在 1024 模力工场 · AI社区日主论坛上，复旦大学中文系硕士、奇点川越创始人程佳伟以《影视制作工具发展史：从胶片机械到生成式 AI，生成方式如何影响审美》为题作主题分享。
程佳伟指出，想让 AI 演出细腻的情绪，与其反复修改提示词，不如自己先演一遍。他分享了团队的一个实用方法：录下真人表演作为参考，让 AI 跟着处理微表情和动作，比单靠文字描述更稳定、更可控。
在他看来，如果只把 AI 当作省钱、省时间的工具，就低估了它对影视创作的影响：非科班创作者也有机会把脑中的镜头做出来，小团队能够承担更多制作任务，创意从构想到验证的距离正在缩短。
但生成变容易了，故事并不会自动变好，人物设定、角色成长和分镜脚本仍要提前想清楚，否则后面就容易跑偏、反复返工。
他介绍道，团队会先确定脚本、测试并固定提示词模板，再交给 Agent 批量生成素材，让人不必始终守着每一次生成。AI 的帮助也不止于“出画面”：音画同步生成可以减轻后期匹配音乐、音效的负担，剪辑软件中的 AI 功能也能处理部分细节工作，但专业影视剪辑目前仍不能完全交给 AI。
与此同时，人物长相保持一致，只解决了部分问题，不同镜头里的场景和空间关系能否接得上，仍是生成影像的一道难关。他还提醒，追求高质量内容不能只看画面效果，也要重视素材授权，审慎对待对他人画风、形象和声音的直接模仿，并保留创作与处理过程的记录。
程佳伟认为，工具技术改变的不只是制作速度，也会改变作品的表达方式，正如手持摄影让镜头走向街头、计算机让影像不再完全依赖实拍，AI 也在打开新的创作空间。但新方式出现，不代表旧方式必须退场：今天仍有人用胶片拍电影，AI 也不会让此前所有创作方式失去价值。
总而言之，工具决定了哪些画面容易拍、哪些效果做得到，也会影响创作者怎么讲故事，最终改变观众看到的影像风格。
- 胶片机械时代：机器笨重，画面相对固定。 摄影机不方便移动，创作者更多在固定场景里，通过演员的表演和场面调度展开故事。
- 设备轻量化之后：镜头走上街头，影像更自由。 手持拍摄带来了移动、抖动和贴近生活的视角。这些原本与设备有关的特点，也逐渐成为表达真实感、现场感的方式。
- 计算机与三维技术普及后：画面不再受实拍条件限制。 创作者可以通过建模、动画和渲染构造场景，让现实中难以拍到的世界出现在银幕上，拓展视觉想象。
- 生成式 AI 阶段：想法更容易变成画面，但细腻表达仍需控制。 非科班创作者也能尝试实现自己的审美，快速比较不同效果；但人物微表情、空间连续性仍难稳定控制。因此，他会用实拍表演作为参考，让 AI 生成更细腻、可信的影像。
想让 AI 演得像“老戏骨”，多写提示词不如先演一遍
非常感谢主办方的邀请，也谢谢大家。
毫无疑问，我们正处在一个生产极大丰富的时代。这种丰富不仅体现在物质层面，也体现在精神内容层面。每个人每天能够获取的内容实在太多了。
单说影视内容，我们可以看电影，也可以看剧；剧又有长剧、中剧、短剧，还有动画、动漫，以及 AI 漫剧。在座应该也有不少人正在看 AI 漫剧。
先简单介绍一下我们在做什么。这就绕不开 AI 漫剧这个话题。不过，我们并不是做漫剧的，也从来没有做过漫剧。
为什么还要谈它？因为有一个现象很有意思：AI 漫剧出现不过一两年，却很快从快速增长的新兴内容形态，进入了竞争激烈的存量市场。
这让我们不禁思考：AI 漫剧真的是 AIGC 内容发展的终点吗？AI 难道只能在这样的场景里发挥作用、进行创作吗？我认为不是。
所以，我们主要做两方面的工作。
一方面，探索漫剧之外的内容表达形式，寻找更容易被观众接受、也更适合传达创作意图的方式。另一方面，探索如何组织高效、稳定的内容生产流程，保证持续、高质量的输出。
这也对应了我的两个身份：我是一名导演，同时也是 AI 时代的开发者。
在这里，我特别想感谢程序员。你们让我们这样的文科生，也有机会通过语言，在计算机世界里表达自己的想法。对我们来说，这是一个很好的时代。
现在，很多人对 AIGC 的理解存在一种偏差，而且往往是低估了它。
这种情况在具体业务场景中尤其常见。无论作为甲方还是乙方，在沟通中，我们遇到的认知大多是：AI 是一个降本增效的工具。
但 AI 对这个时代的意义，真的只停留在降本增效吗？从我作为影视科班出身的创作者的角度来看，完全不是这样。
回到今天的主题，我们能否从历史的角度，看待 AIGC 在当下及未来可能带来的变革？这些想象并非没有依据，回顾历史就能发现一些很有意思的线索。
大家应该都有自己特别喜欢的作品，无论电影还是短剧。喜欢的原因可能很具体：导演调度好，场面安排精彩，美术出色，或者编剧写了一个深入人心、让人共情的故事。
但影视作品最终呈现出来的形态，真的只是由导演、编剧、摄影师决定的吗？
从更宏观的角度看，背后还有一个“幕后推手”，我把它叫作工具与技术。它们在更深的层面上，影响着影视作品的形态。
无论哪种影视生产，本质上都要回答三个问题：如何生产素材，如何组织素材，以及如何组织生产。也就是影像采集、影像加工，以及团队协作的问题。
本文由 AI宝库自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
At the 1024 AI Community Day Master ' s Forum, Cheng Jiawei, a graduate of the Chinese Department of the University of Jordan and founder of Kitagawa University, shared the theme " History of development of video production tools: from film mechanics to generation AI, how to influence aesthetics ".
