Update app.py

app.py

@@ -3,10 +3,12 @@ import os
 from uuid import uuid4
 from pydub.silence import detect_nonsilent
 from pydub import AudioSegment
+from pydub.effects import normalize
  
 
 def get_labels(audio_fp, threshold, min_speech_duration_ms, min_silence_duration_ms, auto_merge, uppper_merge_threshold, max_segment_length):
-    audio = AudioSegment.from_file(audio_fp)
+    audio = AudioSegment.from_file(audio_fp).set_channels(1)
+    audio = normalize(audio)
     speech_timestamps = detect_nonsilent(audio, min_silence_len=min_silence_duration_ms, silence_thresh=-40)
     speech_timestamps = list(filter(lambda x: x[1]-x[0] > min_speech_duration_ms, speech_timestamps))
     speech_timestamps = [{"start": s[0]/1000, "end": s[1]/1000} for s in speech_timestamps]
@@ -61,8 +63,8 @@ interface = gr.Interface(
         gr.Number(label="min_speech_duration_ms", value=250, info="default (250)"), 
         gr.Number(label="min_silence_duration_ms", value=40, info="default (100)"),
         gr.Checkbox(label="Auto merge", value=True),
-        gr.Textbox(label="Gap max threshold value (seconds)", value=0.3),
-        gr.Number(label="Approx Max Segment Length", value=5)
+        gr.Textbox(label="Gap max threshold value (seconds)", value=0.350),
+        gr.Number(label="Approx Max Segment Length", value=10)
     ],
     [
         gr.File(label="VAD Labels"),