原文链接：https://vfxvoice.com/the-tools-shaping-the-future-of-vfx/
标题：THE TOOLS SHAPING THE FUTURE OF VFX
作者：Jim McCullaugh
来源：《VFX Voice》2026 年秋季刊，行业问答圆桌专题
时间：2026 年 10 月 6 日
翻译日期：2026 年 10 月 7 日
Original Caption: Wētā FX’s Kora fire simulation toolset – recognized with a VES Emerging Technology Award – powered the physics-based chemical combustion effects at the heart of Avatar: Fire and Ash, which won the Academy Award for Best Visual Effects. (Image courtesy of Wētā FX and 20th Century Studios)
图1说明：Wētā FX 的 Kora 火焰模拟工具集获 VES 新兴技术奖认可，为《阿凡达：火与烬》（Avatar: Fire and Ash）中的核心效果提供了基于物理的化学燃烧模拟。该片获得奥斯卡最佳视觉效果奖。（图片由 Wētā FX 和 20 世纪影业提供）
The future of visual effects will not hinge on a single breakthrough. It’s being assembled – tool by tool – inside the pipeline. Some of these tools accelerate the work. Others change it entirely. Most do both, often before the industry has time to fully absorb the impact.
视觉效果的未来不会取决于某一项突破。它正在制作管线内部，由一项项工具逐步构建。有些工具加快了工作速度，有些则彻底改变了工作方式。大多数兼具这两种作用，而且往往在行业尚未来得及充分理解其影响时，就已经发挥了作用。
AI may be the headline, but it’s only part of a broader shift. Real-time engines, virtual production, cloud workflows and increasingly intelligent systems are converging into a new kind of toolkit – one that compresses timelines, expands possibilities and shifts who holds creative control.
AI 或许最受关注，但它只是更广泛转变的一部分。实时引擎、虚拟制作、云端工作流和日益智能化的系统正在汇聚，形成一种新的工具集，缩短制作周期、拓展可能性，并改变创意控制权的归属。
We asked a group of VFX artists, supervisors, technologists and others in the VFX and AI ecosystem to identify the tools they believe will shape what comes next. Their answers reveal an industry moving faster than its own definitions, where the future is already in use.
我们请一批视效艺术家、总监、技术专家，以及视效与 AI 生态中的其他人士，指出他们认为将塑造下一阶段发展的工具。他们的回答呈现出一个发展快到已有定义来不及跟上的行业：未来已经开始投入使用。
Elizabeth Willaman, VFX Producer, Los Angeles
Elizabeth Willaman，视效制片人，洛杉矶
“As a Producer, regardless of any specific ML/AI tool, what I’m watching for is the time and money that using a tool saves the vendors, which, in turn, also saves the studios. Can it compress the schedule? Can it reduce change orders, or even eliminate the need to reshoot a missing plate? Traditional VFX vendors are already well underway with deepening their hybrid ML/AI pipelines, and this will keep accelerating as tools change.
“作为制片人，无论具体采用哪种机器学习或 AI 工具，我关注的都是它能为视效供应商节省多少时间和资金，进而也为制片公司带来节省。它能压缩制作排期吗？能减少制作变更单，甚至免去为缺失的素材底板补拍的需要吗？传统视效供应商已经在深化融合机器学习与 AI 的制作管线，随着工具变化，这一进程还会继续加速。
“What’s incredibly interesting to me is that alongside this, a new generation of GenAI-native studios is emerging. They never had our pipelines to begin with, which means they’re using these tools, but answering those same questions about time and money from a completely different starting point. Both are real. I’m paying attention to both.”
“让我格外感兴趣的是，与此同时，新一代生成式 AI 原生工作室正在出现。它们从一开始就没有我们这样的制作管线，因此虽然也在使用这些工具，却从完全不同的起点回答同样的时间和资金问题。两者都真实存在，我都在关注。”
Original Caption: This is what the future of filmmaking looks like from the inside. ILM’s StageCraft LED volume – shown here at full operational scale on Thor: Love and Thunder – has become the defining infrastructure of modern virtual production, with Unreal Engine rendering photorealistic environments in real time as cameras roll. The technology that debuted on The Mandalorian in 2019 is now a permanent fixture of the global production pipeline, and its influence on how stories get told is only accelerating. (Image courtesy of ILM)
图2说明：从内部看去，电影制作的未来就是这样。图中是《雷神4：爱与雷霆》（Thor: Love and Thunder）拍摄现场全面运作的工业光魔（ILM）StageCraft LED 虚拟影棚。它已成为现代虚拟制作的标志性基础设施，摄影机拍摄的同时，虚幻引擎实时渲染出照片级真实的环境。这项技术于 2019 年在《曼达洛人》（The Mandalorian）中首次亮相，如今已成为全球制作管线的固定组成部分，对故事讲述方式的影响还在不断增强。（图片由 ILM 提供）
Rob Bredow, Senior Vice President, Creative Innovation, Digital Production & Technology, ILM, San Francisco
Rob Bredow，创意创新、数字制作与技术高级副总裁，ILM，旧金山
“When I look at where our workflows are heading, I’m most interested in the evolution of art-directable machine learning. When you work with amazing artists, they already have a precise picture in their minds of what they are creating. The best innovations I’m seeing right now are ML tools that act as assistants – filling in the complex details in real-time as an artist draws, driven entirely by artistic choices rather than random generation. By letting these responsive algorithms handle the technical heavy lifting, I hope we can empower teams to do what they do best: focus their energy on storytelling, problem-solving and creative expression.”
“展望工作流的发展方向，我最感兴趣的是能够接受艺术指导的机器学习技术如何演进。出色的艺术家对自己要创作的内容，脑海中已经有了准确的图景。我现在看到的最佳创新，是充当助手的机器学习工具：艺术家绘制时，它们实时补充复杂细节，完全由艺术选择驱动，而非随机生成。我希望让这些能及时响应的算法承担繁重的技术工作，帮助团队发挥所长，把精力放在讲故事、解决问题和创意表达上。”
Mark O. Hammond, VFX Supervisor, MR. X, Toronto
Mark O. Hammond，视效总监，MR. X，多伦多
“The most meaningful shift I’m seeing is AI reducing friction in the production pipeline. Less time hunting for data or toggling between asset management tools, more time working on shots. Departments like rigging are a good early indicator – as the simpler foundational steps get automated, artists can build more sophisticated rigs faster. What won’t change is the need for creative problem-solvers at the top of whatever toolset comes next.”
