FBX动画批量导入与重定向全流程自动化可行性咨询

可行性分析 #

所有步骤都完全可以通过Unreal Engine的Python脚本实现——UE的Python API已经覆盖了资产导入、编辑、序列创建等全流程操作，完全能支撑数百个动画的批量处理需求。

最佳脚本环境 #

直接使用Unreal Engine内置的Python环境即可：

• 先启用UE的Python插件：编辑 > 插件 > 搜索"Python Scripting"并启用，重启编辑器
• 脚本可在UE内置脚本编辑器编写，也能用VS Code等外部编辑器编写后导入执行
• 核心依赖UE自带的unreal模块，无需额外安装第三方库

分步实现思路 #

1. 自定义FBX导入设置 #

通过unreal.FbxImportUI配置导入参数，取消骨骼网格体导入并指定现有目标网格：

import unreal

def setup_fbx_import_settings(target_skeletal_mesh_path):
    import_ui = unreal.FbxImportUI()
    import_ui.import_skeletal_mesh = False  # 取消导入骨骼网格
    target_mesh = unreal.load_asset(target_skeletal_mesh_path)
    import_ui.skeletal_mesh = target_mesh  # 指定现有目标网格
    # 启用动画导入
    import_ui.import_animations = True
    import_ui.animation_import_data.set_editor_property("import_motion", True)
    return import_ui

2. 删除冗余A姿态序列 #

导入后遍历资产，筛选并删除命名含"A姿态"的冗余动画序列：

def delete_duplicate_anim_sequences(imported_assets):
    for asset in imported_assets:
        if isinstance(asset, unreal.AnimationSequence) and "A姿态" in asset.get_name():
            unreal.EditorAssetLibrary.delete_asset(asset.get_path_name())

3. 按规则重命名资产 #

根据自定义规则批量重命名动画序列（示例为前缀+原始名+后缀，可按需调整）：

def rename_anim_sequences(imported_assets, naming_rule):
    for asset in imported_assets:
        if isinstance(asset, unreal.AnimationSequence):
            original_name = asset.get_name()
            new_name = naming_rule.format(original_name=original_name)
            unreal.EditorAssetLibrary.rename_asset(asset.get_path_name(), new_name)

4. 配置身体动画重定向源 #

为身体动画序列指定固定重定向源并保存：

def set_body_retarget_source(anim_seq_path, retarget_source_path):
    anim_seq = unreal.load_asset(anim_seq_path)
    retarget_source = unreal.load_asset(retarget_source_path)
    anim_seq.set_editor_property("retarget_source", retarget_source)
    unreal.EditorAssetLibrary.save_asset(anim_seq_path)

5. 设置面部动画预览姿态 #

为面部动画序列绑定固定预览姿态资产并保存：

def set_face_preview_pose(anim_seq_path, preview_pose_path):
    anim_seq = unreal.load_asset(anim_seq_path)
    preview_pose = unreal.load_asset(preview_pose_path)
    anim_seq.set_editor_property("preview_pose_asset", preview_pose)
    unreal.EditorAssetLibrary.save_asset(anim_seq_path)

6. 创建Level Sequence并绑定动画 #

新建Cinematic Level Sequence，添加MetaHuman动画轨道并绑定身体、面部动画：

def create_level_sequence(seq_name, save_folder, body_anim_path, face_anim_path, metahuman_actor_path):
    # 创建Level Sequence
    seq = unreal.LevelSequenceBlueprintFactory().create_new(
        unreal.EditorAssetLibrary.get_path_name(save_folder), seq_name
    )
    # 添加动画轨道并绑定动画
    body_track = seq.add_master_track(unreal.MovieSceneAnimationTrack)
    body_section = body_track.add_section()
    body_section.set_editor_property("animation", unreal.load_asset(body_anim_path))
    
    face_track = seq.add_master_track(unreal.MovieSceneAnimationTrack)
    face_section = face_track.add_section()
    face_section.set_editor_property("animation", unreal.load_asset(face_anim_path))
    
    # 删除默认控制绑定
    for binding in seq.get_bindings():
        if binding.get_name() == "ControlRig":
            seq.remove_binding(binding)
    
    unreal.EditorAssetLibrary.save_asset(seq.get_path_name())

7. 添加音频轨道并关联资产 #

在已创建的序列中添加音频轨道并绑定对应音频：

def add_audio_to_sequence(seq_path, audio_asset_path):
    seq = unreal.load_asset(seq_path)
    audio_track = seq.add_master_track(unreal.MovieSceneAudioTrack)
    audio_section = audio_track.add_section()
    audio_section.set_editor_property("sound", unreal.load_asset(audio_asset_path))
    unreal.EditorAssetLibrary.save_asset(seq_path)

批量执行整合 #

将上述函数组合，遍历所有FBX文件完成批量处理：

def batch_process_fbx_animations(fbx_file_list, config):
    for fbx_path in fbx_file_list:
        # 1. 导入FBX
        import_ui = setup_fbx_import_settings(config["target_skeletal_mesh"])
        imported_assets = unreal.EditorAssetLibrary.import_asset(
            fbx_path, config["import_folder"], import_ui=import_ui
        )
        # 2. 删除冗余资产
        delete_duplicate_anim_sequences(imported_assets)
        # 3. 重命名动画序列
        rename_anim_sequences(imported_assets, config["naming_rule"])
        # 4-5. 区分并配置身体/面部动画
        for asset in imported_assets:
            if isinstance(asset, unreal.AnimationSequence):
                asset_path = asset.get_path_name()
                if "Body" in asset.get_name():
                    set_body_retarget_source(asset_path, config["body_retarget_source"])
                elif "Face" in asset.get_name():
                    set_face_preview_pose(asset_path, config["face_preview_pose"])
        # 6. 创建Level Sequence
        base_name = fbx_path.split("/")[-1].split(".")[0]
        seq_name = config["seq_naming_rule"].format(base_name=base_name)
        seq_path = f"{config['seq_save_folder']}/{seq_name}"
        body_anim_path = f"{config['import_folder']}/{base_name}_Body"
        face_anim_path = f"{config['import_folder']}/{base_name}_Face"
        create_level_sequence(seq_name, config["seq_save_folder"], body_anim_path, face_anim_path, config["metahuman_actor"])
        # 7. 添加对应音频（假设音频与FBX同名同路径）
        audio_path = fbx_path.replace(".fbx", ".wav")
        add_audio_to_sequence(seq_path, audio_path)

注意事项 #

• 先拿单个文件测试脚本流程，确认参数配置正确后再批量执行
• 批量操作前备份所有资产，避免误删或错误修改
• 需根据自身的命名规则、资产路径调整脚本中的判断逻辑和路径拼接方式

内容的提问来源于stack exchange，提问作者Tom Auger