Update app.py

app.py

@@ -47,12 +47,18 @@ def generate_segment_audio(text, lang, speaker_url, pipe):
     audio_np = audio_data_resampled.cpu().numpy()
     return audio_np
 
-# Function to concatenate audio segments in stereo
 def concatenate_audio_segments(segments):
-    total_length = sum(seg.shape[0] for seg in segments)
+    stereo_segments = []
+    for seg in segments:
+        if seg.ndim == 1 or seg.shape[1] == 1:
+            stereo_seg = np.stack((seg, seg), axis=-1)
+        else:
+            stereo_seg = seg
+        stereo_segments.append(stereo_seg)
+    total_length = sum(seg.shape[0] for seg in stereo_segments)
     concatenated_audio = np.zeros((total_length, 2), dtype=np.float32)
     current_index = 0
-    for seg in segments:
+    for seg in stereo_segments:
         end_index = current_index + seg.shape[0]
         concatenated_audio[current_index:end_index, :] = seg
         current_index = end_index