“我看到的最有意义的变化，是 AI 减少了制作管线中的阻碍。花在查找数据或切换资产管理工具上的时间更少了，制作镜头的时间更多了。绑定等部门很好地体现了这一早期变化：较简单的基础步骤逐渐自动化，艺术家便能更快地构建更复杂的绑定系统。无论接下来出现什么工具集，始终都需要富有创意、善于解决问题的人来驾驭它们。”
Original Caption: The visionaries behind Avatar: Fire and Ash – Kora Fire Toolset, Alexey Dmitrievich Stomakhin, John Edholm, Murali Ramachari and Aleksandr Isakov, accepted the Emerging Technology Award at the 24th Annual VES Awards in February 2026. (Photo courtesy of Moloshok Photography)
图3说明：《阿凡达：火与烬》Kora 火焰工具集背后富有远见的研发者 Alexey Dmitrievich Stomakhin、John Edholm、Murali Ramachari 和 Aleksandr Isakov，于 2026 年 2 月举行的第 24 届 VES 年度颁奖典礼上领取新兴技术奖。（照片由 Moloshok Photography 提供）
Kimball Thurston, CTO, Wētā FX, New Zealand
Kimball Thurston，首席技术官，Wētā FX，新西兰
“The box office has performed strongly in the first half of the year, with several titles shaking up what will make a successful movie. Premium screens and in-person cinema experiences continue to play strongly, and there appears to be a trend of younger audiences wanting analog, in-person experiences – which is exciting for cinema lovers.
“今年上半年票房表现强劲，几部影片正在改变人们对电影成功要素的认识。高端影厅和到影院现场观影的体验继续表现出强大吸引力，年轻观众似乎也越来越想要非数字化、亲身参与的体验。这让电影爱好者感到振奋。
“The desire to ‘AI all the things’ will trigger companies to create new features or tout wholesale replacement of workflows. The challenge is that the AI push is largely driven as a cost-saving mechanism. In the short term, generative AI will likely not be used for final pixels, as governance, quality and token cost issues compound to restrict usage. Instead, the desired financial efficiencies are more likely to be found as AI assistance helps empower artists to collaborate better and iterate more rapidly – AI agents that find and fix errors, integrate with production tracking, or make turnovers between vendors more consistent.”
“‘把一切都 AI 化’的愿望会促使企业开发新功能，或宣称能够全面替代现有工作流。问题在于，推动 AI 应用的主要动力是节省成本。短期内，生成式 AI 很可能不会用于最终成片画面，因为治理、质量和 token 成本问题相互叠加，限制了它的使用。所期待的成本效益，更可能来自 AI 辅助：帮助艺术家更好地协作、更快地迭代，例如由 AI 智能体查找和修复错误、接入制作跟踪系统，或提高供应商之间素材交接的一致性。”
Hoon Kim, CEO, Beeble AI, Seoul
Hoon Kim，首席执行官，Beeble AI，首尔
“One of the most significant shifts in VFX is the transition from AI as a purely generative tool toward AI as a controllable production system. Early AI workflows prioritized novelty and speed, but professional productions require consistency, continuity across shots and precise creative direction. We are seeing hybrid workflows emerge where traditional compositing intersects with learned systems, allowing artists to relight scenes, transform environments and refine imagery after capture while maintaining creative intent.”
“视效领域最重要的变化之一，是 AI 从纯粹的生成工具转向可控的制作系统。早期 AI 工作流注重新奇和速度，但专业制作需要一致性、镜头之间的连贯性，以及准确的创意指导。我们看到混合工作流正在出现，传统合成与基于学习的系统相结合，让艺术家能够在拍摄后重新布光、改变环境、调整图像，同时保留创作意图。”
Original Caption: Houdini in action: Increasing the detail and resolution of pyro simulations using proprietary Houdini ML Model. (Image courtesy of SideFX)
图4说明：Houdini 应用实例：使用专有的 Houdini 机器学习模型，提高烟火模拟的细节和分辨率。（图片由 SideFX 提供）
Original Caption: ILM’s StageCraft technology – seen here on the set of The Creator – represents the most significant shift in how films are physically made since the introduction of digital photography. By wrapping practical sets in photorealistic LED environments powered by Unreal Engine, StageCraft puts the director, the actors and the camera inside the world of the film on the day of shooting – eliminating the creative and logistical gap between production and post that has defined filmmaking for decades. (Image courtesy of ILM)
图5说明：图中是《The Creator》拍摄现场使用的工业光魔 StageCraft 技术，它代表着数字摄影问世以来电影实际拍摄方式最重大的转变。StageCraft 利用虚幻引擎驱动的照片级真实 LED 环境环绕实景布置，让导演、演员和摄影机在拍摄当天就置身于影片世界之中，消除了数十年来一直存在于拍摄与后期之间的创意和组织协调隔阂。（图片由 ILM 提供）
Original Caption: Na’vi warrior in flight through a Kora-built sky – Wētā FX’s physics-based fire simulation toolset transforming the atmosphere of Avatar: Fire and Ash into a world where heat, ash and combustion are as much a part of the environment as water was in The Way of Water, Academy Award winner for Best Visual Effects. (Image courtesy of Wētā FX and 20th Century Studios)
图6说明：一名纳威战士飞过由 Kora 构建的天空。Wētā FX 基于物理的火焰模拟工具集，把《阿凡达：火与烬》的大气环境变成了一个热量、灰烬和燃烧都融入其中的世界，就如同水融入奥斯卡最佳视觉效果奖获奖影片《阿凡达：水之道》（The Way of Water）的环境一样。（图片由 Wētā FX 和 20 世纪影业提供）
Christy Anzelmo, Chief Product Officer, Foundry, London
Christy Anzelmo，首席产品官，Foundry，伦敦
“Cultivating an understanding of the fundamentals of visual storytelling remains as crucial when working with AI as it is for traditional VFX. Artists who understand cinematography, layout, color, lighting and composition can better direct the tools toward a quality result versus being guided by them. When this expertise is coupled with curiosity and problem-solving, artists can combine agentic assistance, fast AI visualization through Griptape, and refinement using traditional tools like Nuke to achieve their creative vision with both speed and control.”
“使用 AI 时，理解视觉叙事的基本原理，依然与传统视效制作中一样重要。懂得电影摄影、布局、色彩、布光和构图的艺术家，能更好地引导工具取得优质结果，而非被工具引导。当这种专业能力与好奇心、解决问题的能力相结合，艺术家就能把智能体辅助、通过 Griptape 进行的快速 AI 可视化，以及使用 Nuke 等传统工具进行的精细调整结合起来，兼顾速度与控制，实现自己的创作构想。”
Sebastien Bergeron, CEO, Pitch Black, Montreal
Sebastien Bergeron，首席执行官，Pitch Black，蒙特利尔
“AI has been a major focus in VFX for years, but what matters most are the tools that hold up in production. That experience evolved into REVIZE, our production-proven machine-learning capability used across our global studios. Originally developed at Rising Sun Pictures and used on delivered projects since 2022, the value isn’t just the technology itself – it’s how naturally it fits into existing workflows, expanding creative possibilities and giving us the confidence to hit delivery deadlines.”
