For a singing Hotel Lobby duet, prepare two clear portraits, choose a short reference segment that already contains vocals, assign lead and support roles, and use its matching audio. Review the mouths and both identities before joining clips or sharing. Two photos plus background music alone do not guarantee singing.
What is the Hotel Lobby AI effect?
The recognizable look is two performers in an orange studio, a hanging microphone and a front-facing camera. The same visual style can be generated as a new scene or recreated with reference-guided motion. These workflows need different inputs.
Searches such as Hotel Lobby AI generator, Hotel Lobby filter and Hotel Lobby template can point to different tools. Check whether the tool accepts only photos, or also a reference performance and audio.
Choose the right workflow
Photo-only generation creates new gestures from portraits and a scene description. It suits an original duet. Motion transfer uses a reference video to guide timing and movement. Audio-driven lip sync uses a voice or singing track to guide mouth movements.
For a recognizable song performance, keep video and audio on the same source interval. Decide explicitly who sings the lead. A scene with music can still show a person standing silently.
Make your clip: six steps
Each Google account can try one free 3-second preview. The source is generated at 4 seconds; playback stops at 3 and there is no download button. A 10-second video is $5.99 and a 15-second video is $9.99; packs are $17.99 for 30 seconds or $54.99 for 100 seconds, spent as 10- or 15-second clips. Prices are before applicable tax. One-time purchase, no subscription. Sign in with Google before trying a preview, buying or accessing your videos. If the studio reaches capacity, its buttons pause until capacity is available again. Purchased credits are kept. The current studio buttons show availability.
Prepare two portraits
Choose one consenting adult per photo. Use JPG, PNG or WebP files up to 5 MB, with a visible face and upper body. Different photo backgrounds are fine; the template supplies the orange scene.
Select the people
Open the studio, sign in with Google, choose Person 1 and Person 2, check both photos and confirm permission. Choosing a file does not upload it. The built-in template supplies the reference video and matching audio.
Choose an actual singing segment
Each Google account can try one free 3-second preview. The source is generated at 4 seconds; playback stops at 3 and there is no download button. A 10-second video is $5.99 and a 15-second video is $9.99; packs are $17.99 for 30 seconds or $54.99 for 100 seconds, spent as 10- or 15-second clips. Prices are before applicable tax. One-time purchase, no subscription. New paid MP4s include the selected Hotel Lobby song excerpt from the same source interval as the performance. The template starts at 00:05 in the original video. We preserve normal speed; adding the audio does not guarantee perfect lip sync.
Make the singing roles explicit
Specify the left person as lead and the right person as support for this template. Check the chosen source segment, rather than inferring its roles from a later verse. Keep identities on their intended sides.
Wait for the same task
After submission, Making your clip appears and duplicate submissions are blocked. Check progress queries the same task. If submission is unknown, wait for its task and charge to be checked before trying another generation.
Review and save the MP4
Play the entire result with sound. Check both faces, mouth movement, hands and framing. Download promptly from the completed task. An additional generation can incur another charge; changing a prompt does not repair the existing file.
Music and mouth movement are different
Adding the Hotel Lobby song to a finished silent-looking clip does not make its people lip-sync. The mouth animation must be generated from the corresponding vocals; the final audio must use the same start point and duration.
In our local native 5-second test, the provider returned about 5.17 seconds. We removed the small ending surplus at normal speed and paired it with the same 5-second source audio. The left person opened her mouth in close shots, but the opening wide shot still did not meet the intended continuous-singing result. A successful task alone is not a quality check.
If you choose another soundtrack in TikTok or Instagram, check the available library for your account and region. Lower or mute the existing track when necessary so two tracks do not compete.
Check the result before sharing
Our tests show that body shape and singing assignment need explicit instructions. A model may inherit a reference performer's build or preserve a silent source mouth. Review an existing result before paying for another attempt.
- Both identities stay recognizable from beginning to end.
- The intended singer visibly opens their mouth when the vocal starts.
- Mouth timing follows the audio, not merely random facial movement.
- Hands, microphone and clothing remain coherent; both people stay in frame.
- The saved MP4 has the expected duration and an audible, matching soundtrack.