"Fantasy Rhapsody" - Project Description
"Fantasy Rhapsody" originates from an unconventional take on a traditional Western holiday, seeking to unleash a realm of imagination through a carnival-esque atmosphere. Visually, it employs a collage of fantasy and childlike wonder to deconstruct and reimagine familiar figures. Rather than following a linear narrative, the piece acts as a visual odyssey of flowing imagery—an improvised dreamscape set to a driving, rhythmic beat, celebrating the joy of breaking free from constraints.
AI Creation Workflow
Concept & Visual Development: In the initial phase, deep discussions were held with a Large Language Model (LLM) to establish the distinct theme of a "joyful and grotesque Halloween" and to draft the video's initial script framework. This process also involved exploring and ultimately defining "collage aesthetics" as the core visual language.
Core Visual Asset Generation: Visual materials were developed and planned using the Midjourney platform, which defined the film's overall color palette, texture, and aesthetic tone, ensuring a consistent visual style throughout the entire piece.
Dynamic Scene Generation: Selected static images were imported into Tiamat AI (即梦AI) to generate video from the images, animating them and bringing them to life.
Original Music Generation: Believing that music is the bridge to the vitality of the visuals, prompts were designed using an LLM. Based on these prompts, the original theme song was composed using Suno AI to perfectly match the mood and rhythm of the visuals.
Post-production & Final Assembly: Finally, the creator personally used CapCut (剪映) to perform the final editing, meticulously synchronizing the visuals with the music's beat (beat-matching), and integrating the subtitles. This stage integrated all the fragmented, AI-generated assets into a cohesive and complete work.
Built With
- capcut
- midjourney
- suno-ai
- tiamat-ai
Log in or sign up for Devpost to join the conversation.