“多年来，AI 一直是视效领域的重要关注点，但最关键的是工具能否经得起实际制作的检验。这些经验逐渐形成了 REVIZE，也就是我们在全球各工作室中使用、经过制作验证的机器学习能力。它最初由 Rising Sun Pictures 开发，自 2022 年起已用于完成交付的项目。它的价值不仅在于技术本身，更在于能够自然融入既有工作流，拓展创作可能性，并让我们有信心按期交付。”
Cristin Barghiel, Vice President, Product Development/R&D, SideFX, Toronto
Cristin Barghiel，产品开发与研发副总裁，SideFX，多伦多
“The most significant shift in recent versions of Houdini is the integration of machine learning into the very fabric of the software. By training neural networks on existing high-quality simulation data, Houdini’s ML tools can enhance the look and feel of physics behaviors in a fraction of the time – including the ability to up-res pyroclastic simulations in near real-time, allowing artists to iterate rapidly and explore many variations of smoke and fire in the time it once took to produce a single result.
“Houdini 近期版本最显著的变化，是机器学习融入了软件本身。通过利用已有的高质量模拟数据训练神经网络，Houdini 的机器学习工具只需过去一小部分的时间，就能改善物理行为的视觉表现和质感，包括以接近实时的速度提升火山碎屑模拟（pyroclastic simulations）的分辨率。这样一来，在过去只能产生一个结果的时间里，艺术家便能快速迭代，探索烟雾和火焰的多种变化。
“Houdini’s procedural approach continues to expand: KineFX brings procedural workflows to character rigging and animation; Solaris extends that philosophy into look development and lighting through OpenUSD; Copernicus introduces procedural image-processing directly inside Houdini. As ML becomes increasingly pervasive, these systems will combine AI-assisted workflows with procedural generation to accelerate creativity and production.”
“Houdini 的程序化方法还在不断扩展：KineFX 为角色绑定和动画引入程序化工作流；Solaris 通过 OpenUSD 将这一理念扩展到外观开发与布光；Copernicus 则把程序化图像处理直接引入 Houdini。随着机器学习的应用日益普遍，这些系统将把 AI 辅助工作流与程序化生成相结合，加快创作和制作。”
Gary Mundell, CEO, LOFTAPPS, San Francisco
Gary Mundell，首席执行官，LOFTAPPS，旧金山
“What feels most impactful is how AI can open up the creative process for artists who are not necessarily technical. With LOFTai, I’m interested in shortening the distance between the idea and the image. The most immediate seismic shift will be in animation – where AI can dramatically change how ideas are developed, iterated and brought to screen at a fraction of traditional cost. The next big opportunity is not just making more images faster but building tools that help people make sense of everything they are creating.”
“我觉得最有影响力的，是 AI 如何让未必擅长技术的艺术家也更容易开展创作。在 LOFTai 上，我关心的是缩短想法与图像之间的距离。最先出现的剧变将发生在动画领域：AI 能以传统成本的一小部分，大幅改变创意的开发、迭代和呈现方式。接下来的重大机会，除了更快地产生更多图像，还在于开发工具，帮助人们理解和梳理自己创作的一切。”
Mathieu Raynault, Founder/Owner, Raynault, Montreal
Mathieu Raynault，创始人兼所有者，Raynault，蒙特利尔
“The future of our industry will be shaped less by VFX tools than by the tools used by directors, editors and cinematographers. AI will soon let them block layouts, animation and concepts directly in the cutting room or on set. But our role as VFX artists won’t fundamentally change – we’ll still be expected to bring creativity and artistic judgment to turn ideas into final imagery. The tools that matter aren’t the ones generating the most buzz. They’re the ones that shorten iteration cycles, reduce technical friction and improve communication with clients.”
“塑造行业未来的，与其说是视效工具，不如说是导演、剪辑师和摄影师使用的工具。AI 很快就能让他们直接在剪辑室或拍摄现场，初步安排布局、动画和概念方案。但我们作为视效艺术家的角色不会发生根本变化：我们仍需运用创造力和艺术判断，把想法变成最终画面。真正重要的工具，是那些能缩短迭代周期、减少技术阻碍、改善与客户沟通的工具，而非引起最多热议的工具。”
Chris Barnett, Entrepreneur/Technologist, Christie Digital/Virtual Production, Cypress, California
Chris Barnett，企业家兼技术专家，Christie Digital／虚拟制作，加利福尼亚州赛普里斯（Cypress）
“Virtual production is evolving fast, but the right tool depends on the project. Christie’s projection-based approach – Virtual Projection – offers a compelling alternative to LED stages, particularly where budget and image quality both matter. High-quality RGBH projection creates large, seamless backgrounds across curved and unconventional surfaces while preserving color accuracy, eliminating the red-lobster color cast common in LED environments, and removing moiré artifacts entirely. Christie sees projection as part of a broader hybrid virtual production ecosystem – bridging the gap between practical sets and VFX.”
“虚拟制作发展很快，但合适的工具取决于项目。Christie 基于投影的 Virtual Projection 方案，为 LED 影棚提供了一个有吸引力的替代选择，尤其适合同时重视预算与画面质量的项目。高质量 RGBH 投影能在曲面和非常规表面上形成大幅无缝背景，同时保持色彩准确，消除 LED 环境中常见的‘红龙虾’式偏色，并完全去除摩尔纹。Christie 将投影视为更广泛的混合虚拟制作生态的一部分，连接实景布置与视效。”
Fernando Rivas-Manzaneque, Co-Founder & CEO, Volinga, Los Angeles
Fernando Rivas-Manzaneque，联合创始人兼首席执行官，Volinga，洛杉矶
“The most grounded way to think about AI’s current impact is not as a replacement for production, but as a layer of acceleration on top of it. Where things get genuinely exciting is in 3D Gaussian splatting – ML applied to 3D reconstruction. What used to require expensive photogrammetry pipelines can now be captured and integrated into virtual production workflows in a fraction of the time. Beyond capture, Gaussian splatting is becoming the key that unlocks generative AI in 3D environments – genuine three-dimensional spaces, not flat images or video. When that layer meets generative AI built on top of it, the result won’t just be faster pipelines. It will be a fundamentally different way of building worlds.”
