diff --git a/paddlespeech/server/engine/tts/paddleinference/tts_engine.py b/paddlespeech/server/engine/tts/paddleinference/tts_engine.py
index 43b0df407..97c6744f1 100644
--- a/paddlespeech/server/engine/tts/paddleinference/tts_engine.py
+++ b/paddlespeech/server/engine/tts/paddleinference/tts_engine.py
@@ -409,7 +409,9 @@ class PaddleTTSConnectionHandler(TTSServerExecutor):
 
         # wav to base64
         buf = io.BytesIO()
-        wavfile.write(buf, target_fs, wav_speed)
+        # wavfile.write(buf, target_fs, wav_speed)
+        sf.write(buf, wav_speed, target_fs, format="wav")
+        buf.seek(0)
         base64_bytes = base64.b64encode(buf.read())
         wav_base64 = base64_bytes.decode('utf-8')
         logger.debug("Audio to string successfully.")
diff --git a/paddlespeech/server/engine/tts/python/tts_engine.py b/paddlespeech/server/engine/tts/python/tts_engine.py
index 4d1801006..5031519e6 100644
--- a/paddlespeech/server/engine/tts/python/tts_engine.py
+++ b/paddlespeech/server/engine/tts/python/tts_engine.py
@@ -173,7 +173,10 @@ class PaddleTTSConnectionHandler(TTSServerExecutor):
 
         # wav to base64
         buf = io.BytesIO()
-        wavfile.write(buf, target_fs, wav_speed)
+        # wavfile.write(buf, target_fs, wav_speed)
+        sf.write(buf, wav_speed, target_fs, format="wav")
+        buf.seek(0)
+
         base64_bytes = base64.b64encode(buf.read())
         wav_base64 = base64_bytes.decode('utf-8')
         logger.debug("Audio to string successfully.")