Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
|
@@ -223,4 +223,114 @@ class VoiceSynthesizer:
|
|
| 223 |
|
| 224 |
return filepath, None
|
| 225 |
|
| 226 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 223 |
|
| 224 |
return filepath, None
|
| 225 |
|
| 226 |
+
def create_interface():
|
| 227 |
+
synthesizer = VoiceSynthesizer()
|
| 228 |
+
|
| 229 |
+
with gr.Blocks() as interface:
|
| 230 |
+
gr.Markdown("# 🎙️ Advanced Voice Synthesis")
|
| 231 |
+
|
| 232 |
+
with gr.Row():
|
| 233 |
+
with gr.Column():
|
| 234 |
+
gr.Markdown("## 1. Capture Reference Voice")
|
| 235 |
+
reference_audio = gr.Audio(sources=["microphone", "upload"], type="numpy")
|
| 236 |
+
process_ref_btn = gr.Button("Process Reference Voice")
|
| 237 |
+
process_ref_output = gr.Textbox(label="Reference Voice Processing")
|
| 238 |
+
|
| 239 |
+
with gr.Column():
|
| 240 |
+
gr.Markdown("## 2. Generate Speech")
|
| 241 |
+
text_input = gr.Textbox(label="Enter Text to Speak")
|
| 242 |
+
|
| 243 |
+
# Model Selection
|
| 244 |
+
model_dropdown = gr.Dropdown(
|
| 245 |
+
choices=[
|
| 246 |
+
"bark (Suno AI)",
|
| 247 |
+
"speecht5 (Microsoft)"
|
| 248 |
+
],
|
| 249 |
+
label="Select TTS Model",
|
| 250 |
+
value="bark (Suno AI)"
|
| 251 |
+
)
|
| 252 |
+
|
| 253 |
+
# Voice Preset Dropdowns
|
| 254 |
+
with gr.Row():
|
| 255 |
+
bark_preset = gr.Dropdown(
|
| 256 |
+
choices=[
|
| 257 |
+
"v2/en_speaker_6 (Female)",
|
| 258 |
+
"v2/en_speaker_3 (Male)",
|
| 259 |
+
"v2/en_speaker_9 (Neutral)"
|
| 260 |
+
],
|
| 261 |
+
label="Bark Voice Preset",
|
| 262 |
+
visible=True
|
| 263 |
+
)
|
| 264 |
+
|
| 265 |
+
speecht5_preset = gr.Dropdown(
|
| 266 |
+
choices=[
|
| 267 |
+
"Default Speaker"
|
| 268 |
+
],
|
| 269 |
+
label="SpeechT5 Speaker",
|
| 270 |
+
visible=False
|
| 271 |
+
)
|
| 272 |
+
|
| 273 |
+
generate_btn = gr.Button("Generate Speech")
|
| 274 |
+
audio_output = gr.Audio(label="Generated Speech")
|
| 275 |
+
error_output = gr.Textbox(label="Errors", visible=True)
|
| 276 |
+
|
| 277 |
+
# Process reference audio
|
| 278 |
+
process_ref_btn.click(
|
| 279 |
+
fn=synthesizer.process_reference_audio,
|
| 280 |
+
inputs=reference_audio,
|
| 281 |
+
outputs=process_ref_output
|
| 282 |
+
)
|
| 283 |
+
|
| 284 |
+
# Dynamic model and preset visibility
|
| 285 |
+
def update_model_visibility(model):
|
| 286 |
+
if "bark" in model.lower():
|
| 287 |
+
return {
|
| 288 |
+
bark_preset: gr.update(visible=True),
|
| 289 |
+
speecht5_preset: gr.update(visible=False)
|
| 290 |
+
}
|
| 291 |
+
else:
|
| 292 |
+
return {
|
| 293 |
+
bark_preset: gr.update(visible=False),
|
| 294 |
+
speecht5_preset: gr.update(visible=True)
|
| 295 |
+
}
|
| 296 |
+
|
| 297 |
+
model_dropdown.change(
|
| 298 |
+
fn=update_model_visibility,
|
| 299 |
+
inputs=model_dropdown,
|
| 300 |
+
outputs=[bark_preset, speecht5_preset]
|
| 301 |
+
)
|
| 302 |
+
|
| 303 |
+
# Speech generation logic
|
| 304 |
+
def generate_speech_wrapper(text, model, bark_preset, speecht5_preset):
|
| 305 |
+
# Map model name
|
| 306 |
+
model_map = {
|
| 307 |
+
"bark (Suno AI)": "bark",
|
| 308 |
+
"speecht5 (Microsoft)": "speecht5"
|
| 309 |
+
}
|
| 310 |
+
|
| 311 |
+
# Select appropriate preset
|
| 312 |
+
preset = bark_preset if "bark" in model else speecht5_preset
|
| 313 |
+
|
| 314 |
+
return synthesizer.generate_speech(
|
| 315 |
+
text,
|
| 316 |
+
model_name=model_map[model],
|
| 317 |
+
voice_preset=preset
|
| 318 |
+
)
|
| 319 |
+
|
| 320 |
+
generate_btn.click(
|
| 321 |
+
fn=generate_speech_wrapper,
|
| 322 |
+
inputs=[text_input, model_dropdown, bark_preset, speecht5_preset],
|
| 323 |
+
outputs=[audio_output, error_output]
|
| 324 |
+
)
|
| 325 |
+
|
| 326 |
+
return interface
|
| 327 |
+
|
| 328 |
+
if __name__ == "__main__":
|
| 329 |
+
interface = create_interface()
|
| 330 |
+
interface.launch(
|
| 331 |
+
share=False,
|
| 332 |
+
debug=True,
|
| 333 |
+
show_error=True,
|
| 334 |
+
server_name='0.0.0.0',
|
| 335 |
+
server_port=7860
|
| 336 |
+
)
|