“理解 AI 当前影响的一种最切合实际的方式，是把它看作叠加在制作之上的加速层，而非制作的替代品。真正令人兴奋的是三维高斯泼溅（3D Gaussian splatting），即把机器学习应用于三维重建。过去需要昂贵摄影测量管线完成的内容，如今只需原来一小部分的时间，就能完成采集并融入虚拟制作工作流。除采集之外，高斯泼溅也正在成为让生成式 AI 进入三维环境的关键。这些是真正的三维空间，而非平面图像或视频。当这一层与构建在其上的生成式 AI 相结合，结果将不只是更快的制作管线，还会是构建世界的方式发生根本变化。”
Jamie Hearing, DFX Supervisor, Clear Angle Studios, London
Jamie Hearing，数字视效总监（DFX Supervisor），Clear Angle Studios，伦敦
“The most exciting recent development for us has been the integration of Gaussian splatting and volumetric capture, built on top of our pre-existing Structure from Motion pipelines. For environments, this provides instant realism with a level of spatial awareness that previously required exhausting stages of post-processing – rich, layout-ready spaces delivered directly from capture to the VFX pipeline. For characters, volumetric capture arrays allow for temporally consistent data that remains entirely true to the original performance, providing instant access to spatial performance data without traditional rigging bottlenecks.”
“最近最让我们兴奋的进展，是在既有的运动恢复结构（Structure from Motion）管线基础上，将高斯泼溅与体积捕捉结合起来。对于环境，它能立即提供真实感与空间感知，以往达到这样的程度需要繁复的后期处理。采集完成后，丰富且可直接用于布局的空间就能进入视效管线。对于角色，体积捕捉阵列可以提供时间上连贯的数据，完全忠实于原始表演，让我们立即获得表演的空间数据，绕开传统绑定的瓶颈。”
Original Caption: ILM’s StageCraft LED volume – seen here on the set ofThor: Love and Thunder – has become one of the defining tools of modern VFX production, placing photorealistic digital environments directly behind actors in real time and eliminating the wall between shooting and post. Powered by Unreal Engine and proprietary camera tracking technology, StageCraft is no longer an emerging technology. It is the new standard – and its influence on how films and series are conceived, budgeted and shot will only deepen through 2026 and beyond. (Image courtesy of ILM)
图7说明：图中是《雷神4：爱与雷霆》拍摄现场的工业光魔 StageCraft LED 虚拟影棚。它已成为现代视效制作的标志性工具之一，将照片级真实的数字环境实时呈现在演员身后，消除了拍摄与后期之间的壁垒。StageCraft 由虚幻引擎和专有摄影机跟踪技术驱动，已不再是一项新兴技术，而是新的标准。它对电影和剧集的构思、预算编制和拍摄方式的影响，在 2026 年及以后只会继续加深。（图片由 ILM 提供）
Original Caption: The atmospheric haze, combustion layers and ember behavior in this sequence from Avatar: Fire and Ash are the work of Kora – Wētā FX’s physics-based fire simulation toolset and winner of the VES Emerging Technology Award – which powered more than 1,000 digital fire shots across the film. (Image courtesy of Wētā FX and 20th Century Studios)
图8说明：《阿凡达：火与烬》这一段落中的大气薄雾、燃烧层次和余烬运动，都由 Kora 实现。Kora 是 Wētā FX 基于物理的火焰模拟工具集，获得了 VES 新兴技术奖，并为全片超过 1,000 个数字火焰镜头提供了支持。（图片由 Wētā FX 和 20 世纪影业提供）
Original Caption: Beeble Canvas: This is what the future of compositing looks like on a laptop. Beeble Canvas combines AI video models, node-based workflow automation and traditional compositing tools in one browser-based environment – shown here using the SwitchX model to transform live-action footage with full alpha, prompt and reference control. No render farm required. No cloud upload. Just the shot, the model and the artist’s intent. (Image courtesy of Beeble AI)
图9说明：Beeble Canvas：笔记本电脑上的合成未来就是这样。Beeble Canvas 在一个基于浏览器的环境中，结合了 AI 视频模型、节点式工作流自动化和传统合成工具。图中使用 SwitchX 模型处理实拍素材，可全面控制 alpha 通道、提示词和参考内容。无需渲染农场，也无需上传云端，只有镜头、模型和艺术家的创作意图。（图片由 Beeble AI 提供）
Jase Lindgren, Principal User Advocate, Perforce, Minneapolis
Jase Lindgren，首席用户倡导专员，Perforce，明尼阿波利斯
“I spent more than a decade in VFX before joining Perforce, and what I notice now is how quickly AI tools have moved from novelty to daily use – and how little the asset side of our pipelines has changed to match. It’s the leakiest era of asset management VFX has ever seen. The most impactful innovation of the next phase isn’t another generative model – it’s the layers underneath. The studios that pull ahead won’t be the ones with the best models. They’ll be the ones who treat every output, the one that shipped and the 10 that didn’t, as part of their creative library.”
