artificial intelligence technology is helping the domestic micro-short drama industry become high-quality.
Recently, the micro-short drama "Sanxingdui: Apocalypse of the Future" was unveiled at the "Exploring the Ancient Shu Civilization - Sanxingdui and Jinsha" exhibition held at the Grand Canal Museum in Beijing. The play is an online micro-short drama created and produced using generative artificial intelligence technology, showing the charm of ancient Shu civilization. Almost at the same time, the country's first AIGC original fantasy micro-short drama "Mountains and Seas: Splitting the Waves" held an offline viewing event. The scenes of the drama were completely produced by AI, constructing a bizarre ancient mythical world for the audience.
As of July 18, "Sanxingdui: Future Apocalypse", produced by Bona Pictures (001330.sz), co-produced by Douyin, and provided chief technical support by Douyin's Jimeng AI, has been online for 11 days with 10 episodes played. 24.807 million; the Xingmang short drama "Mountains and Seas Strange Mirror: Cutting Waves", which is provided with in-depth technical support by Kuaishou Keling AI, has been online for 5 episodes and 7.451 million views in 6 days.
aigc reshapes the video industry chain, and a new production method may emerge. Many practitioners told reporters that AIGC has a significant impact on the industry's operating model, cost structure, production cycle and other aspects. However, the technology is not yet fully mature and still requires manual participation. The industry needs to "catch up" first and then " Beyond”.
has poured into the ai micro-short drama track
. As participants, Bona, Douyin, and Kuaishou all attach great importance to the research and development and promotion of aigc technology.
Jiang Defu, CEO of Bona Pictures, told China Business News that under the current state of technological development, the effect of "Sanxingdui: Future Apocalypse" can be said to be "good", but technological development will never end, and there is always a way to improve. Process, if we wait for the technology to fully mature, we will always be in a "waiting" state, because the technology is always being updated and iterated. Therefore, Bona’s technical team is currently working overload in order to be the first in the industry to launch aigc generative short series. “You can’t wait until the end of the year to launch it, because by then new technologies will appear again, and your products will Falling behind again.”
Quji Xiaojiang, production director of Bona Pictures and head of Bona Aigms Production Center, said that in 2023 Bona decided to use AI technology to produce the series, and in 2024 it confirmed its cooperation with Jimeng AI.
In the middle of this year, Kuaishou put kolors, a large Vincent picture model, and kling, a large video generation model, into the market. Chen Kun, the director of "Mountains and Seas: Cutting the Waves", said in an interview that he made the first trailer of the film in 2023, and signed a cooperation with Kuaishou Xingmang Short Drama in the first half of 2024 to enter the production process.
In February this year, the Open Artificial Intelligence Research Center (openai) launched the Vincent video large model sora, triggering discussions about "AI subverting the traditional film and television industry."
Regarding the current impact of AI on the film and television industry, Qu Ji Xiaojiang said that he recognized the view that "without AI, there would be no productivity". The film and television industry needs to be prepared. Earth-shaking changes in the times have arrived. If you do not face it, you will definitely be left behind. behind.
According to the reporter's observation, although "Sanxingdui: Apocalypse of the Future" and "Mountains and Seas: Mirror of the Mountains and Waves" were both created by AIGC, the styles of the final films were different. The former has a stronger sense of story, logic and suspense; the latter has a straightforward script but a higher degree of picture completion.
Talking about script creation, Chen Kun said that in the early stages of development of AI film and television, it must choose the pain points of traditional film and television for creation. In film and television creation, especially in micro-short drama creation, the two major tracks of fantasy and science fiction are limited by the high cost and long cycle of CG special effects, and the development speed is obviously slow. However, audience demand for both tracks is overwhelming. The needs of the audience are the source of creation. The popularity of
video AI has directly brought changes to the film and television industry and practitioners. Chen Kun said that since the broadcast of "Mountains and Seas", many people have come to discuss cooperation, including content cooperation and financing cooperation.
Huajin Securities research report analysis, in the early stage of aigc video development, sora relied on its own video generation and editing capabilities to widen the gap with the original model through scaling law (scale law), pushing the industry into a new stage.aigc video reshapes the video industry chain. Based on the prompts of the generation tool, the interactive video production paradigm will reshape the traditional video production process, while promoting the development of the application layer and interaction layer, improving user experience and participation, and bringing benefits to the b-side and c-side to develop space.
