
Titanic: Ship of Legends - A Pioneering Documentary Series with AI
In 2023-2024, our video production team embarked on an extraordinary journey to create
"Titanic: Ship of Legends," a compelling documentary series spanning 10 episodes. This series delves into the incredible stories of survival and tragedy aboard the Titanic, focusing to answer a profound psychological question: why did some individuals with seemingly no means survive against all odds, while others with wealth and advantages did not?

Overcoming Challenges with Innovation​
Recreating the scenes of the Titanic with the budget of a major feature film was beyond our reach. Similarly, producing all the scenes for 10 episodes using 3D animation would have been too expensive and time-consuming. Initially, we planned to use artist-created drawings enhanced with 2D animations, a technique long employed by our colleagues in the documentary film industry. However, during the production of the first episode, we faced an unexpected setback when our artist disappeared, leaving us with only a few incomplete drawings. With the deadline looming and no other artists available to meet our timeline, we urgently sought alternative solutions.
At this critical juncture, new updates from Midjourney and Kaiber.AI were generating significant online buzz. We decided to experiment with Midjourney to visualize some Titanic scenes. Although the results were not perfect or historically accurate, they captured the atmosphere and terror of the event in a fascinating way as well as realistic images of people looking very convincing.


​While the AI-generated visuals may not seem as impressive now, 1.5 years later, given the rapid advancements in AI technology, they were groundbreaking at the time. We decided to blend these AI generations with real footage, primarily using them as b-roll and transitions between scenes.


Integrating AI and Traditional Techniques​
Taking a bold step, we decided to incorporate a mix of classic drawings, 2D and 3D animation, AI-generated images, and AI-animations into various episodes. Each episode focuses on a different individual, and we aimed to give each one a unique style and atmosphere.



As the series progressed over several months, AI technology advanced rapidly. By 2024, Runway ML was producing animations with realistic movement, comparable to high-quality 3D animations that typically require extensive time and resources. What once took days or weeks could now be accomplished in minutes.
A Blend of Reality and Imagination
The result is a documentary that seamlessly combines footage shot with professional cameras, featuring a presenter narrating stories, with b-roll and historical events recreated using a blend of images, animations, and archival materials. We filmed scenes in the Titanic museum and various locations in New York and other states connected to the Titanic's story. These scenes are interwoven with real photographs, newspapers, 2D and 3D animation, stock footage, drawings, and AI-generated content, creating a mesmerizing blend where the line between reality and generations blurs.
Cutting-Edge Tools and Techniques
Throughout the production, we utilized 8 AI-powered tools:
-
DaVinci Resolve for editing and color grading
-
Adobe After Effects for 2D animation, text, and transitions
-
Midjourney and Leonardo.AI for image generation
-
Kaiber.AI for flip-book style animations
-
Runway ML for realistic footage animations
-
Adobe Photoshop with Adobe Firefly for enhancing historical images
-
Topaz AI Video for upscaling images and animations to 4K resolution
-
Blender for 3D animation
A special mention goes to Adobe Photoshop, which played a crucial role in restoring and enhancing historical photographs. We used real, copyright-free photographs of people and events, many of which were low quality and did not fit the 16:9 film format. Adobe Firefly technology allowed us to add rº0ealistic environments around these images, seamlessly fitting them into the desired aspect ratio, significantly enhancing the film's visual cohesion.

Original:

With generated environment around the original image
With the release of Runway ML Gen-3, we achieved another breakthrough by animating historical photographs in a realistic manner, transforming old, low-quality images into stunning video footage. This innovation is a dream come true for documentary filmmakers.