Choose the smallest performance scope that covers what viewers must see. A spoken close-up needs readable mouth movement and expression. A gesture-led explanation needs hands and upper-body timing. A dance shot may need foot placement, turns and weight changes. A convincing talking face does not demonstrate that a tool can reproduce the rest of a performance.
Start with the visible information
Watch the intended shot without sound, then listen without the image. What information disappears in each test? A line of dialogue may carry the meaning through words. A shrug may communicate hesitation visually. A step-and-turn phrase depends on relationships among hips, knees and feet.
Write the missing information as the requirement. “The character speaks” is incomplete if a hand must point to an object at a particular moment. “The character dances” is too broad if only a shoulder reaction will be shown.
This choice also determines the reference you need. A mouth close-up cannot provide evidence for a full-body pivot.
Read body coverage as a real constraint
Act-Two's best-practice guidance frames a single performer waist-up or closer. It is not a universal full-body dance specification. Kling Motion Control's documentation covers body movement, gestures and expressions.
Keep those provider-specific statements separate from your own acceptance criteria. A supported category still needs a test with your framing, costume and movement.
| Shot requirement | Evidence that must remain visible | First review |
|---|---|---|
| Spoken close-up | Mouth, eyes and face outline | Speech timing and expression |
| Hand-led explanation | Face and complete gesture | Correct hand and object relationship |
| Dance phrase | Relevant body and ground contact | Action order, turns and balance |
| Walking reaction | Body orientation and face | Movement and expression together |
Test one layer without hiding the others
For speech, choose a brief line containing a pause and a visibly different mouth position. Review the audible beginning and ending against the face. Then inspect whether the expression suits the line rather than merely moving continuously.
For a gesture, use an action with a clear start, peak and return. Check which hand performs it and whether the fingers disappear at the edge.
For dance, identify a small recognizable phrase. Watch the legs and torso before concentrating on the face. Do not cover a failed turn with a close-up and call the original full-body requirement satisfied.
These are separate review exercises, not a provider benchmark.
Decide whether to divide the shot
A single shot combining dialogue, a rapid spin and interaction with another person creates several independent reasons to fail. Consider a speaking close-up followed by a wider movement shot when the story permits it. The cut should serve the scene, not conceal a crucial action viewers were promised.
Retain the same appearance notes across both shots. If speech and body movement must happen together, test that combination explicitly; success on two separate clips does not establish synchronization.
Questions before generation
Does lip sync include facial acting?
Check the actual feature description. Mouth synchronization and an emotionally appropriate performance are different observations, even when one product combines them.
Can I use a waist-up performance for a full-body character?
Only judge the body information the performance actually supplies. If feet or leg timing matter, obtain a suitable recording rather than inventing unseen choreography in your review notes.