“加入 Perforce 前，我在视效行业工作了十多年。现在我注意到，AI 工具已经迅速从新奇事物变成日常工具，但制作管线中的资产环节为适应这种变化所作的调整却很少。这是视效资产管理有史以来流失最严重的时期。下一阶段最有影响力的创新，将来自底层环节，而非另一个生成模型。取得领先的工作室，会是那些把每个产出都视为创意资源库一部分的工作室，包括交付出去的那一个，以及未交付的另外十个，而非拥有最佳模型的工作室。”
Original Caption: For creators already working on compatible Wacom Cintiq and Intuos Pro models, Wacom’s new Art Pen 2 gives their trusted tablets a fresh creative upgrade, adding more expressive control for drawing, editing, sketching and visual work without having to overhaul the tools they already know. The Wacom Art Pen 2 brings 360-degree barrel rotation to Wacom’s latest pen technology, allowing artists to shift the thickness, character and feel of a line simply by rotating the pen in their hand. (Image courtesy of Wacom)
图10说明：对于已使用兼容型号 Wacom Cintiq 和 Intuos Pro 的创作者，Wacom 新推出的 Art Pen 2 为信赖的数位板带来了新的创作升级。无需全面更换已有工具，就能在绘画、编辑、草图和视觉工作中获得更丰富的表现控制。Wacom Art Pen 2 将 360 度笔杆旋转功能引入 Wacom 最新的数位笔技术，让艺术家只需转动手中的笔，就能改变线条的粗细、特征和质感。（图片由 Wacom 提供）
Original Caption: Gaussian splat made in Volinga Suite of a Viking village film set captured in Iceland. (Image courtesy of Volinga)
图11说明：使用 Volinga Suite 制作的高斯泼溅表示，源自对冰岛一处维京村落电影布景的采集。（图片由 Volinga 提供）
Original Caption: Unreal Engine 5 remained the backbone of virtual production in 2026 – its Nanite virtualized geometry and Lumen global illumination now standard tools on LED volume stages from top-budget streaming series down to broadcast. UE5.7 and 5.8 updates have concentrated on performance, animation pipelines and the demands of large-scale open worlds – while Epic plans to launch Unreal Engine 6 by the end of 2026. (Image courtesy of Epic Games)
图12说明：2026 年，虚幻引擎 5 仍是虚拟制作的支柱。其 Nanite 虚拟化几何和 Lumen 全局光照，已成为 LED 虚拟影棚的标准工具，应用范围从高预算流媒体剧集延伸到广播电视制作。UE5.7 和 5.8 的更新着重于性能、动画管线和大规模开放世界的需求；与此同时，Epic 计划在 2026 年底前推出虚幻引擎 6。（图片由 Epic Games 提供）
David McGavran, CEO, Maxon, Bad Homburg, Germany/ Thousand Oaks, California
David McGavran，首席执行官，Maxon，德国巴特洪堡／加利福尼亚州千橡市
“The next phase of VFX will be shaped by tools that make production faster, more connected and more artist-driven. Real-time simulation, faster rendering, smarter asset workflows and more native 3D pipelines are giving artists and smaller teams the ability to take more of a project from concept through final output without losing creative control. AI-assisted tools are becoming valuable where they remove friction – denoising, upscaling, rotoscoping, depth maps, smarter search – rather than where they try to replace the artist. The future is not about pushing creators out of the process; it is about helping them work faster, iterate more freely and preserve creative intent.”
“视效的下一阶段，将由能让制作更快、连接更紧密、更加由艺术家主导的工具塑造。实时模拟、更快的渲染、更智能的资产工作流，以及更多原生三维管线，让艺术家和较小团队有能力在保留创意控制的同时，承担一个项目从概念到最终输出的更多环节。AI 辅助工具的价值，正在降噪、提升分辨率、逐帧描摹、深度图和更智能的搜索等减少阻碍的环节中显现，而非体现在试图替代艺术家的地方。未来要帮助创作者更快地工作、更自由地迭代，并保留创作意图，而不是把他们排除在过程之外。”
Benjamin Field, Co-Founder & CEO, Deep Fusion, Cardiff, Wales
Benjamin Field，联合创始人兼首席执行官，Deep Fusion，威尔士卡迪夫
“What is VFX, exactly? At its core, it’s a trick – a visual tool designed to immerse an audience in a world or a story. Nothing specifies how that trick has to be done. The craft and skill are human. The technology has always driven the execution. AI is the latest evolution of that relationship, not a break from it. The more consequential shift is economic: if AI can reduce VFX budgets by even a significant fraction, more productions become financially viable – which means more work, not less. The real promise of AI efficiency isn’t replacement – it’s restoration of sustainable margins in an industry that has priced itself out of its own business.”
“视效究竟是什么？从根本上说，它是一种视觉把戏，一种让观众沉浸于世界或故事的视觉工具。并没有规定这种把戏必须怎样实现。技艺和技能来自人，技术则一直推动着实现方式。AI 是这种关系的最新演进，并未打破它。更重要的变化发生在经济层面：如果 AI 哪怕能让视效预算降低相当一部分，更多项目就能在财务上具备可行性，这意味着工作会增加，而非减少。AI 提升效率的真正潜力，在于让一个因自身定价而难以维系业务的行业恢复可持续的利润空间，而非替代人员。”
Zeev Farbman, CEO, LTX, Jerusalem
Zeev Farbman，首席执行官，LTX，耶路撒冷
“Every major creative tool goes through the same three stages before it becomes part of how an industry works. The first is novelty – early adopters push tools to the extreme, often treating them as separate from the creative process. After years of testing, we’ve reached the adoption phase: industries are shifting their use of AI from demos or brief scenes to entire films, pre- and post-production and visual effects. For AI video to earn a permanent place in production, it has to meet professional standards. Motion needs to be stable across frames, timing and sound need to stay aligned, and outputs need to be usable in post-production environments where artists expect precision. That’s what we’re building at LTX – designing models for production workflows rather than isolated demos.”
“每一种重要的创作工具，都要经历同样的三个阶段，才会成为行业运作方式的一部分。首先是新奇阶段：早期使用者把工具推向极限，往往将它们与创作过程分开看待。经过多年测试，我们已进入采用阶段：各行业正把 AI 的应用从演示或短场景扩展到整部电影、前期、后期和视觉效果。要让 AI 视频在制作中占据长期位置，就必须符合专业标准。运动在帧间需要保持稳定，时序与声音需要保持对齐，输出也必须能用于后期制作环境，因为这里的艺术家要求精确。这正是我们在 LTX 构建的东西：为制作工作流设计模型，而非只做独立演示。”
Dan May, President, Blackmagic Design, Australia, U.S., U.K., Asia
Dan May，总裁，Blackmagic Design，澳大利亚、美国、英国、亚洲
“The ability to collaborate and truly work together in real-time across VFX and the rest of post is a game changer that will continue to shape the future of the VFX industry. With tight timeframes and tighter budgets, linear workflows no longer work for post – different departments need to be able to work together, in the same project, in the same software, at the same time. Whether live-syncing projects across VFX, color, audio and editorial; sharing media globally across different VFX houses; or being able to make changes live while remotely receiving client feedback; cloud-based collaboration is the key to working together in an environment where creativity drives the decisions, not technological limitations. That being said, not all cloud-based collaboration solutions are created equal. To protect their work and IP, it’s important for VFX artists and post houses to select solutions that keep their media private and do not use it for AI training.”
