GPT Image 2+Seedance Prompt Case Study
DavidmComfort · x · 2026-07-14
This share features a prompt case study using GPT Image 2 + Seedance 2.0 for a contemporary performance piece involving live singing and choreography. The author attempted to describe the performance elements using IPA + FACS.
Technically, the author is unsure how much of this the model actually "digested," but the final result is quite interesting. A key technique was applying color coding to the storyboard annotations; otherwise, the model would mix up the different information. The colors successfully helped distinguish body movements, camera motions, composition, and lighting from the environment and character rendering. The initial storyboard is included in the replies.
Related event: Seedance 2.0 prompt recipes: from retro VHS to cinematic realism(6 posts)→
More from Multimodal
- Getting Started with AI Video: Solving Consistency and Censorship — cynicalnewenglander · 2026-07-22
- Storyboard-first workflows are making AI dance videos and influencers more consistent — aftahi_ai · 2026-07-22
- Runpod MCP and Claude help spin up image and video generation workflows — 802high · 2026-07-22
- Midjourney prompt turns a bee into a glitching pixel explosion — michaelrabone · 2026-07-22
- A physics reward can improve video generation without creating a real physics engine — Dapper-Drawer4546 · 2026-07-22
- HeyGen adds a media-sourcing skill for coding agents with 75k images and 10k tracks — HeyGen · 2026-07-22