This guide addresses the lip sync AI search by preparing a scene with dialogue generated by Veo 3.1. Describe the character and put the spoken line in quotation marks. It does not synchronise an existing voice recording or replace a dedicated dubbing service. Check pronunciation, mouth movement and visual continuity; exact synchronisation is not guaranteed.
A concrete subject and one main action reduce ambiguity in the request.
Action
A concrete subject and one main action reduce ambiguity in the request.
Setting
The setting and lighting establish a scene without combining incompatible ideas.
Camera
One camera or framing instruction makes the composition easier to assess.
Lighting
The setting and lighting establish a scene without combining incompatible ideas.
Style
Separating the subject from the background helps you check each part of the result.
Sound
Distinguishing what stays from what changes makes comparisons between attempts more useful.
Character with dialogue
1. Character with a generated line
A fictional adult workshop host faces the camera in a quiet room. In one stable medium shot, the host says exactly: "The workshop opens tomorrow morning." Relaxed expression, soft daylight, no other speakers, no subtitles and no background music.
Suggested settings: 720p · 16:9 · 8 s
Why it helps: The line is supplied as text to the model, not as a real voice track or dubbing request. Review the result without promising synchronisation.
A fictional adult gardener stands beside a greenhouse and looks toward the camera. The gardener says: "These seedlings are ready for the garden." Fixed medium shot, gentle daylight, one speaker and a quiet background. Keep the face visible throughout the shot.
Suggested settings: 720p · 16:9 · 8 s
Why it helps: One speaker and a short line make voice, mouth movement and facial continuity easier to check, without requesting cloning.
A compact reading room contains a low bookcase, a fabric armchair and a small round table. Afternoon light falls across the floor from one tall window. Keep furniture proportions believable and the walking space clear. Use a wide composition with natural materials and a quiet, lived-in atmosphere. A curtain moves gently while the camera slowly reveals the room.
Suggested settings: 720p · 16:9 · 8 s
Why it helps: Separating the subject from the background helps you check each part of the result.
A narrow footpath crosses a meadow toward a distant hill at sunrise. Fine mist stays close to the grass while the hill remains visible. Use a wide composition with the path leading through the frame. Keep the horizon level, the light soft and the colours consistent with early morning. Mist drifts across the grass; keep the camera fixed.
Suggested settings: 720p · 16:9 · 8 s
Why it helps: The setting and lighting establish a scene without combining incompatible ideas.
A close view shows water droplets resting on a dark green leaf. Keep one central droplet in focus and the surrounding veins easy to recognise. Light enters from the side and produces a small highlight without bleaching the leaf. Separate the background with shallow depth and a simple composition. One droplet slowly runs down the leaf while the macro camera stays fixed.
Suggested settings: 720p · 16:9 · 8 s
Why it helps: Separating the subject from the background helps you check each part of the result.
A fictional adult explorer wears a mustard raincoat, dark boots and a small canvas backpack. The explorer pauses beside a stone bridge and looks toward a forest path. Preserve the coat, backpack and hairstyle as identifying details. Use soft morning light and avoid introducing additional characters into the composition. The explorer turns toward the path and takes one careful step.
Suggested settings: 720p · 16:9 · 8 s
Why it helps: Distinguishing what stays from what changes makes comparisons between attempts more useful.
A cyclist wearing a plain blue jacket moves steadily along an empty riverside path. Keep the bicycle proportions coherent and the rider fully inside the composition. Show the river as a quiet background with a few trees. Use gentle daylight and describe one continuous movement without a sudden change of location. Track the cyclist steadily from the side without sudden cuts.
Suggested settings: 720p · 16:9 · 8 s
Why it helps: One camera or framing instruction makes the composition easier to assess.
A folded linen cloth lies beside a smooth stone and a small unfinished wooden bowl. Make the weave, grain and stone surface distinguishable without exaggerating their texture. Use soft light from a large window and a simple overhead composition. Keep the background neutral and leave clear space around each object. A soft breeze moves one edge of the linen; keep the bowl and stone still.
Suggested settings: 720p · 16:9 · 8 s
Why it helps: Separating the subject from the background helps you check each part of the result.
Before handing over an asset, check names, visible text, pronunciation where relevant and any detail that could mislead an audience. Verify important claims against your own sources and confirm your rights to the input material.
A concrete subject and one main action reduce ambiguity in the request.
The setting and lighting establish a scene without combining incompatible ideas.
Separating the subject from the background helps you check each part of the result.
One camera or framing instruction makes the composition easier to assess.
Short quoted text lets you check whether the output contains the requested words.
Distinguishing what stays from what changes makes comparisons between attempts more useful.
When to choose another model
Distinguishing what stays from what changes makes comparisons between attempts more useful.
Opening the Studio and preparing your text is free. Generating requires an account and credits; the Free plan does not include credits. You see the cost before confirming.
Can I upload a recorded voice to synchronise it?
This Veo 3.1 workflow does not accept audio references. The character and voice are generated from the description and spoken line; it does not synchronise an existing recording.
Does a reference face determine the voice?
It does not provide a voice recording. Describe the character and line, then listen to the generated performance. This workflow does not offer voice cloning.
How do I check whether the dialogue works?
Use a short line and review pronunciation, duration, mouth movement and facial continuity throughout the clip. Rewrite if needed; instructions do not guarantee exact synchronisation.