Youku's First AI-Assisted Special Effects TV Series 'Di Renjie of the Tang Dynasty' Premieres
-
Youku's newly launched TV series 'Di Renjie of the Tang Dynasty' has attracted attention because it is the platform's first to utilize artificial intelligence to assist in generating special effects scenes, especially in showcasing the Chang'an cityscape.
It is reported that the series employed AI technology to generate scenes such as crowds, streetscapes, and architecture, completely replacing traditional hand-drawn methods. Youku staff explained that AI technology ensures strict restoration of street widths, brick colors, and architectural styles. These buildings, environments, and crowds are all digitally created assets produced through artificial intelligence. "It's like building with Lego blocks," the staff said. "Various visual elements can be quickly assembled as needed. Pedestrians, buildings, even dense clouds and sunset glows can be freely added or removed - what you see is what you get." They revealed that nearly 20,000 digital assets were used for scene effects in 'Detective Dee', far exceeding the scale of conventional digital scenes.
This AI-based scene construction allows directors to make adjustments anytime and quickly realize their creative visions. Chief director Li Yunliang stated that this realistic effect enables him to freely zoom in and out during shooting without worrying about the "blurry distance shots and fake close-ups" common in traditional special effects. Unlike conventional effects, digital scenes allow directors to choose any shooting angle. Moreover, compared with traditional special effects, the production of AI digital assets is more efficient, more precise, and can be reused. A Youku staff member stated: "In the past, it required a significant amount of time and cost for manual production, but now with AI, automatic construction can be achieved. Work that previously took a month to complete can now be done in just ten days, increasing production efficiency by nearly 70%." With continuous algorithm optimization, production time will be further shortened.