
To make a Hotel Lobby AI trend edit, replace the two performers in the orange-studio scene with two character photos. Keep the shared microphone, distinct roles and performance rhythm, then inspect both people throughout the generated clip.
Make the edit step by step
1. Prepare two character photos
Choose one person per image, with clear facial features and even lighting. An upper-body or full-body photo also provides clothing information. Avoid covered faces, heavy filters and cropped heads.
2. Assign the left and right performers
Open the Hotel Lobby template and replace its two character images. Input 1 corresponds to the performer starting on the left; Input 2 corresponds to the performer starting on the right. Confirm both thumbnails before generating.
3. Keep the performance references
For a first attempt, retain the supplied video, audio and scene instructions. Change the cast first so you can judge whether the two photos work. Changing the camera, background and performance at the same time adds more problems to diagnose.
4. Generate and review both characters
Read the current credit quote before submitting once. Wait for the completed result, then check faces during turns, hands near the microphone, clothing consistency and audio alignment. If one person looks wrong, improve that portrait first.
Save the completed video using the result's download control. Playback duration is separate from generation time; the published example below is about 15 seconds long.
5. Prepare the vertical composition
The example is 4:3. For a 9:16 post, fit the full frame inside a vertical canvas before trying a crop. Check both performers through the entire clip; a centered crop can remove the person standing farther from the microphone. Add captions away from faces and hands.
Compare the photos with the result
The scene recreates the orange-studio performance associated with Quavo and Takeoff's “HOTEL LOBBY” on A COLORS SHOW. This published example replaces the duo with Einstein and Chaplin characters.
Compare the references with the moving result: the white hair and gray sweater on the left, and the bowler hat and dark suit on the right, give you specific details to inspect. Watch through the clip to judge whether the faces, outfits, and gestures remain convincing.
This is a published template example, not an independent quality test or a guarantee that every generation will match it. A fresh run can produce different details, even with the same inputs.
Optional Hotel Lobby AI prompt
For a custom edit in a tool that accepts two character images, video and audio, this prompt describes the cast replacement. The prepared template needs no extra prompt.
Recreate the performance in @Video1. Use @InputImage1 for the performer who starts on the left and @InputImage2 for the performer who starts on the right. Preserve each character's face, hairstyle, clothing, and proportions throughout. Follow the reference staging, camera framing, orange background, hanging microphone, and performance timing. Use @Audio1 as the audio reference.
This is an editing starting point; the displayed example was not generated as a test of this rewritten prompt. Attach the actual media and map the reference names to your tool's inputs. Writing a filename in the prompt does not upload it.
Common problems and what to change
| Problem | What to try next |
|---|---|
| The two characters start on the wrong sides | Check the two Input slots against the left/right instructions before changing the prompt. |
| A face drifts or loses its identity | Use a sharper, less obstructed photo and inspect moments when the subject turns. |
| Clothing changes during the clip | Try a reference showing more of the intended outfit, with a clear silhouette. |
| Hands or the microphone distort | Inspect the entire result; a good first frame can still lead to an unusable section. Trim a short flawed section if the edit still works. |
| Lip movement feels out of sync | Compare the generated video with its audio before trimming. Reference audio does not guarantee perfect lip sync. |
| One person disappears in the vertical edit | Fit the full 4:3 frame inside the 9:16 canvas and check both subjects throughout. |
An audio reference guides performance but does not guarantee an unchanged recording or perfect lip sync. If you need a specific soundtrack, align it in your finishing editor and review the export with sound. Increasing export resolution does not add detail missing from the generated source.