# Bulin 47 AI 街头 Freestyle 视频生成器

URL: https://aieffectlab.com/zh/video-effects/bulin-47-freestyle

用人物照片、直立宠物角色或两张独立人像制作街头表演；需要跟随指定动作时，上传动作参考片段。

## 这个效果能做什么

街头 freestyle 效果让主角在音箱搭档与围观圈旁完成肩部律动、小手势和收尾定格。照片变体会制作一段新的表演诠释；动作参考变体则跟随上传片段的动作。两种方式在同一页选择，按你已有的素材决定流程。

## 准备哪些素材

人物和原创角色变体各用一张清楚的身份图。宠物需要事先准备好肩部、爪部清晰的直立形象。双人变体中，照片 A 是表演主角，照片 B 是持音箱的搭档；两张身份图不是首帧与尾帧。动作参考变体需要一张身份图和一段 MP4 表演。

### 人物 freestyle

一张人物照片，制作新的肩部律动街头片段。

Use @Image1 as the sole lead identity, preserving face or markings, outfit and proportions. Make an original street-freestyle scene with a stable portable speaker held by an adult partner, an open spectator circle, visible hands and feet, and restrained background movement. The consenting adult turns toward the speaker, makes two alternating shoulder bounces, acknowledges the circle with a small hand gesture and holds the finish. Medium-wide continuous shot. Do not imitate a named artist’s face or voice.

### 宠物 freestyle

一张直立宠物角色图，让肩部与爪部动作清楚。

Use @Image1 as the sole lead identity, preserving face or markings, outfit and proportions. Make an original street-freestyle scene with a stable portable speaker held by an adult partner, an open spectator circle, visible hands and feet, and restrained background movement. Animate the prepared upright stylized pet character. Preserve ear shape, muzzle, coat and eye color. It makes two gentle shoulder accents and a small paw gesture, with stable articulated feet and a held finish.

### 原创角色

一张原创角色图，保留轮廓与颜色。

Use @Image1 as the sole lead identity, preserving face or markings, outfit and proportions. Make an original street-freestyle scene with a stable portable speaker held by an adult partner, an open spectator circle, visible hands and feet, and restrained background movement. Use the original mascot or character as designed. Preserve its silhouette, colors and proportions. Head turn, two shoulder accents, open hand gesture and a held finish; do not convert it into a realistic human.

### 自定义场景

一张主角照片，补充地点、搭档和手势。

Use @Image1 as the sole lead identity, preserving face or markings, outfit and proportions. Make an original street-freestyle scene with a stable portable speaker held by an adult partner, an open spectator circle, visible hands and feet, and restrained background movement. Build the performer, speaker partner, outfit, street setting and framing around the additional scene direction. Use a short continuous shot with a glance, two shoulder movements, crowd acknowledgment and a pause.

### 自带动作

一张身份图与一段动作视频，只替换主角。

Edit @Video1. Replace only its principal performer with the whole character in @Image1, preserving identity, outfit and proportions. Follow source movement, camera, pose order and action timing. Keep background performers unchanged and hands and feet coherent. Do not blend faces or add another lead.

### 主角与搭档

两张身份图：A 表演，B 持音箱。

Use @Image1 ONLY as the lead performer and @Image2 ONLY as the adult speaker partner. These are separate identities, never first/last frames. Both remain in a shared street composition: A glances at B, makes two shoulder accents and acknowledges the crowd; B holds a stable portable speaker at waist height with both hands. Never exchange faces or outfits. Continuous medium-wide framing with space between limbs.

## 如何制作视频

先选变体再上传。照片表演可选择时长和画幅，补充地点与手势要求，避免新增多余身份。动作参考上传后会核实秒数，输出时长必须覆盖片段；想保留构图时，选择与参考相同的画幅。双人表演先核对两个位置，再查看报价。提交一次，重新打开本页只查询保存的任务。

## 检查生成结果

分别检查主角与音箱搭档。脸和衣服应保持在各自角色上，音箱与手部握姿应稳定。留意肢体融合、脚部出画和背景突然移动。宠物与原创角色应保留原有轮廓和毛色。

脸部混合时，使用每张只包含一人的照片，并选择双人变体。动作偏差大时，改用清晰的短动作参考片段，而非堆叠更多文字。宠物随拍先处理成直立角色图。查询状态只跟踪已有任务，不会新建表演。

## 常见问题

### 哪个变体会跟随我自己的动作？

选择「自带动作」。上传一张身份照片和一段 MP4，照片提供主角外观，视频提供动作、节奏与构图。

### 两张人物照片可以有不同分工吗？

可以。在双人变体里，A 是表演主角，B 是持音箱的搭档。两张图分别作为身份参考，不作为开头和结尾帧。

### 会附带指定艺人的歌曲或声音吗？

不附带指定歌曲或人物声音。照片表演开启声音后会请求原创音效；有动作片段时优先参考其音轨，也可另传声音参考。发布前请检查生成结果。

[制作这个效果](https://aieffectlab.com/zh/video-effects/bulin-47-freestyle#generator)