Update app.py

app.py

@@ -73,7 +73,7 @@ vocos = load_vocoder()
 
 target_sample_rate = 24000
 n_mel_channels = 100
-hop_length = 256 - 8
+hop_length = 256 - 6
 target_rms = 0.1
 nfe_step = 16  # 16, 32
 cfg_strength = 2.0
@@ -378,7 +378,7 @@ def infer(ref_audio_orig, ref_text, gen_text, exp_name, remove_silence, cross_fa
     gr.Info("Converting audio...")
     with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as f:
         aseg = AudioSegment.from_file(ref_audio_orig)
-        aseg = remove_silence_edges(aseg) + AudioSegment.silent(duration=50)
+        aseg = remove_silence_edges(aseg) + AudioSegment.silent(duration=150)
 #        non_silent_segs = silence.split_on_silence(
 #            aseg, min_silence_len=1000, silence_thresh=-42, keep_silence=1000
 #        )