diff --git a/audio_playback.py b/audio_playback.py
index eb2a8a5..cdd66b6 100644
--- a/audio_playback.py
+++ b/audio_playback.py
@@ -8,6 +8,7 @@ from scipy.io.wavfile import write
 import subprocess
 import sounddevice
 import numpy as np
+import torchaudio
 from pydub import AudioSegment
 from .utils import get_audio
 audio_path = os.path.join(folder_paths.get_input_directory(),"audio")
@@ -72,7 +73,7 @@ class SaveAudio:
         file_path = os.path.join(audio_path,str(time.time()).replace(".","")+".wav")
         print(audio_path)
         outfile = os.path.join(audio_path,str(time.time()).replace(".","_")+".wav")
-        open(file_path,"wb").write(audio())
+        torchaudio.save(file_path,audio["waveform"],audio["sample_rate"])
         Fs, data = wavfile.read(file_path)
         n = data.size
         t = n / Fs
@@ -104,6 +105,7 @@ class PlayBackAudio:
 
     def play_audio(self,audio):
         file = BytesIO(audio())
+		torchaudio.save(file,audio["waveform"],audio["sample_rate"])
         audio_file = AudioSegment.from_file(file, format="wav")
         audio = AudioData(audio_file)
         sounddevice.play(audio.audio_data,audio.sample_rate)