In the current domestic market, Jimeng AI, a subsidiary of Douyin, supports the generation of images and videos through natural language and image input. It provides intelligent canvas, story creation mode, and AI editing capabilities such as first and last frames, lip sync, camera movement control, and speed control.
html In June, Kuaishou Keling's large video generation model was launched, using sora-like technology. The video resolution can be up to 1080p, the duration can be up to 2 minutes, and it provides functions such as first and last frame control, lens control, etc.; in July, Alibaba Damo The institute released the AIGC product light-seeking video creation platform to improve video production efficiency, solve video post-editing problems, and achieve precise control of video content through shot organization and video editing capabilities.Huajin Securities analysts believe that the application of AI in the entertainment industry has developed from simple automation tasks to content creation and personalized experiences. In terms of short video production, AI can significantly improve video production efficiency and reduce the time and cost of short drama production.
It will take time for the technology to mature
The benefits of choosing aigc to support film and television production are obvious - cycle shortening, cost reduction, and efficiency improvement.
Qu Ji Xiaojiang said that many projects were afraid to "move" in the past. On the one hand, the cost was very high. On the other hand, the current film market and audiences are also very picky. The cost of trial and error was too high. Unlike in the past, there was a general industry law. . But now with the help of AI, the company can "run two laps" at a lower cost first. If the audience's response is not enthusiastic enough or they are not interested in the IP, they can directly change projects and try again later.
Due to confidentiality clauses, Chen Kun cannot disclose the specific production cost amount, but he said that if compared with the traditional film and television production process, the cost of the current process is about less than a quarter of the past. In the specific process of
, aigc does not participate in the whole chain. Chen Kun said that for example, the early script creation was still done by people.
"At present, if we rely heavily on scripts written by AI, it will not meet our requirements." Chen Kun said that the specific links that will improve efficiency include early-stage shooting, costumes, lighting and stage design. Post-production still relies heavily on manual editing, which is not much different from the traditional production process.
Regarding the specific participation of AI, Chen Kun revealed that the film did not use Vincentian video technology, but Vincentian pictures and graphic videos. Because of the character consistency, scene consistency, and style consistency issues required for film and television drama production, the current Vincent video technology cannot yet achieve it. Specifically, if the video is generated directly from text, the multiple possibilities will be superimposed, and the controllability will become worse. Chen Kun also tried Vincent Video products from other manufacturers, but the final results were not satisfactory.
At a recent preview screening in theaters, Chen Kun said: "I am very nervous about what will be presented on the big screen today. Maybe 70 to 80 percent of the shots meet the minimum requirements of this screen. There are still 23 percent. Ten’s lens needs to be improved.”
Currently, there is a common problem in the AI video industry that is difficult to accurately simulate the physical principles of complex scenes. According to the reporter’s observation, “Sanxingdui: Future Apocalypse” also has this problem, such as the character’s mouth shape and pronunciation do not match, The body movements are not coordinated enough and the characters’ facial expressions are rigid. Cheng Weizhong, founder and CEO of
Zhongke Shenzhi, previously told reporters that all current AI video models have problems such as models that are unclear about the relationship between people and the environment, and unclear about the logical relationship between the front and back. This problem will cause the model to fail in practice. When applied, it cannot achieve the level of perfection envisioned by the user. The relevant person in charge of
Jimeng AI told reporters that AI iteration is very fast, and the maturity and stability of AIGC will be improved in the future. During this cooperation, Jimeng AI improved the "video generation" function, including supporting 24fps, 30fps, 60fps frame filling and the ability to double the super-resolution, adding horizontal movement of the lens, up and down movement, and support for the direction of lens movement. and amplitude control, etc.The person in charge of related technologies at
Keling also told reporters that judging from the perspective of technological development, it is possible to solve many of the problems that currently arise during video generation. However, the iterative process of Keling AI is not just a purely technical issue. It needs to be discussed with the creators to determine what product functions and forms the technology should be presented in in order to truly produce good works.
Chen Kun acknowledges the current popularity of video AI, but he believes that the current development of the industry is far from reaching the top of the mountain and is still on the mountainside. At present, there are many excellent works in the traditional film and television industry. Putting aside the time cost and capital cost, purely from the perspective of visual presentation effect, AI film and television films are not much better than traditional film and television works. Some special effects scenes are modeled, It can also be done by pasting materials, rendering drivers, etc., or even doing it better.
Therefore, Chen Kun is even more looking forward to the development of AI-native technology. When technology advances step by step, the first thing the industry must do is to "catch up." When the technology reaches a completely highly controllable and highly precise production effect, then we can do "transcendence". Of course, Chen Kun said that this will take some time, but not too long.
(This article comes from China Business News)
-
-
-
-
-
-
-
-
-
-