“视效与后期制作其他环节之间能够实时协作、真正共同工作，会改变行业格局，并继续塑造视效行业的未来。在时间紧、预算更紧的情况下，线性工作流已不适合后期制作。不同部门需要能够同时在同一个项目、同一款软件中共同工作。无论是在视效、调色、音频和剪辑之间实时同步项目，在世界各地不同视效公司之间共享媒体素材，还是远程接收客户反馈的同时实时修改，云端协作都是共同工作的关键，让创意推动决策，而非受技术限制左右。不过，各种云端协作方案并不相同。为了保护作品和知识产权，视效艺术家与后期公司应选择能够让媒体素材保持私密、且不会将其用于 AI 训练的方案。”
Sarah Watling, Co-Founder & CEO of JALI Research, Toronto
Sarah Watling，联合创始人兼首席执行官，JALI Research，多伦多
“As character-driven storytelling expands across film, games and immersive media, one area of growing importance is the ability to deliver high-fidelity character performance – particularly facial animation and multilingual dialogue – at scale, without losing directability. Emerging audience engagement opportunities need expressive, believable characters that can operate consistently across platforms, languages and interactive contexts, often within real-time or near-real-time constraints. This is driving a shift toward more integrated workflows – where capture, creation, animation and interaction increasingly operate as part of a continuous system rather than discrete stages. Within these environments, performance is no longer a downstream output, but a foundational layer that must remain coherent, adaptable and responsive – supporting both authored storytelling and live or participant-driven experiences. As immersive formats evolve, the demand for subtle facial nuance in digital characters and avatars will only grow, reinforcing the need for systems that allow artists to direct performance with precision while sustaining consistency at scale.”
“随着以角色为中心的叙事向电影、游戏和沉浸式媒体扩展，规模化实现高保真角色表演，同时保留指导和控制能力，变得越来越重要，尤其是面部动画和多语言对白。新出现的观众互动机会，需要富有表现力、可信的角色，能够在不同平台、语言和交互情境中保持一致，而且往往需要满足实时或接近实时的要求。这正推动工作流进一步整合，让捕捉、创作、动画和交互逐渐成为连续系统的组成部分，不再各自作为独立阶段。在这些环境中，表演已不再只是下游输出，而是基础层：它必须保持连贯、适应性和响应能力，既支持预先创作的叙事，也支持现场或参与者驱动的体验。随着沉浸式形式发展，对数字角色与化身细微面部表现的需求只会增加，也更加需要这样的系统：既让艺术家准确指导表演，又能在规模化制作中保持一致性。”
Mohsen Mousavi, VP, Visual Pioneering, Eyeline, Vancouver
Mohsen Mousavi，视觉创新副总裁，Eyeline，温哥华
“At Eyeline, we see the future of VFX not in automated prompts, but in pioneering cutting-edge techniques that enhance the artistry of traditional filmmaking. We treat generative AI as a new creative tool, and one that can only be guided by our artists who have been honing their unique expertise for decades. The heart of our philosophy centers on a deep respect for what we call the ‘sacred relationship’ between directors and actors: human performance, emotion and identity remain the centerpiece of every project. So, when we choose to leverage these emerging tools, it’s to enable actors and directors to work together effectively with production-ready AI systems that prioritize high fidelity, reliability and precise creative control.”
“在 Eyeline，我们认为视效的未来在于开拓前沿技术，提升传统电影制作的艺术表现，而非自动化提示词。我们将生成式 AI 视为一种新的创作工具，而且这种工具只能由我们的艺术家来指导，他们已经磨炼独有的专业技能数十年。我们的核心理念，是深深尊重导演与演员之间那种我们称为‘神圣关系’的联系：人的表演、情感与身份，始终是每个项目的中心。因此，当我们选择运用这些新兴工具时，目的是让演员与导演能够借助可投入制作的 AI 系统有效合作，而这些系统将高保真、可靠性和准确的创意控制放在首位。”
Original Caption: GSplat generated with Houdini 22: A furry cat, with eye reflections and refractions all properly captured. (Image courtesy of SideFX)
图13说明：使用 Houdini 22 生成的 GSplat：一只毛茸茸的猫，眼睛的反射与折射都得到了准确捕捉。（图片由 SideFX 提供）
Original Caption: A side-by-side comparison of OpenVDB pyro and its GSplat equivalent. Artists can notice how much lighter the GSplat is – yet how closely it matches the quality of the VDB. (Image courtesy of SideFX)
图14说明：OpenVDB 烟火模拟与对应 GSplat 的并排比较。艺术家可以看到 GSplat 轻量了多少，同时其画质又与 VDB 十分接近。（图片由 SideFX 提供）
Original Caption: A cloud GSplat generated from an OpenVDB asset. (Image courtesy of SideFX)
图15说明：由 OpenVDB 资产生成的云朵 GSplat。（图片由 SideFX 提供）
Original Caption: Achieving cinematic quality with Neural Layer in Autodesk Flow Studio. (Image courtesy of Autodesk)
图16说明：使用 Autodesk Flow Studio 中的 Neural Layer 实现电影级画质。（图片由 Autodesk 提供）
Original Caption: Houdini in action: ML training for Gaussian splat generation using Houdini’s Karma renderer and Task Operators (TOPs). (Image courtesy of SideFX)
图17说明：Houdini 应用实例：使用 Houdini 的 Karma 渲染器和任务算子（Task Operators，TOPs），开展用于生成高斯泼溅的机器学习训练。（图片由 SideFX 提供）
Original Caption: Gaussian splatting is one of the five shifts reshaping the VFX industry in 2026 – AI moving from denoising into pipeline-interior tasks, with Gaussian splatting enabling rapid reconstruction of photorealistic 3D environments from captured footage. What used to require expensive photogrammetry pipelines or full CG builds can now be captured, reconstructed and integrated into virtual production workflows in a fraction of the time – and Nuke 17.0 now enables artists to non-destructively mask and manipulate Gaussian splats within volumetric workflow. (Image courtesy of Foundry)
图18说明：高斯泼溅是 2026 年重塑视效行业的五大变化之一。AI 正从降噪走向制作管线内部的任务，高斯泼溅则让人们能够利用拍摄素材，快速重建照片级真实的三维环境。过去需要昂贵的摄影测量管线或完整 CG 构建才能完成的内容，如今只需原来一小部分的时间，就能完成采集、重建并融入虚拟制作工作流。Nuke 17.0 现在还支持艺术家在体积工作流中，对高斯泼溅进行非破坏性的遮罩处理和操控。（图片由 Foundry 提供）
Original Caption: Nuke 17: Support for Gaussian splats, enabling artists to import, view, manipulate, render and export Gaussian splats opens up new environment workflows for set extensions and matte painting. Griptape, acquired by Foundry and integrating with Nuke, enables creation and management of secure, programmable AI agents that can support tasks across the entire pipeline – turning AI from an experiment into a controlled production instrument. (Image courtesy of Foundry)
图19说明：Nuke 17：支持高斯泼溅，让艺术家能够导入、查看、操控、渲染和导出高斯泼溅，为场景扩展和数字绘景开辟新的环境工作流。已被 Foundry 收购、正在与 Nuke 整合的 Griptape，支持创建和管理安全、可编程的 AI 智能体，辅助整个制作管线中的任务，让 AI 从实验转变为可控的制作工具。（图片由 Foundry 提供）
Neishaw Ali, Founding Partner, CEO & Executive Producer, SPIN VFX, Toronto
Neishaw Ali，创始合伙人、首席执行官兼执行制片人，SPIN VFX，多伦多
“For decades, our industry has been organized around separate sectors such as film, television, gaming, animation, visual effects, advertising and immersive media. Each had its own workflows, technology, talent and business models. That distinction is rapidly disappearing. We’re entering a screen-based content economy where content is no longer defined by where it’s viewed but by how efficiently it can move across platforms. The same characters, environments and digital assets can power a feature film, a streaming series, a game, a live event or an immersive experience. The value is no longer in creating assets once; instead, it’s in creating them once and using them everywhere.
