01The possession in one paragraph
The clip opens wide on a floodlit outdoor court at night. The attacker, in a light-blue blazer, white top, black trousers and heels, crosses over and drives past the defender, who wears a red suit-style outfit. By about 1.5 seconds she is hanging on the rim for the first dunk. The ball drops through the net, and from about 3.5 to 5.8 seconds she stands near the baseline with the ball bouncing and celebrates toward the camera, a line of spectators behind her. A second drive starts at about 5.8 seconds, and from about 7.2 to 10.3 seconds she rises over the defender in slow motion, legs split, ball held above the rim. A second celebration runs from about 11.9 to 14 seconds before the camera whips up to the floodlights and back to the attacker walking toward the lens.
02Role 1: the Attacker
The attacker is the star. She is on screen for almost the whole 15 seconds and is the only person seen facing the camera for long stretches: during both celebrations she stands still, square to the lens, at full-body or three-quarter size. That makes her face and outfit the most visible thing in the video. Put the person who should get the credit, the bragging rights or the joke in this slot. A sharp, front-facing photo with the face clearly visible and good resolution gives the template the most to work with.
03Role 2: the Defender
The defender is seen very differently. She appears in profile during the two drives (about 0.5 to 1.2 seconds and 6.0 to 6.8 seconds) and then spends both dunks in the foreground with her back to the camera, watching the ball go through the hoop. Viewers recognise her mostly by hair, build and outfit rather than by her face. When you cast a friend here, choose a photo that shows the hair and clothing clearly as well as the face; a tight face crop gives the template little to carry into those over-the-shoulder shots.
04Two players required, the court optional
Every 1v1 Dunk video needs two photos: one for the attacker and one for the defender. The reference players are not reused in your video, so both roles are always cast from the photos you upload, and the studio will not start a job without both. The Court slot is the only optional one: leave it empty to keep the original floodlit night court, or add a location photo to move the game somewhere else. That makes two or three photos per video. The credit quote always includes preparing both player photos; a location photo adds no character preparation. Read the quote before you generate.
05Choosing who scores on whom
The roles are not symmetrical, so the casting is the joke. The attacker gets two dunks and two celebrations; the defender gets beaten twice and watches from behind. If you want a rematch with the roles reversed, that is a second video with the photos swapped between the two slots, not a setting inside one video. Use photos of yourself or of people who have agreed to appear, one person per photo.
06What your photos do not change
The choreography, cuts, slow-motion timing, single ball and hoop are part of the template. Your photos change who plays each part; they do not add moves, change who wins or put the two players face to face at the end. The source's sound is kept: there is no speech, and the audio carries the label "original sound - Scrambled.ai" with no named song. The preview on the trend page is the credited source clip, not an Oakgen result.
