Toc Toc Dance From One Photo

The template already contains the room, the camera move, the knee lift, the bounce and the knock audio. Nothing about the staging comes from you. What the model has to read from your photo is the person: the face, the hair, the outfit, the silhouette and the shoes. The template page lists exactly what is fixed on the other side of that split.

That is why a full-body or half-body shot is more reliable than a head-only portrait. The model can only carry what it can see; the rest is inference, and inference is where artefacts come from.

• One person. Crop everyone else out. Two faces leave the model without a subject.
• Front-facing. The face should be readable, not turned away or behind a phone.
• Even lighting. A window in front of the subject beats a hard lamp to one side.
• Whole body if you can. Legs and shoes tell the model how the dancer is built.
• No heavy filters. Smoothing, beauty filters and sunglasses remove the detail the model needs.
• Clean background. A busy frame is not copied into the clip, but it does make the subject harder to isolate.

Your photo becomes the person in the reference clip. The model tracks the motion, the camera move and the timing from the template, and keeps the template's original sound, so the knee lift and the bounce land on the same knock as the source.

It is motion control, not a text prompt. You are not describing a scene and hoping the model finds it; you are placing one person into a clip that already exists. That is also why the output is predictable — the room does not change between renders.

Most bad results trace back to the photo, not to the model. A head-only crop produces guessed legs. A dim photo loses the face. Two people in frame produce a blended subject. A photo where the subject is turned sideways gives the model almost nothing to track.

The other failure mode is expectation. The clip is about 12 seconds, it keeps the template's song, and it is 9:16. It will not become a 30-second routine, a different track or a different room. If those are the requirement, filming your own version is the right route. A render also costs 120 credits, so check the cost breakdown before spending one on a photo you are unsure about.

One adult, 18 or over, and one person only. Upload only a photo you own or have permission to use, with the consent of anyone shown. Uploaded images are checked by an automated moderation service before generation, and a rejected image cannot be used.

Do not present a generated clip as a real recording of someone's actions. The output is synthetic, and the terms require you to keep it that way. If a photo keeps getting refused, the troubleshooting guide walks through the rejection order.