“数十年来，我们的行业一直按电影、电视、游戏、动画、视觉效果、广告和沉浸式媒体等不同领域组织。每个领域都有自己的工作流、技术、人才和商业模式。这种区分正在迅速消失。我们正进入一种以屏幕为基础的内容经济：内容不再由观看地点定义，而由它能以多高的效率跨平台流转来定义。同样的角色、环境和数字资产，可以用于一部电影长片、一部流媒体剧集、一款游戏、一场现场活动或一种沉浸式体验。价值已不在于把资产做出来一次，而在于一次创建、处处使用。
“That requires a completely different production model. The traditional linear pipeline where departments work in isolation and hand work off one stage at a time is giving way to a connected ecosystem built on USD, real-time engines, AI tech and cloud collaboration. Production is becoming simultaneous rather than sequential. Artists, supervisors, clients, and technology are all interacting with the same data at the same time.
“这需要完全不同的制作模式。传统线性管线中，各部门各自工作，按阶段逐步交接；如今，它正在让位于由 USD、实时引擎、AI 技术和云端协作构建的互联生态。制作正从先后依次进行转向同时进行。艺术家、总监、客户和技术系统，都在同时与同一套数据交互。
“AI is accelerating this shift, but I don’t believe its greatest impact is replacing artists. Its real value is removing the repetitive work that prevents artists from doing what humans do best: making creative decisions, solving complex problems and telling compelling stories.
“AI 正在加快这一转变，但我不认为它最大的影响是替代艺术家。它真正的价值，是去除那些妨碍艺术家发挥人类所长的重复工作，让他们作出创意决策、解决复杂问题，并讲述引人入胜的故事。
“I believe the next generation of studios won’t be defined by whether they’re VFX studios, animation studios or game studios. They’ll be content technology companies that combine world-class artists with intelligent, interoperable pipelines capable of creating content for every screen. The competitive advantage will no longer come from having the most artists or even the newest technology. It will come from how well technology enhances human creativity, how quickly ideas move from concept to screen, and how effectively studios can adapt to a market that is changing faster than ever before. It isn’t an evolution of our business; it is a reinvention of it.”
“我相信，下一代工作室不会再由其属于视效、动画还是游戏工作室来定义。它们将成为内容技术公司，把世界一流的艺术家与智能、可互操作的制作管线结合起来，为各种屏幕创作内容。竞争优势将不再来自拥有最多的艺术家，甚至也不来自拥有最新的技术，而取决于技术如何提升人的创造力、想法从概念走向屏幕的速度，以及工作室应对这个变化比以往更快的市场的能力。这已不只是业务的演进，而是重新塑造我们的业务。”
Diana Colella, Executive Vice President of Media & Entertainment, Autodesk
Diana Colella，媒体与娱乐业务执行副总裁，Autodesk
“The industry conversation is shifting from ‘if’ we use AI to ‘how’ we can use it responsibly. This is why Autodesk is augmenting the VFX tools creators already use and trust – Maya, 3ds Max, Flame and Flow Studio – with AI-powered capabilities designed to empower artists, not replace them, helping them continue to push creative boundaries and tell more ambitious stories.”
“行业讨论正在从‘是否’使用 AI，转向‘如何’负责任地使用。这也是 Autodesk 为创作者已经使用并信任的视效工具 Maya、3ds Max、Flame 和 Flow Studio，增添 AI 能力的原因。这些能力旨在帮助艺术家，而非替代他们，让他们继续拓展创作边界，讲述更具雄心的故事。”
Dr. Yanpei Cao, Chief Scientist & Co-Founder, Tripo AI, Shanghai & San Francisco
Yanpei Cao 博士，首席科学家兼联合创始人，Tripo AI，上海与旧金山
“I believe the future of VFX will be shaped by a shift from editing shots and assets to authoring controllable worlds. Today, VFX already combines many forms of creation and transformation: building assets, modifying images and video, simulating motion, compositing layers and refining shots until they serve the story. The deeper transformation is that these elements will increasingly be created and controlled within dynamic, editable, simulation-ready generated worlds, rather than as disconnected assets or isolated frames. This can change the production process: previs, virtual production, layout, lighting, post and interactive experiences can begin to share the same underlying scene intelligence. That shift will not remove the need for artists; it will raise the importance of taste, direction and supervision. The next generation of VFX tools should help artists explore more ideas, preserve creative intent, and maintain control over geometry, motion, lighting, physics and style. In my view, the most valuable AI systems in VFX will not be black boxes that generate a final answer. They will be collaborators inside the production pipeline: fast, controllable, interoperable and grounded in the language of filmmaking. The future is not less craft; it’s craft amplified by much more powerful creative infrastructure.”
“我相信，从编辑镜头和资产转向创作可控世界，将塑造视效的未来。如今，视效已经结合了多种创作与变换方式：构建资产、修改图像和视频、模拟运动、合成图层，以及不断调整镜头，使其服务于故事。更深层的变化在于，这些元素将越来越多地在动态、可编辑、可供模拟的生成世界中被创建和控制，而非作为互不关联的资产或孤立画面存在。这可以改变制作过程：预演、虚拟制作、布局、布光、后期和交互体验，可以开始共享同一套底层场景智能。这一转变不会消除对艺术家的需求，反而会提高审美、指导和监督的重要性。下一代视效工具应帮助艺术家探索更多想法、保留创作意图，并掌控几何、运动、光照、物理和风格。在我看来，视效领域最有价值的 AI 系统，不会是生成最终答案的黑箱，而会成为制作管线内部的协作者：快速、可控、可互操作，并以电影制作的语言为基础。未来的技艺不会减少，而会借助强大得多的创作基础设施得到增强。”
Marco Ragozzino, Writer/Director/Visual Effects Producer, Los Angeles
Marco Ragozzino，编剧、导演、视效制片人，洛杉矶
“We’re witnessing the evolution of the VFX industry in, no pun intended, real time! I see the future of VFX as being a more iterative and malleable process thanks to gen AI tools incorporated into pipelines along with real-time rendering. Take Gaussian splats for example; they’re basically 3D environments generated from 2D photographic data and can be turned around quickly. This is exciting to me since I no longer have to rethink writing a scene that takes place in a cost-prohibitive location. G-Splats are like Rosco softdrops on steroids; when used in conjunction with a virtual production volume, I can shoot without the usual restrictions of a complicated location (such as an airport), retain the proper parallax and dynamic lighting, and make reasonable adjustments on the fly.
