What is an AI Hotel Lobby video?
Here, “AI Hotel Lobby” means an orange-stage, two-performer video inspired by the visual setting of Quavo & Takeoff’s HOTEL LOBBY performance for COLORS, published June 17, 2022. It is a scene-production idea, not a guide to designing hotel interiors. The original performance is a visual reference; this page does not supply its footage, song or performers’ likenesses.
The practical starting point is one image containing both original characters, already placed on a simple orange stage. Animate that composition with small gestures and a stationary camera. This creates a new shot; it does not reproduce the original performance or synchronize mouths to its recording.
Compose the first frame before planning motion
Prepare your own photograph, illustration or authorized composite outside the video generator. AI Genjutsu accepts an optional single first-frame image rather than combining two separate portraits. Use JPEG or PNG and settle the duo composition before submitting it.
Keep both faces large enough to read. Give each person a clear silhouette, leave space between their arms, and include room around their heads and feet. Different original outfits help you notice identity swaps. Put both performers on the same floor plane so one does not appear to float behind the other.
For the orange-stage look, reduce the scene to a flat background, coherent floor and restrained lighting. Avoid signage, extra performers and elaborate props on the first attempt. Save a clean version of the frame; comparing revisions becomes difficult when the composition changes every time.
For more on placing two characters, read the AI rap duo workflow.
Direct one short, readable action
Write what each character does, rather than asking for “the Hotel Lobby dance.” Name them by position and describe a small motion sequence. A useful original prompt is:
Locked medium-wide shot of two original adult performers on a seamless burnt-orange stage. The performer on the left makes a small shoulder bounce and lowers one hand. The performer on the right nods once, then gives a restrained open-palm gesture. Both stay in their original positions. Keep the background simple, the lighting steady and the camera stationary. No additional people or scene changes.
Start with a shorter supported duration and one gesture per person. A locked camera makes drift easier to detect and avoids adding an orbit or zoom to an already complex two-person shot. Prompt instructions guide the model; they cannot guarantee exact hand shapes, identities or timing.
AI Genjutsu’s workflow uses Seedance 1.5 Pro text or single-image generation, with audio disabled and watermark enabled. Check the studio's current availability before rendering. The prompt above is a planning example, not a tested result.
Check the result as a sequence
When generation is available, inspect the whole clip at normal speed before looking at isolated frames. Check whether the performers cross, exchange clothing details, grow extra fingers or change size. Compare the first, middle and final frames for background and face drift.
If identities mix, simplify arm movement and increase the space between bodies in the source image. If the stage mutates, remove competing lighting and scenery instructions. Change one variable for each retry and retain unsuccessful attempts when estimating cost. These are troubleshooting suggestions, not guaranteed repairs.
Music and mouth movement are separate tasks
A visually convincing rap scene can still have inaccurate mouth movement. This workflow has no audio upload, lyric generation or dedicated rap lip-sync control. Add only music you are licensed to use in an editor; merely naming a song in a prompt does not grant rights or produce synchronized vocals.
ByteDance’s Seedance 1.5 Pro announcement describes model-level audiovisual capabilities. Those capabilities do not mean audio is enabled in every application, including this one. See AI rap video planning for the broader production steps.
Common questions
Can I keep the exact original dance?
A first frame does not encode an entire performance. Exact choreography requires a different motion workflow and appropriate source permissions.
Can I use two celebrity photos?
The composition input is one image. That technical limit does not settle likeness, consent or publication rights; original characters are the straightforward planning choice.
Is this an official COLORS or artist tool?
No. AI Genjutsu is independent and has no claimed affiliation with Migos, Quavo, Takeoff, COLORS or Higgsfield.