Maybe I Was Never Sure At All

Inspiration

This project was inspired by two concepts: existential uncertainty and the effect of digital presentation on personal narrative. We wanted to explore the quiet, internal moment when one questions identity, and simultaneously investigate how utilizing different visual frames and formats affects the story's emotional delivery in the social media era.

What it does

"Maybe I Was Never Sure At All" is a poetic, cinematic short film that visualizes a young man's internal journey of self-discovery and finding peace in uncertainty. It deliberately shifts between cinematic and social-media-style frames to analyze how platform aesthetics impact the story's emotional resonance and audience interpretation.

How we built it

This film was built on a resourceful blend of generative AI and traditional editing:

Visuals (Video Generation): We utilized Veo 3.1 Fast for all video content generation.

Narration (AI Voice): The philosophical internal monologue was created using ElevenLabs, lending a consistent, introspective voice to the persona.

Post-Production & Assembly: Filmora was used for final assembly, integrating the voiceover, text overlays, and managing the rhythmic coherence between the intentionally varied frame sizes and aspect ratios.

Challenges we ran into

The technical hurdles focused on achieving both visual and rhythmic consistency:

Visual Fidelity: We struggled with the Veo 3.1 Fast model's handling of complex human action. While elements like the cat were generated perfectly, scenes requiring intricate hand movements—specifically the matcha whisking and pouring—resulted in weird or inconsistent hand gestures, requiring extensive selective generation and editing.

Aesthetic Cohesion: Maintaining color tone and white balance was difficult due to the intentional use of different frame sizes and the underlying model's limitations, creating challenges in achieving rhythmic cohesion during the rapid cuts.

Accomplishments that we're proud of

We are proud of successfully blending a deep, philosophical narrative with an experimental approach to frame analysis. We managed to overcome challenging AI limitations (like poor hand generation) through creative editing, demonstrating how a single creator can produce a visually and emotionally resonant short film using accessible generative tools.

What we learned

We learned the critical need for post-production mastery when dealing with AI-generated human anatomy, especially with subtle actions like hand gestures. Furthermore, the project clarified the profound effect that different viewing "frames" (aspect ratios, color grading) have on narrative impact and how to intentionally use this for maximum emotional effect in the social media age.

What's next for Maybe I Was Never Sure At All

We plan to submit this film to festivals focused on experimental and short-form narratives. The insights gained from the frame and aesthetic analysis will be applied to our next projects, focusing on developing new techniques to generate consistent, high-fidelity human actions within generative AI video.

Built With

• elevenlabs
• english
• veo3.1