apps/worker/video_processing/render_adapter.py

@@ -483,6 +483,7 @@ class RenderAdapter:
         progress_cb: ProgressCallback | None = None,
         rendered_clip_ids: list[str] | None = None,
         failed_clip_ids: list[str] | None = None,
+        voiceover_audio_path: str | None = None,
     ) -> RenderAdapterResult:
         """执行统一渲染核心流程（BGM + ASR + 渲染 + 缩略图 + 上传）。
 
@@ -491,6 +492,7 @@ class RenderAdapter:
         Args:
             rendered_clip_ids: 成功下载/准备的 clip id 列表（render_plan 从下载阶段传入）
             failed_clip_ids: 失败的 clip id 列表
+            voiceover_audio_path: 配音素材库音频本地路径（一键生成场景使用）
 
         Returns:
             RenderAdapterResult
@@ -525,6 +527,7 @@ class RenderAdapter:
             output_height=output_height,
             bgm_path=bgm_path,
             asr_service=asr_service,
+            voiceover_audio_path=voiceover_audio_path,
         )
         result = render_svc.render()
 
@@ -593,6 +596,7 @@ class RenderAdapter:
         job_id: str = "",
         work_dir: Path | None = None,
         progress_cb: ProgressCallback | None = None,
+        voiceover_audio_path: str | None = None,
     ) -> RenderAdapterResult:
         """使用内存中的 plan/clips/asset_path_map 直接渲染。
 
@@ -606,6 +610,7 @@ class RenderAdapter:
             job_id: 关联的 Job ID
             work_dir: 工作目录，不传则用临时目录
             progress_cb: 进度回调
+            voiceover_audio_path: 配音素材库音频本地路径
 
         Returns:
             RenderAdapterResult
@@ -649,6 +654,7 @@ class RenderAdapter:
                 plan_id=actual_plan_id,
                 job_id=job_id,
                 progress_cb=progress_cb,
+                voiceover_audio_path=voiceover_audio_path,
             )
 
         except subprocess.CalledProcessError as exc:

apps/worker/video_processing/unified_render_service.py

@@ -175,6 +175,7 @@ class UnifiedRenderService:
         transition_duration: float = DEFAULT_TRANSITION_DURATION,
         asr_service: Any = None,  # ASRService 实例，用于自动生成字幕
         bgm_path: str | None = None,  # BGM 本地文件路径
+        voiceover_audio_path: str | None = None,  # 配音素材库音频本地路径
     ):
         self.plan = plan
         self.clips = clips
@@ -186,6 +187,7 @@ class UnifiedRenderService:
         self.transition_duration = transition_duration
         self.asr_service = asr_service
         self.bgm_path = bgm_path
+        self.voiceover_audio_path = voiceover_audio_path
         self._transition_engine = TransitionEngine(default_duration=transition_duration)
         self._speed_engine = SpeedEngine()
         self._asr_timeline_cache: Any = None  # ASR 字幕结果缓存，避免重复调用
@@ -226,6 +228,9 @@ class UnifiedRenderService:
         # 3.5 TTS 配音生成（如果配置了）
         self._maybe_add_voiceover_layer(layers, video_duration=video_duration)
 
+        # 3.6 配音素材库音频（如果传入了本地路径）
+        self._maybe_add_voice_library_layer(layers, video_duration=video_duration)
+
         # 4. 生成 ASS 字幕文件（如果有 title/subtitle 配置）
         ass_path = self._maybe_generate_ass(video_duration)
 
@@ -843,6 +848,69 @@ class UnifiedRenderService:
             logger.warning("TTS 配音异常，跳过: %s", e)
             return False
 
+    def _maybe_add_voice_library_layer(
+        self,
+        layers: list[RenderLayer],
+        *,
+        video_duration: float,
+    ) -> bool:
+        """将配音素材库音频作为整段配音加到 audio 图层.
+
+        与 TTS 配音共享同一套 audio 图层混音架构，
+        支持与 BGM、TTS 的音量平衡，不再走独立的后处理 mux 链路。
+
+        Returns:
+            是否成功添加了配音音轨
+        """
+        if not self.voiceover_audio_path:
+            return False
+
+        audio_path = Path(self.voiceover_audio_path)
+        if not audio_path.exists() or audio_path.stat().st_size == 0:
+            logger.warning("配音素材库音频文件不存在或为空，跳过: %s", self.voiceover_audio_path)
+            return False
+
+        try:
+            # 找到或创建 audio 图层
+            audio_layer = None
+            for layer in layers:
+                if layer.role == "audio":
+                    audio_layer = layer
+                    break
+
+            if audio_layer is None:
+                from video_processing.unified_render_service import _LAYER_Z_INDEX  # type: ignore
+
+                z_index = _LAYER_Z_INDEX.get("audio", 2)
+                audio_layer = RenderLayer(role="audio", z_index=z_index)
+                layers.append(audio_layer)
+
+            # 配音素材作为整段配音：从 0 开始，覆盖整个视频时长
+            # 音频不足视频时长时，混音层会按实际长度处理（amix 不自动循环）
+            vo_clip = ResolvedClip(
+                clip_id="voice_library_main",
+                asset_id="voice_library",
+                local_path=audio_path,
+                clip_type="audio",
+                order=len(audio_layer.clips),
+                start_time=0.0,
+                duration=video_duration,
+                config={"volume": 1.0, "voice_library": True},
+                actual_duration=video_duration,
+            )
+            audio_layer.clips.append(vo_clip)
+
+            logger.info(
+                "配音素材库音频已添加到 audio 图层: plan_id=%s duration=%.2fs",
+                self.plan.id,
+                video_duration,
+            )
+            return True
+
+        except Exception as e:
+            logger.warning("配音素材库音频添加失败，跳过: %s", e)
+            return False
+
     @staticmethod
     def _resolve_watermark_config(plan_config: dict[str, Any] | None) -> WatermarkConfig | None:
         """从 plan config 中解析水印配置，兼容两种存储格式.

apps/worker/worker_app/tasks/generation.py

@@ -1238,6 +1238,7 @@ def _render_video(
             plan_id=f"gen_{task_id}",
             job_id=task_id,
             work_dir=temp_path,
+            voiceover_audio_path=voice_path,
         )
     finally:
         db.close()
@@ -1256,17 +1257,8 @@ def _render_video(
         render_duration,
     )
 
-    # 配音混音（素材库音频，后处理混音）
-    if voice_path:
-        final_path = temp_path / f"final-{task_id}.mp4"
-        try:
-            _mux_audio_track(render_output_path, voice_path, final_path)
-            output_path = final_path
-        except Exception as mux_err:
-            logger.warning("[task_id=%s] [混音] 音频混合失败，使用无音频版本: %s", task_id, mux_err)
-            output_path = render_output_path
-    else:
-        output_path = render_output_path
+    # 配音素材库音频已在统一渲染引擎内部通过 audio 图层混音处理
+    output_path = render_output_path
 
     return output_path, render_duration