Reviving CG History: Enhancing Paul Debevec's Fiat Lux

Reviving CG History: Enhancing Paul Debevec's Fiat Lux

Fiat Lux: Enhancing a Piece of CG History

In this article, we will delve into the fascinating world of CG history and explore the possibilities of enhancing a video that has left a lasting impression on the industry. Join us as we take a trip down Memory Lane and uncover the techniques that revolutionized the way we approach lighting and CG. From compressing the original footage to exploring AI-based image enhancers, we will discuss the challenges and possibilities of creating a better version of this iconic video. While the road ahead may be filled with obstacles, we are determined to make the most of what we have and explore the potential of video-enhancing tools such as Video Enhanced AI. So, strap in as we embark on this journey to breathe new life into "Fiat Lux."

Table of Contents

  • Introduction: A Piece of CG History
  • The Brainchild: Paul de Becque's Fiat Lux
  • Enhancing the Video: Challenges and Roadblocks
  • AI-Based Image Enhancers: Pros and Cons
  • The Software: Video Enhanced AI
  • Previewing the Results
  • Fine-Tuning in Final Cut
  • Adding the Hazy Effect
  • Comparing Before and After
  • Adding Noise for a Natural Look
  • The Final Verdict: The Best Version
  • The YouTube Compression Challenge
  • A Separate 4K Version: Overcoming Copyright Claims
  • Final Thoughts and Invitation

Introduction: A Piece of CG History


"Fiat Lux" is a legendary video that made waves in the CG community almost 20 years ago. Its groundbreaking techniques and jaw-dropping results redefined the way we perceive lighting and CG. However, the only available high-quality copy is a compressed 720 by 480 pixel video. While it's better than nothing, it falls short of doing justice to such an iconic piece. This leads us to ponder the following question: Is there a way we can create a better version of this video?

The Brainchild: Paul de Becque's Fiat Lux


Before we dive into enhancing the video, let's take a quick history lesson. "Fiat Lux" is the brainchild of Paul de Becque, a visionary who, along with his team, pioneered techniques still used today. Their earlier short showcased how realistic results can be achieved by projecting photos and videos onto 3D geometry. Then came "Fiat Lux," which departed from regular CG lights and instead used high dynamic range photos for lighting a VCG environment. Combined with camera mapping, it yielded stunning results. Now that we know the origins of the video, let's explore how we can enhance it.

Enhancing the Video: Challenges and Roadblocks


Enhancing a low-resolution video like "Fiat Lux" comes with its fair share of challenges. Regular image resizing in Photoshop falls short of providing adequate results. The sheer lack of usable detail in the footage poses another roadblock to overcome. Furthermore, the original footage's quality compromises, such as poor geometry and interlaced frames, impede our progress. Despite these obstacles, we won't lose hope and will give it our best shot, utilizing the power of AI-based image enhancers.

AI-Based Image Enhancers: Pros and Cons


AI-based image enhancers show promise when it comes to improving images. However, they are primarily trained on real-world footage and may not work as effectively with CG elements. Considering the low-resolution nature of "Fiat Lux," extracting useful detail seems like an uphill battle. Additionally, the software we will be using, Video Enhanced AI, has limited options for enhancement and processing speed. While it might not be perfect, let's see how well it performs on the actual images of "Fiat Lux."

The Software: Video Enhanced AI


Our tool of choice for upscaling the video is Video Enhanced AI. Despite its poorly ported Mac version that relies solely on CPU processing, it possesses a crucial advantage: it has been trained not only with real-world footage but also with CG images. While the software's simplicity might limit our options in terms of enhancement and magnification, let's preview a couple of frames to assess its impact on the video artifacts and the quality of edges and surfaces.

Previewing the Results


Upon previewing the frames, we observe that Video Enhanced AI effectively removes video artifacts, offering smoother edges and surfaces. This improvement surpasses what regular image resizing in Photoshop can achieve. Notably, the software performs better in certain scenes, especially with architectural elements, creating a painterly effect. Although the level of detail in the original footage is limited, Video Enhanced AI does its best with what it has.

Fine-Tuning in Final Cut


Now that we have the enhanced image sequence after hours of processing, we move on to fine-tuning the footage in Final Cut. First, we overlay the original video to assess the perceptible improvement. While there may not be additional detail, the enhanced version appears cleaner and smoother, devoid of video artifacts. However, it seems slightly oversharpened, affecting the haze effect Present in the original video. To restore this effect, we need to recreate it in post and strike a balance between processed and natural-looking imagery.

Adding the Hazy Effect


To achieve the hazy effect, we duplicate the video sequence, apply a blur, and reduce the opacity. This creates the desired atmospheric haze, giving the footage its unique aesthetic. We employ two layers of haze: one with HEAVIER blur for the overall scene and another with a narrower blur to smooth out surrounding edges and lessen harsh painterly artifacts in the background. Balancing the opacity between 10 and 20 percent, we retain the original's Charm while elevating it with enhanced visuals.

