[低延迟端点]：完成首句保留和快速结束修正，包含ACK缓冲策略、批量读帧和主说话人端点回归测试

src/owner_voice_pet/config.py

@@ -25,10 +25,11 @@ class AppConfig:
     wake_kws_threshold: float = 0.15
     wake_kws_score: float = 1.0
     wake_ack_text: str = "我在"
-    post_playback_drain_ms: int = 50
+    post_playback_drain_ms: int = 0
     pipeline_mode: str = "live_turn_based"
     endpoint_mode: str = "primary_speaker"
     speaker_profile_ms: int = 600
+    speaker_profile_min_ms: int = 120
     speaker_absent_ms: int = 300
     speaker_similarity_threshold: float = 0.70
     speaker_min_rms: float = 0.012
@@ -73,10 +74,11 @@ class AppConfig:
             wake_kws_threshold=float(get("WAKE_KWS_THRESHOLD", "0.15") or "0.15"),
             wake_kws_score=float(get("WAKE_KWS_SCORE", "1.0") or "1.0"),
             wake_ack_text=get("WAKE_ACK_TEXT", "我在") or "我在",
-            post_playback_drain_ms=int(get("POST_PLAYBACK_DRAIN_MS", "50") or "50"),
+            post_playback_drain_ms=int(get("POST_PLAYBACK_DRAIN_MS", "0") or "0"),
             pipeline_mode=(get("PIPELINE_MODE", "live_turn_based") or "live_turn_based").lower(),
             endpoint_mode=(get("ENDPOINT_MODE", "primary_speaker") or "primary_speaker").lower(),
             speaker_profile_ms=int(get("SPEAKER_PROFILE_MS", "600") or "600"),
+            speaker_profile_min_ms=int(get("SPEAKER_PROFILE_MIN_MS", "120") or "120"),
             speaker_absent_ms=int(get("SPEAKER_ABSENT_MS", "300") or "300"),
             speaker_similarity_threshold=float(
                 get("SPEAKER_SIMILARITY_THRESHOLD", "0.70") or "0.70"
@@ -216,6 +218,7 @@ class AppConfig:
             )
         for name, value in {
             "OWNER_SPEAKER_PROFILE_MS": self.speaker_profile_ms,
+            "OWNER_SPEAKER_PROFILE_MIN_MS": self.speaker_profile_min_ms,
             "OWNER_SPEAKER_ABSENT_MS": self.speaker_absent_ms,
         }.items():
             if value <= 0: