Update app.py

app.py

@@ -13,6 +13,11 @@ def get_audio_length(audio_path):
         rate = audio.getframerate()
         return frames / float(rate)
 
+# Ensure that the file starts with "RIFF" (WAV header)
+def is_wav_file(filepath):
+    with open(filepath, 'rb') as f:
+        return f.read(4) == b'RIFF'
+
 # Generate precise SRT entries for a text batch
 def generate_accurate_srt(text, start_time, batch_index):
     srt_entries = []
@@ -45,17 +50,21 @@ async def batch_process_srt_and_audio(script_text, voice, batch_size=500, progre
         batch_text = script_text[i:i+batch_size]
         audio_file = f"audio_batch_{i}.wav"
         
-        # Generate audio for each batch
+        # Generate audio for each batch and save as WAV
         tts = edge_tts.Communicate(batch_text, voice, rate="-25%")
         await tts.save(audio_file)
         
+        # Check if saved file is a valid WAV file
+        if not is_wav_file(audio_file):
+            raise ValueError(f"Audio file {audio_file} is not a valid WAV file.")
+        
         # Get precise audio length for synchronization
         batch_duration = get_audio_length(audio_file)
         srt_entries, cumulative_time = generate_accurate_srt(batch_text, cumulative_time, batch_index)
         
         # Append entries and audio for the batch
         total_srt_entries.extend(srt_entries)
-        batch_audio = AudioSegment.from_file(audio_file)
+        batch_audio = AudioSegment.from_file(audio_file, format="wav")  # Explicitly specify format
         combined_audio += batch_audio
         batch_index += len(srt_entries)