Comparing Before and After


Let's compare some before-and-after images to gauge the impact of our enhancements. While not every sequence requires the blurring effect, some close-ups already possess the desired level of blur, retaining their original sharpness. By comparing the original video to the new footage, we witness visible improvements. The enhanced version, while not groundbreaking, presents a cleaner and sharper image overall, making it a significant enhancement.

Adding Noise for a Natural Look


To enhance the perceived sense of detail and lend a more natural appearance, we add noise to the footage. By treating each sequence differently, employing different noise levels for close-ups and water scenes, we mimic the characteristics of lenses used during shooting. While the effect may vary in different sequences, the addition of noise contributes to a more organic and pleasing visual experience.

The Final Verdict: The Best Version


After considering various factors, we determine that the hazy version without additional noise achieves the best outcome. While all three versions possess their merits, the hazy rendition strikes the right balance between the AI-enhanced image and the original video's aesthetics. Its softer, yet faithful, look resonates well with the visual language of "Fiat Lux." With this decision, we can confidently present our enhanced video to the world.

The YouTube Compression Challenge


Before sharing our enhancements with the world, we must address the challenge of YouTube compression. Uploading several versions of the video, we analyze the impact of different bit rates on image quality. After thorough testing, we find that a 56-megabit-per-Second bitrate strikes the best balance between size and quality. Furthermore, we determine that the hazy version without added noise fares better under YouTube's compression, retaining more fidelity and avoiding degradation.

A Separate 4K Version: Overcoming Copyright Claims


Due to copyright claims triggered by the usage of certain Music, we have prepared a separate 4K version of "Fiat Lux." This ensures that the YouTube algorithm won't bury this video, allowing more people to witness the process of enhancing the footage. You can find the link to the separate video in the description below and on the end card of this video.

Final Thoughts and Invitation


In conclusion, we have done our best to enhance the available footage of "Fiat Lux." While the results may not be earth-shattering, we invite you to be the judge and share your thoughts in the comments. We appreciate your support and hope you enjoy the final result of this journey to Revive a piece of CG history.



  1. Going back in time to explore the CG masterpiece "Fiat Lux"
  2. Unveiling Paul de Becque's revolutionary techniques in CG lighting
  3. Overcoming challenges in enhancing low-resolution footage
  4. Exploring the potential of AI-based image enhancers like Video Enhanced AI
  5. Fine-tuning the enhanced video in Final Cut with the hazy effect
  6. Comparing the before-and-after improvements in image quality
  7. Adding noise for a natural and detailed look
  8. Deciding on the best version and addressing the YouTube compression challenge
  9. Sharing a separate 4K version to overcome copyright claims
  10. Final thoughts and invitation for viewers to judge the enhanced result


Q: What is "Fiat Lux"?

"Fiat Lux" is a CG video that made waves in the industry almost 20 years ago. It showcased groundbreaking techniques in CG lighting and continues to inspire and influence artists today.

Q: Who is Paul de Becque?

Paul de Becque is the visionary behind "Fiat Lux". Along with his team, he pioneered techniques that are still used in the CG industry today. His innovative approach to lighting and CG redefined the way we perceive and create visuals.

Q: What challenges were faced in enhancing the video?

Enhancing a low-resolution video like "Fiat Lux" poses several challenges. The lack of detail in the footage, poor geometry, and interlaced frames were among the roadblocks that needed to be overcome.

Q: Which software was used for enhancing the video?

The software used for upscaling and enhancing the video was Video Enhanced AI. Although it had limitations in terms of enhancement options, it proved effective in removing artifacts and improving the overall image quality.

Q: Was noise added to the enhanced video?

Yes, noise was added to the enhanced video to enhance the perceived sense of detail and make it look more natural. Different noise levels were applied to different sequences to replicate the characteristics of different lenses used during shooting.

Q: How did the enhanced video fare on YouTube?

The enhanced video was uploaded to YouTube, taking into account the compression challenges. After analyzing different bitrates, it was found that a 56-megabit-per-second bitrate struck the best balance between size and quality. The hazy version without added noise performed better under YouTube's compression, maintaining fidelity and avoiding degradation.

Q: How can I watch the final enhanced version of "Fiat Lux"?

You can access the final enhanced version of "Fiat Lux" in the link provided in the description and end card of this video. Enjoy the journey of reviving a piece of CG history!

Find AI tools in Toolify

Join TOOLIFY to find the ai tools

Get started

Sign Up
App rating
AI Tools
Trusted Users
No complicated
No difficulty
Free forever
Browse More Content