Cheng Jia Wei pointed out that instead of changing prompt over and over again, it would be better to do it yourself. He shared a practical approach of the team: recording reality performances as a reference, allowing AI to follow micro-images and movements, more stable and manageable than textual descriptions alone.
In his view, if only AI was used as a tool to save money and time, its impact on video creation was underestimated: Non-scientific creators also have the opportunity to create shot in their minds, with small teams taking on more production tasks, and the distance between conception and validation of creativity is being reduced.
However, it becomes easier to generate stories that do not automatically change, and character creation, role growth and storyboard scripts still need to be thought out in advance, otherwise they are easy to run and return to work.
He explained that the team would first identify the script, test and fix the prompt template and then give it to Agent, which would produce asset in bulk, so that it would not have to keep an eye on each generation. AI helps more than "out of the picture": Synchronization of the sound can lighten the burden of post-production voiceover, sound, and the AI functionality in the editing software can handle some of the details, but the professional video clip is still not fully available to AI.
At the same time, the identity of the person is consistent and only partially resolved, and the possibility of a different scene in shot and of a spatial relationship remains a challenge in the generation of images. He also cautioned that the pursuit of high-quality content should not be limited to visual effects, but should also take into account the asset mandate, be careful about the direct imitation of the style, image and voice of others and keep records of the process of creation and processing.
Cheng Jia Wei believes that tool technology changes not only the speed of production, but also the way in which the work is expressed, just as hand-held photography takes shot to the streets, computers make images less fully dependent on reality, and AI is opening up new creative spaces. However, the emergence of a new way does not mean that the old way has to leave: Today there are still people making films on film, and AI won't make any of the previous ways of creating any value.
In short, the tools determine which images are easy to film and which effects can be achieved, and they influence how the creators tell stories and eventually change the image style that the audience sees.
- Age of film mechanics: machines are heavy and the picture is relatively fixed. Cameras are not easily mobile, and creators are more present in fixed scenes, spreading stories through actors ' performances and scenes.
- After the equipment was light quantization: shot went to the streets and the images were freer. Handheld photography brings a perspective of movement, tremors and proximity to life. These characteristics, which were originally related to the equipment, are also becoming a way of expressing a sense of truth and a sense of being present.
- When computers and three-dimensional technologies become available: the picture is no longer subject to the conditions of live filming. The creators can expand visual imagination by modelling, animation and replicating tectonic scenes so that the world that is difficult to capture in reality can appear on screen.
- Generating AI phase: Ideas are easier to translate into images, but fine expression still needs to be controlled. Non-scientific creators can also try to achieve their aesthetics, quickly comparing different effects; however, micro-impressions of people and space continuity remain difficult to control. So he'll use it as a reference to make AI more sophisticated and credible.
If you want AI to act like an old play, write more about prompt.
Thank you very much for the invitation and thank you very much.
There is no doubt that we are in an era of great productivity. This enrichment is reflected not only in material but also in spiritual content. Everyone has too much to go through every day.
We can watch movies and dramas alone; they have long dramas, melodramatics, short operas, animations, animations, and Al comics. There's probably a lot of people in this room watching the AI comic.
Let me briefly describe what we are doing. That's not going to get around the subject of the AI comic. But we don't do comics, and we never do.
Why talk about it? Because one of the interesting things is that, after only a year or two, Ai's comics quickly moved from a rapidly growing new form of content to a competitive stock market.
And that makes us wonder: Is it really the end of AIGC's content? AI, can it only work and create in this scenario? I don't think so.
So, we do two main things.
On the one hand, it explores the expression of content beyond the comics and the ways in which it is more acceptable to viewers and more appropriate to convey creative intentions. On the other hand, it explores how to organize efficient and stable content production processes to ensure continuous and high-quality output.
It also corresponds to my two identities: I'm a director, and I'm also an developer of the AI era.
Here, I would like to thank the programmer in particular. You give humanities like us the opportunity to express our thoughts through language in the computer world. This is a good time for us.
Now, there's a deviation in the understanding of AIGC, and it's often underestimated.
This is particularly common in the context of specific operations. Whether we're on the A or the B side, most of the perception we encounter in communication is that AI is a tool for reducing efficiency.
But does AI really mean anything to this era, just to reduce efficiency? From the point of view of me as the creator of the film class, that's not true.
Back to today's theme, can we look at the changes that AIGC could bring in the present and the future from a historical perspective? These imaginations are not without merit, and a review of history reveals some interesting leads.
You should all have your favorite work, both film and short play. The reasons for this may be very specific: the director is well-planned, well-organized, well-articulated, or he's written a deep, heart-friendly story.
But is it true that the final form of video work is determined by the director, the writer, the photographer?
From a more macro point of view, there's a "back-to-back" behind it, which I call tools and technologies. At a deeper level, they influence the shape of visual works.
In any video production, three questions are essentially answered: how to produce asset, how to organize asset and how to organize production. Image collection, image processing and teamwork.
Auto-fetched and translated by AI Gem Box. All rights belong to the original author.