“我们正实时见证视效行业的演进，这里没有双关的意思！随着生成式 AI 工具与实时渲染一同融入制作管线，我认为未来的视效流程会更易迭代、更有可塑性。以高斯泼溅为例，它基本上是从二维摄影数据生成的三维环境，而且能快速完成。这让我感到兴奋，因为我不再需要因场地成本过高，而重新考虑是否写一场发生在那里的戏。高斯泼溅就像大幅增强版的 Rosco SoftDrop 实景背景幕。将它与虚拟制作影棚结合使用，我就能在拍摄时免受复杂场地，例如机场，通常带来的限制，保留正确的视差和动态光照，并在现场随时作出合理调整。
“I also believe that with advancements in performance capture, even more of the actors’ work will translate into the finished film, whether the character is fully digital or a hybrid. Reviewing a take with the actor(s), as their full digital form in real-time, helps not only dial in the performance, but allows more room to improvise, adjust blocking and shot composition, and incorporate new ideas without affecting the production schedule. This is all under the assumption that gen AI tools are incorporated responsibility, like assisting in tedious and repetitive tasks, so the creative teams can jump right into their artistic skill sets faster.”
“我还相信，随着表演捕捉技术进步，演员的更多表演都将呈现在最终影片中，无论角色是完全数字化的，还是混合形式的。与演员一起实时回看一条拍摄素材，并以其完整的数字形态呈现，不仅有助于调整表演，还能为即兴发挥、调整走位和镜头构图、融入新想法留下更多空间，而不影响制作排期。这一切的前提，是以负责任的方式把生成式 AI 工具引入制作，例如用它辅助繁琐、重复的任务，让创作团队能更快地运用自己的艺术专长。”
Original Caption: For IT: Welcome to Derry, the Live Action team at RodeoFX brought one of the most unsettling creatures to life – the Pickle Monster, a grotesque amalgam of decaying organic matter and writhing appendages that embodies the shapeshifting horror at the heart of Derry’s darkness. Working in close collaboration with the production’s practical effects team, RodeoFX seamlessly blended digital enhancements with on-set elements to create a creature that feels viscerally real and deeply wrong in equal measure. (Image courtesy of RodeoFX and HBO)
图20说明：在《小丑回魂：欢迎来到德里镇》（IT: Welcome to Derry）中，RodeoFX 的实拍团队让最令人不安的怪物之一活了起来。这只‘腌黄瓜怪物’（Pickle Monster）由腐败的有机物和扭动的附肢怪异地混合而成，体现了德里镇黑暗核心中能够变换形态的恐怖。RodeoFX 与剧组实物特效团队紧密合作，将数字增强与现场元素无缝融合，让这只怪物给人强烈的真实感，同时又令人深感不对劲。（图片由 RodeoFX 和 HBO 提供）
Original Caption: Fusion Studio 21 now includes more than 100 new motion graphics effects and tools, including updated USD support for 3D matte objects and textures, an integrated Krokodove VFX library, a new Macro Editor for developing custom effects and templates faster, and more. (Image courtesy of Blackmagic Design)
图21说明：Fusion Studio 21 现已加入超过 100 项新的动态图形效果和工具，包括更新后的 USD 支持，可用于三维遮罩物体和纹理；集成的 Krokodove 视效库；可更快开发自定义效果和模板的新宏编辑器（Macro Editor），以及其他功能。（图片由 Blackmagic Design 提供）
Original Caption: Houdini in action: Neural terrain generation using interactive paint brushes in Houdini. (Image courtesy of SideFX)
图22说明：Houdini 应用实例：使用 Houdini 中的交互式绘画笔刷进行神经网络地形生成。（图片由 SideFX 提供）
Jon Mason, CEO, Hotspring, U.K.
Jon Mason，首席执行官，Hotspring，英国
“As certain AI tools designed to do particular tasks get embedded in VFX workflows we will see savings in time and money and a general contraction and realignment of the production and VFX industry. Beyond that, we will see the birth of more AI production companies using AI-generated plates instead of shot material, then enhancing those plates using VFX. We will begin to see more of a mix of both shot content enhanced by AI and VFX alongside fully AI-generated content enhanced by VFX and some live-action elements. There will also be a premium on hand-crafted, human-centric skills like stop-frame animation, some CG work, and anything new and unique created by skilled craftspeople.”
“随着用于特定任务的某些 AI 工具嵌入视效工作流，我们将看到时间和资金的节省，以及制作与视效行业整体的收缩和调整。除此之外，还会出现更多 AI 制作公司，用 AI 生成的底板替代拍摄素材，再用视效强化这些底板。我们将开始看到更多混合内容：一方面是经 AI 和视效增强的拍摄内容，另一方面是完全由 AI 生成、再经视效和部分实拍元素增强的内容。同时，定格动画、某些 CG 工作，以及熟练创作者制作的任何新颖、独特内容等手工完成、以人为核心的技能，也将获得更高价值。”
Jon Wood, Head of VFX, Untold Studios, U.K.
Jon Wood，视效负责人，Untold Studios，英国
“The most important developments driving the next phase of the industry are the integration of AI into existing VFX toolsets and more controllable, production-ready workflows. These advances will allow artists to work faster while still maintaining the precision and control we require. Rather than standalone or overly simplified ‘one-click’ AI solutions, the real value comes from tools that integrate into established pipelines and preserve the level of control artists are used to in traditional workflows. Clients expect precise, pixel-level control, and simply regenerating the entire image with random results each time is not a viable approach.”
“推动行业下一阶段发展的最重要进展，是将 AI 融入既有视效工具集，以及更可控、能投入实际制作的工作流。这些进展将让艺术家工作得更快，同时保留我们所要求的精度和控制能力。真正的价值，来自能够融入成熟制作管线、并保留艺术家在传统工作流中所习惯的控制程度的工具，而非独立存在或过于简化的‘一键式’AI 方案。客户要求准确到像素级的控制，每次都随机生成整幅图像，并不是可行的做法。”
转载请注明来源。欢迎留言评论，欢迎对文章中的引用来源进行考证，欢迎指出任何有错误或不够清晰的表达。