At the 2024 Annual China Game Industry Conference held today, Giant Network unveiled the "QianYing" audio game generation model. This model consists of the game video generation model YingGame and the video dubbing model YingSound, aimed at lowering the barriers to game development, allowing ordinary people to easily create games.
YingGame is a video generation model designed for open-world games, jointly developed by Giant Network's AI Lab and Tsinghua University's SATLab. This model claims to be the first to achieve interactive control of diverse character actions, allowing users to customize game characters while offering improved game physics simulation features. This means that users can manipulate game characters like a "director," creating more vivid and interactive game scenes.
Complementing YingGame is the self-developed video dubbing model YingSound by Giant Network. This model was developed in collaboration with Giant Network's AI Lab, Northwestern Polytechnical University’s ASLP Lab, and Zhejiang University, supporting the generation of sound effects from video or video text combinations. YingSound features time alignment and video semantic understanding capabilities, enabling end-to-end generation of various types of sound effects for different kinds of videos, including real-life scenes, animation, and game videos, showcasing versatile application scenarios. This means that the sounds in game videos are no longer just simple background music, but complex sound effects that perfectly match the visuals and enhance immersion, injecting "soul" into the gaming experience.
Earlier this year, Giant Network Chairman Shi Yuzhu proposed that the company is exploring the creation of an AI game incubation platform, aimed at lowering the barriers to game development, allowing game enthusiasts without programming or development backgrounds to realize their ideas, enabling ordinary people to develop games. The release of the "QianYing" model is undoubtedly an important step towards this dream.
In April last year, Shi Yuzhu also stated that AI can serve as an auxiliary tool in the game development phase, enhancing efficiency and saving labor. However, he believes that the true essence of "Game + AI" lies in game planning. He pointed out that AI can help planners quickly find the best values, which is more accurate and efficient than relying on human intuition.
It is reported that Giant Network achieved operating revenue of 1.427 billion yuan in the first half of 2024, a year-on-year decrease of 1.13%; net profit attributable to the parent company was 718 million yuan, a year-on-year increase of 8.16%; and net profit excluding non-recurring items was 884 million yuan, a year-on-year increase of 27.92%. In the context of increasingly fierce competition in the gaming industry, Giant Network is actively investing in AI technology, which may bring new growth points.
The release of the "QianYing" model not only demonstrates Giant Network's strength in the field of AI technology but also brings new imaginative possibilities to the entire gaming industry. With AI empowering game creation, the dream of "everyone can develop games" may indeed become a reality. In the future, we may see more innovative game products driven by AI emerging, providing players with richer and more diverse experiences.