To make a Hotel Lobby-inspired AI video in faceswap, prepare your own performance MP4 and a clear portrait, upload the footage, assign portraits to the scenes you want to generate, and review the assembled result. The current tool uses character motion transfer. It is not a dedicated two-photo preset that independently replaces both performers in the original duet.
Start with the Hotel Lobby AI video overview to see the reference and workflow summary, or open the video workspace when your files are ready.
What is the Hotel Lobby video reference?
The orange-studio concept refers to Quavo and Takeoff performing HOTEL LOBBY on A COLORS SHOW. A simple colored background, a hanging microphone, and a restrained performance setup make the scene recognizable. That performance is a creative reference; it is not a faceswap-generated example or an editable clip supplied by this site.
For this workflow, prepare a performance you recorded or footage you can edit. You can borrow the idea of a minimal studio without reproducing the original performers, song, or choreography. Use your own music or choose music separately for your intended publication.
Choose between a solo test and a duet concept
| Your idea | How to prepare it for the current tool |
|---|---|
| One person in an orange studio | Record a clear solo shot with the intended background and camera angle |
| Two people appearing in the same edit | Record separate solo shots, put them into one MP4, and assign portraits to their separate segments |
| Two people overlapping in one shot | Keep the segment unchanged or prepare simpler footage; independent group-person targeting is unsupported |
| A portrait with no reference performance | Record or prepare the MP4 first; the photo alone does not supply motion |
Different portraits per scene are supported. Assigning a separate identity to each of two people inside the same shot is a different control and is not available here. Do not upload a complicated duet expecting the scene controls to separate its performers automatically.
Step 1: Prepare the studio footage and portrait
Record a short performance against a plain orange background or another background you want in the result. Keep the camera steady and the face visible. Let the microphone sit away from the face where possible: a cable or microphone crossing the eyes or mouth can make the generated boundary harder to review.
A roughly 5–10 second solo shot is a useful first experiment, not a required length or a tested performance benchmark. If you want a vertical post, prepare the source composition before uploading. This motion transfer workflow does not automatically build an orange studio from a text prompt.
Prepare a standard MP4 up to 100 MB and a clear JPG, PNG, or WebP portrait up to 10 MB. Compare your portrait with the reference angle and lighting. See the portrait quality checklist for image preparation.
Step 2: Upload and review the scene timeline
Sign in to the video workspace, choose the MP4, and select Upload and analyze. Wait for scene analysis, then preview each segment on the timeline.
Check whether the intended performer remains clear throughout the segment. Look for a hand crossing the face, a turn away from the camera, a second performer entering, or a microphone blocking facial details. Decide which segments to generate and which to keep before assigning portraits.
Step 3: Assign portraits or keep original clips
Add your portrait to the portrait library. Select a segment, choose the portrait, and use Apply to this scene. For a repeated appearance, use Apply to selected scenes or Apply to all generated scenes.
For a two-person edit made from separate solo shots, assign portrait A to the first performer’s segments and portrait B to the other performer’s segments. Confirm the selections in each preview rather than assuming that adding two photos creates a duet template.
Select Keep original for a title, room detail, or difficult group scene. Wait for all assignments to save. Every segment selected for generation needs a portrait before the generation step becomes available.
Step 4: Generate and check the moving result
Submit the scene plan with Generate video. The service queues the generated segments and assembles them with the original clips. Check the current trial and cost notice in the tool before submitting; processing time depends on the footage and queue.
Inspect the segment previews and the full result. Character motion transfer can alter clothing, body details, and background objects. It does not promise a face-only replacement that preserves every other part of the footage.
| What to inspect | Why it matters |
|---|---|
| Eyes and mouth during turns | A stable front-facing frame can hide identity changes later in the shot |
| Hands, microphone, and cable | Foreground objects can merge with or cross the generated face |
| Edges against the orange background | Plain backgrounds make flicker and outlines easier to see |
| Changes between performers | Separate segments may differ in lighting or appearance |
| Scene transitions and audio | A good individual scene does not guarantee a clean assembled edit |
If one generated segment needs another attempt, regenerate that segment. Failed tasks also have a retry option. For footage that remains too difficult, keep the segment or prepare a clearer reference instead of repeatedly processing the same shot.
Step 5: Check the soundtrack and download
The service joins the clips in their original order and merges the uploaded MP4’s audio. It does not automatically add HOTEL LOBBY, clone the original artists’ voices, or guarantee lip-sync to the track.
If the upload has your own beat, that is the source audio used. Select any replacement music in your editor or publishing platform according to the options available for your intended use. Preview the full export, then download the MP4 and clearly label it as AI-edited when sharing.
Frequently asked questions
Can I make the Hotel Lobby video from two photos alone?
Not through this reference-video workflow. It needs an MP4 and portrait assignments. Photos alone do not establish the movement, camera framing, or shot sequence.
Can the app create the orange background automatically?
Prepare the desired background in your source footage first. Generation may change background details, but the tool is not a text-to-studio editor or a guaranteed background-preservation effect.
Can I use the official COLORS video as my uploaded MP4?
The official player is linked for reference. This site does not provide the file or permission to edit and republish the performance. Use footage you recorded or have permission to process.
Does keeping the source audio mean the new character will sing in sync?
No. Audio merging and singing lip-sync are separate capabilities. This workflow does not provide dedicated control over a generated vocal performance.
What other video ideas use the same workflow?
The Chuttamalle-inspired dance tutorial covers preparing personal dance footage, while the AI world tour tutorial explains planning several stage scenes. Both use the same principle: a clear reference video, readable portraits, and review across frames.
When your source performance is ready, start a Hotel Lobby-inspired edit with one simple scene before expanding the project.
