back to file based audio output

app.py

@@ -13,27 +13,38 @@ def synthesize_speech(text):
  # Check for NSFW content
  nsfw_result = nsfw_detector(text)
  if nsfw_result[0]['label'] == 'NSFW':
- yield "NSFW content detected. Cannot process."
- return
+ return "NSFW content detected. Cannot process.", None
 
  model_path = hf_hub_download(repo_id="aigmixer/speaker_00", filename="speaker_00_model.onnx")
  config_path = hf_hub_download(repo_id="aigmixer/speaker_00", filename="speaker_00_model.onnx.json")
  voice = PiperVoice.load(model_path, config_path)
 
- # Synthesize speech and stream audio
- for audio_chunk in voice.synthesize(text, chunk_size=2048):
- yield audio_chunk.tobytes()
+ # Create an in-memory buffer for the WAV file
+ buffer = BytesIO()
+ with wave.open(buffer, 'wb') as wav_file:
+ wav_file.setframerate(voice.config.sample_rate)
+ wav_file.setsampwidth(2) # 16-bit
+ wav_file.setnchannels(1) # mono
+
+ # Synthesize speech
+ voice.synthesize(text, wav_file)
+
+ # Convert buffer to NumPy array for Gradio output
+ buffer.seek(0)
+ audio_data = np.frombuffer(buffer.read(), dtype=np.int16)
+
+ return audio_data.tobytes(), None
 
 # Using Gradio Blocks
 with gr.Blocks(theme=gr.themes.Base()) as blocks:
  gr.Markdown("# Text to Speech Synthesizer")
  gr.Markdown("Enter text to synthesize it into speech using PiperVoice.")
  input_text = gr.Textbox(label="Input Text")
- output_audio = gr.Audio(label="Synthesized Speech", type="numpy", streaming=True)
+ output_audio = gr.Audio(label="Synthesized Speech", type="numpy")
+ output_text = gr.Textbox(label="Output Text", visible=False) # This is the new text output component
  submit_button = gr.Button("Synthesize")
 
- submit_button.click(synthesize_speech, inputs=input_text, outputs=output_audio)
+ submit_button.click(synthesize_speech, inputs=input_text, outputs=[output_audio, output_text])
 
 # Run the app
-blocks.launch()
-
+blocks.launch()