跳到正文
Academy Software Foundation·· 8 天前AI 评分32

Rodeo FX 如何用 AI 辅助工程把开源审片工具 xSTUDIO 带入日常制作

The Last Mile: How Rodeo FX Brought xSTUDIO Into Daily Production

AI 导读

Rodeo FX 视效总监 Julien Hery 在 ASWF Open Source Days 上介绍,团队仅用两名工程师、基于 xSTUDIO 的插件架构而非分叉核心代码,将这款开源媒体播放与审片工具接入工作室日常制作流程。

正文 · 原文

At ASWF Open Source Days, Julien Hery of Rodeo FX shared how a small team used AI-assisted engineering and a plugin-based approach to bring an open source review tool into the studio’s daily production workflow.

Open source software gives studios a strong starting point. Getting it to work with the way a studio actually operates is another challenge.

In visual effects production, every studio has its own workflows, naming conventions, databases, review rooms, and technical requirements. An application can be excellent at what it does and still need considerable work before artists can use it comfortably every day.

That gap was the focus of the talk of Julien Hery at ASWF Open Source Days. As a visual effects supervisor at Rodeo FX, Hery explained how the studio adapted xStudio, an open source media playback and review application hosted at the ASWF, to fit its production needs without forking the core codebase.

The project offers a practical look at what it takes to bring open source tools into production, and how AI-assisted engineering can help a small team move faster without losing sight of quality.

Key Takeaways

  1. The last mile makes the difference. Open source tools provide a strong foundation, but studios need to adapt them to their own databases, naming conventions, and production workflows. At Rodeo FX, integrating these pieces into the review process helped make xStudio a practical tool for artists and supervisors working on a show.
  2. Build on what already works. Instead of forking xStudio and modifying its core, Rodeo FX used its plugin architecture to add studio-specific features, alongside two native C++ components. This kept the original application intact while allowing a small team of just two engineers to move quickly.
  3. Let production needs drive development. The team gathered feedback directly from artists, leads, and supervisors to identify what was slowing down daily work. By testing improvements in production and releasing updates frequently, they could fix bugs, respond to feedback, and deliver useful features while the needs were still fresh.
  4. Use AI with engineering discipline. AI helped accelerate coding, prototyping, refactoring, and repetitive tasks. However, speed alone wasn’t enough. The team documented rules based on past mistakes, created specialized agents grounded in the actual codebase and API documentation, and kept humans responsible for architectural decisions and testing.
  5. Test in real production. Putting new features into use quickly helped the team discover problems that prototypes alone might not reveal. The rapid pace also made thorough testing more challenging, particularly for UI behavior. The experience reinforced the importance of unit tests, sound architecture, and capturing lessons learned along the way.
  6. Give back to the community. While some integrations were specific to Rodeo FX, others could benefit studios more broadly. The team contributed a DeckLink broadcast plugin ported to macOS and a heads-up display plugin for onion-skin annotation. Sharing reusable improvements helps strengthen the open source tools that studios depend on.

The bigger takeaway

Rodeo FX’s experience shows that bringing open source software into production doesn’t always require a large engineering team or a heavily customized fork.

By building on xStudio’s plugin architecture, working directly with artists and supervisors, and using AI to accelerate implementation, two engineers were able to deliver improvements at a pace that kept up with production needs.

The approach still required engineering discipline, architectural decisions, and testing. AI made more work possible in less time, but the team had to establish the rules and practices that made that speed sustainable.

For studios looking to get more out of open source tools: start with a strong shared foundation, focus on the workflows that matter most to your artists, and contribute general-purpose improvements back to the community.

The last mile may be specific to each studio. The lessons learned along the way don’t have to be.


Thanks to Julien and the Rodeo FX team for a candid look at what adoption really takes. Follow ASWF on LinkedIn, Bluesky and X. 

来源:Academy Software Foundation · aswf.io