuan in 21 years. With the popularity and acceptance of virtual digital people, the virtual idol market will further expand.
Manufacturers at home and abroad layout the virtual digital human track from multiple angles. Chinese manufacturers focus on the application and development of the downstream of the industrial chain, while overseas manufacturers focus on the construction of the basic layer and the development of AI technology. China: Tencent, Alibaba, Netease and other comprehensive manufacturers support their relevant business units to make efforts to the track. Entertainment content focuses on VuP and Mango Excellent Media Co.Ltd(300413) (virtual host); Marketing operations focus on Bluefocus Intelligent Communications Group Co.Ltd(300058) , Lehua entertainment, Three’S Company Media Group Co.Ltd(605168) , Zhewen Interactive Group Co.Ltd(600986) , Hangzhou Anysoft Information Technology Co.Ltd(300571) , Xinhuanetco.Ltd(603888) , Beijing Jetsen Technology Co.Ltd(300182) , Inly Media Co.Ltd(603598) , Hylink Digital Solution Co.Ltd(603825) , etc; The foundation layer construction focuses on Iflytek Co.Ltd(002230) , Zhuoyi technology, Beijing Jetsen Technology Co.Ltd(300182) , Xinhuanetco.Ltd(603888) , Three’S Company Media Group Co.Ltd(605168) , etc. Overseas: AI technology focuses on soul machine, oben and room ai； Dynamic capture focuses on vicon, optitrack, xsens, etc; Rendering modeling focuses on epicgames, unity, NVIDIA, etc. Due to the increasing attention of the virtual digital human industry, manufacturers of relevant industries at home and abroad have obtained multiple rounds of financing in a short time, vigorously promoting the rapid development of the industry.
Investment suggestion: with the upsurge of meta universe, virtual digital human, as a first mover scene industry, has attracted much attention. “Five horizontal and two vertical” forms the basic technical architecture of virtual digital human, in which modeling dynamic light field technology, optical and inertial dynamic capture scheme, PBR and heavy lighting rendering are the core technologies, and unreal and unity are the main rendering engines. In terms of application scenarios, considering the improvement of the demand for digital avatars and the public’s acceptance of virtual digital people, focus on industries such as games, social networking, marketing and virtual idols. Overseas focus on upstream technology manufacturers in the industrial chain, while China recommends focusing on operators and downstream application manufacturers in the middle of the industrial chain.
Risk tip: the technological innovation and development of virtual digital human are not as expected; User acceptance is lower than expected; Industry policy supervision risk; Copyright management and protection risks; User privacy risk
