How to Generate a Two-Person POV Couple Video

This tutorial explains how to create a handheld couple POV video with two uploaded identity photos. The target result is a romantic phone video: the woman leads the man by the hand, turns back, the man smoothly rotates the phone into selfie mode, and both people appear together in a close, intimate couple moment.

Best settings

• Model: HappyHorse 1.1 reference-to-video
• Resolution: 720P
• Ratio: 3:4
• Duration: 5 seconds
• Inputs: two people, two photos
• Photo order: image 1 is the man, image 2 is the woman

Photos to upload

Use one clear male identity photo and one clear female identity photo. The faces should be visible, sharp, and close enough for identity reference. Avoid tiny faces, heavy filters, sunglasses, or photos where the hairstyle is hidden.

The prompt should tell the model to preserve facial features, face shape, hairstyle, skin tone, age impression, body proportions, and overall personality. This reduces identity drift and prevents the model from mixing the two people.

The key POV logic

The most important part is the camera logic. The viewer should feel like they are watching footage captured by the man's phone camera.

• The video is not a third-person camera shot.
• The video is not an observer filming the couple.
• The phone device itself should not appear in the frame.
• In the opening, only the man's held hand, forearm, or a small edge of his body should appear.
• The man should enter the selfie frame only after he extends his arm and rotates the phone.

If the prompt does not define this transition, the model often places the man in the wrong position, makes him appear too early, or turns the scene into a normal third-person couple shot.

Prompt structure

Use this order when writing a two-person POV prompt:

1. Identity reference rules
2. Character styling and mood
3. Scene and atmosphere
4. POV phone camera logic
5. Opening hand-held action
6. Smooth transition into selfie
7. Close couple interaction
8. Negative constraints

Example prompt

Create a realistic adult couple POV phone video using two uploaded reference images. Reference image 1 is the male lead identity. Reference image 2 is the female lead identity. Preserve both people's facial features, face shape, hairstyle, skin tone, age impression, body proportions, and overall personality. Do not mix identities.

Scene: a bright private beach at golden hour, wet reflective sand, gentle waves, warm ocean wind, and soft sun on natural skin. The man is shirtless and wears dark beach shorts. The woman wears a sensual but tasteful high-end bikini, barefoot, with wind-blown hair. Her mood is feminine, cute, affectionate, playful, and relaxed, like a girlfriend being filmed by her boyfriend during a beach vacation.

Camera logic: this is the actual footage captured by the man's phone camera. The man holds the phone in his right hand. The viewer sees the phone camera footage itself, not a visible phone screen or a third-person camera. Keep a natural handheld feeling with slight walking movement and small wrist adjustments, but keep faces clear.

Opening action: the man walks behind the woman while she holds his left hand and leads him along the beach. The frame mainly shows the woman ahead of him, her body, hair, bikini, the beach, and the waves. Only the man's held left hand and a small part of his forearm may appear at the lower edge. Do not show the man's full body, face, chest, or shorts in the opening.

Transition: the woman slows down, turns back toward the phone camera, smiles with soft teasing eye contact, and gently pulls him closer. The man extends his right arm forward and outward, raises the phone, pulls it away from his body, and smoothly rotates it inward into a couple selfie angle. As the phone rotates, the man's face, right shoulder, and shirtless upper body gradually enter from the right side of the frame. The transition must feel like a real handheld couple selfie, continuous and without a jump cut.

Selfie action: the man stays on the right side of the frame and the woman stays on the left or slightly center-left. Their shoulders touch. The woman places one hand on his chest, shoulder, or neck and leans into him. They look at each other closely, smile for a brief moment, then move into a clear lingering romantic French kiss. Their heads tilt naturally in opposite directions, lips stay together, and the kiss feels warm, intimate, and consensual. Keep it tasteful and cinematic, not explicit or vulgar. Use a medium close-up showing faces, shoulders, and upper bodies, not an extreme mouth close-up.

After the kiss, they remain close in the selfie, smiling softly and looking into the phone camera while the man keeps recording.

Avoid: third-person camera, observer shot, visible phone device, broken POV logic, long walking sequence, no kiss, only a quick lip touch, stiff kiss, extreme mouth close-up, exaggerated tongue, face merging, male full body visible too early, male entering from the left, torso blocking the camera, sudden jump cut, mixed identities, changed facial features, deformed hands, extra fingers, distorted clothing, severe motion blur, warped background.

Practical tips

Start with a short walking introduction. In a 5-second video, the walking part should only establish the POV and relationship; the main result should be the selfie interaction.

Write the transition clearly. The best phrase is: "the man extends his right arm forward and outward, raises the phone, and rotates it inward into a couple selfie angle." This tells the model how the man moves from behind the camera into the frame.

Keep the intimacy readable, not graphic. A medium close-up with hands on shoulder, chest, or neck usually looks more natural than a mouth-only close-up.

If the man appears in the wrong place, strengthen the constraint: "the man must enter from the right side only after the phone rotates into selfie mode."

If the result becomes a normal couple shot, repeat: "this is the actual footage captured by the man's phone camera, not a third-person camera."