58TH FLOOR / FIELD GUIDE
What Is the 58th Floor AI Video Format?
Understand the two-person teal-studio format, how A and B map to the reference, and what changes when new people take the roles.
Explore the performanceQuick answer
The short version
58th Floor AI is a photo-to-video trend inspired by Anderson .Paak and Cordae’s “58th Floor” performance on A COLORS SHOW. For this preset, it means a two-person studio performance recreated with identities from two photos. The visual signature is a teal set, one suspended silver microphone, expressive gestures and a solo close-up. It is a reference-guided AI recreation, not an original recording of the uploaded people.
Explore the studioThe scene is already set
Our selected reference runs for 15 seconds in vertical framing. It begins with two people at a hanging microphone, moves into the right performer’s close-up, widens to show the circular platform and returns to the pair.
These details come from the reference used for this preset. Other videos using the same search term may have different cuts or lengths, so use this site’s role descriptions when preparing your photos.
Two identities, two continuous roles
Photo A replaces the role initially on the left, identified in the reference by a black tuxedo. Photo B replaces the role on the right, identified by a white dress; B also owns the solo close-up.
Those descriptions locate the characters in the reference. They do not require the uploaded people to have a particular gender or wear wedding clothes. Each supplied identity should remain attached to the same performance role through every cut.
The face changes; the expression still performs
Replacing a person does not mean freezing their uploaded expression. A portrait supplies facial identity. The reference supplies the changing gaze, eyebrows, mouth shapes, head turns, gestures and reactions.
The intended result combines both: a recognizable new person performing the original role. It should not turn both roles into identical movements or transfer B’s expressions to A.
Reference-guided generation is not a frame-perfect face swap
A face overlay tries to change a facial region within existing footage. This workflow generates new video frames using identity photos and a performance reference. It can carry appearance cues such as hair and clothing into the scene.
Because the model recreates the frames, facial details, fingers, gestures and lip timing can vary. Keeping the original performance is the direction given to the model, not a guarantee that every frame will match.
Share a creative remake with clear context
Use photos with the people’s permission and describe a synthetic performance honestly. Do not present it as footage of them actually recording in the studio or as an endorsement by the original performers.
58th Floor AI is an independent tool. Confirm the rights for any soundtrack and intended use before sharing; a trend name or a generated video does not itself grant music rights.
FAQ
Frequently Asked Questions
Can I choose different choreography?
This preset is built around one fixed performance. It is not a general prompt editor or an arbitrary video upload tool.
Will every gesture and mouth shape match exactly?
The reference guides them, but AI generation can vary. Review the full result for identity, movement and timing.
Is the illustrated studio an actual generated example?
Yes. The hero and separate example were generated with this website. The two displayed portraits correspond to the hero video. Scene illustrations explain the shots; they are not additional generated results.
58th Floor AI
Choose your two performers
58th Floor AI is an independent creative tool. Use photos with permission and review your result before sharing.
Explore the studioIf payment or delivery needs attention, use the support details and published refund policy.