Treat each source dancer as a continuing character, even when the pair trades positions. This makes the intended replacement easier to describe and the result easier to judge.
Choose the Paired Performance
Select Smalltown Boy AI from the video examples to load the supplied male-and-female performance and its custom replacement prompt. Pick the exchange you want to mimic and trim the reference if prompted.
Choose Two Distinct Appearances
Use one photo per character field. For pet experiments, select images where different facial features and coat markings are easy to distinguish.
Map the Images to the Original Pair
The supplied prompt maps Character Image 1 to the male performer and Character Image 2 to the female performer. Keep those assignments through changes of place, then review generation settings and the credit estimate.
Inspect the Generated Exchange
Preview the moment the characters move past each other. Check that both moving bodies belong to the intended cast and save a take with usable identities and motion.
Choose the Paired Performance
Select Smalltown Boy AI from the video examples to load the supplied male-and-female performance and its custom replacement prompt. Pick the exchange you want to mimic and trim the reference if prompted.
Choose Two Distinct Appearances
Use one photo per character field. For pet experiments, select images where different facial features and coat markings are easy to distinguish.
Map the Images to the Original Pair
The supplied prompt maps Character Image 1 to the male performer and Character Image 2 to the female performer. Keep those assignments through changes of place, then review generation settings and the credit estimate.
Inspect the Generated Exchange
Preview the moment the characters move past each other. Check that both moving bodies belong to the intended cast and save a take with usable identities and motion.
A still opening frame can look convincing while the later exchange loses the new identities. Evaluate what happens when the pair moves, overlaps, and separates again.
AnyMimic requests a generated replacement of the reference pair. Each image supplies appearance, while the video guides gestures, reactions, overlap, and the shared environment.
Separate images let you specify two distinct characters rather than asking the model to infer a pair from a single background photo.
Explicit instructions keep the same assignment when the characters switch sides. That is the intended behavior; imperfect generation can still require another take.
Inspect subject count, likeness or coat markings, body integrity, overlap, and any remaining source characters. Download a usable clip for final timing and audio edits.
Use the result as a short performance asset, then finish the context and soundtrack in your own editing workflow.
订阅套餐与价格
标准积分包
$10
一次付款获得 15 积分
解锁更长的视频生成时长
无需订阅
无自动续费
进阶积分包
$30
一次付款获得 60 积分
解锁更长的视频生成时长
无需订阅
无自动续费
超值积分包
$80
一次付款获得 200 积分
解锁更长的视频生成时长
无需订阅
无自动续费
480p 视频生成每秒约消耗 1 积分，15 秒视频约需 15 积分。
不同模型的生成成本不同。
生成页面会自动计算积分消耗，并显示在生成按钮旁。