Spaces:
Sleeping
Sleeping
File size: 20,167 Bytes
5a65ad6 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 |
"""
Voice Cloning Module
This module provides voice cloning and text-to-speech capabilities using
Coqui TTS and other state-of-the-art TTS models.
"""
import os
import logging
from typing import Dict, List, Optional, Union, Any
from pathlib import Path
import json
import torch
import numpy as np
from TTS.api import TTS
from TTS.tts.configs.xtts_config import XttsConfig
from TTS.tts.models.xtts import Xtts
import soundfile as sf
from ..config import TTS_MODEL, VOICE_CLONE_SAMPLES_MIN, VOICE_CLONE_DURATION_MIN, SAMPLE_RATE
from ..audio_processing.processor import AudioProcessor
class VoiceCloner:
"""Voice cloning using Coqui TTS models."""
def __init__(
self,
model_name: str = TTS_MODEL,
device: str = "auto",
use_gpu: bool = True
):
"""
Initialize voice cloner.
Args:
model_name: TTS model name
device: Device to run model on
use_gpu: Whether to use GPU acceleration
"""
self.model_name = model_name
self.device = self._setup_device(device, use_gpu)
self.tts = None
self.model = None
self.audio_processor = AudioProcessor()
self.logger = logging.getLogger(__name__)
# Voice sample management
self.voice_samples = {}
self.speaker_embeddings = {}
def _setup_device(self, device: str, use_gpu: bool) -> str:
"""Setup device configuration."""
if device == "auto":
if use_gpu and torch.cuda.is_available():
return "cuda"
else:
return "cpu"
return device
def load_model(self) -> None:
"""Load the TTS model."""
try:
self.logger.info(f"Loading TTS model: {self.model_name}")
# Initialize TTS
self.tts = TTS(
model_name=self.model_name,
progress_bar=True,
gpu=(self.device == "cuda")
)
self.logger.info("TTS model loaded successfully")
except Exception as e:
self.logger.error(f"Failed to load TTS model: {str(e)}")
raise RuntimeError(f"TTS model loading failed: {str(e)}")
def register_voice(
self,
speaker_name: str,
voice_samples: List[Union[str, Path]],
validate: bool = True
) -> Dict[str, Any]:
"""
Register a new voice with audio samples.
Args:
speaker_name: Unique identifier for the speaker
voice_samples: List of paths to voice sample files
validate: Whether to validate voice samples
Returns:
Dictionary with registration results
"""
try:
self.logger.info(f"Registering voice: {speaker_name}")
if validate:
validation_result = self._validate_voice_samples(voice_samples)
if not validation_result['valid']:
raise ValueError(f"Voice sample validation failed: {validation_result['errors']}")
# Process voice samples
processed_samples = []
total_duration = 0.0
for sample_path in voice_samples:
# Load and process audio
audio_data = self.audio_processor.load_audio(sample_path, normalize=True)
# Calculate duration
duration = len(audio_data) / SAMPLE_RATE
total_duration += duration
processed_samples.append({
'path': str(sample_path),
'audio_data': audio_data,
'duration': duration
})
# Store voice information
self.voice_samples[speaker_name] = {
'samples': processed_samples,
'total_duration': total_duration,
'num_samples': len(processed_samples),
'registered_at': self._get_timestamp()
}
# Generate speaker embedding if using XTTS
if "xtts" in self.model_name.lower():
self._generate_speaker_embedding(speaker_name)
result = {
'speaker_name': speaker_name,
'num_samples': len(processed_samples),
'total_duration': total_duration,
'status': 'registered'
}
self.logger.info(f"Voice registered successfully: {speaker_name} "
f"({len(processed_samples)} samples, {total_duration:.1f}s)")
return result
except Exception as e:
self.logger.error(f"Voice registration failed: {str(e)}")
raise RuntimeError(f"Voice registration failed: {str(e)}")
def _validate_voice_samples(self, voice_samples: List[Union[str, Path]]) -> Dict[str, Any]:
"""Validate voice samples."""
validation_result = {
'valid': True,
'errors': [],
'warnings': [],
'info': {}
}
if len(voice_samples) < VOICE_CLONE_SAMPLES_MIN:
validation_result['errors'].append(
f"Need at least {VOICE_CLONE_SAMPLES_MIN} voice samples, got {len(voice_samples)}"
)
validation_result['valid'] = False
total_duration = 0.0
valid_samples = 0
for sample_path in voice_samples:
try:
# Validate individual file
file_validation = self.audio_processor.get_audio_info(sample_path)
total_duration += file_validation['duration']
valid_samples += 1
# Check sample quality
if file_validation['duration'] < 3.0:
validation_result['warnings'].append(
f"Short sample ({file_validation['duration']:.1f}s): {sample_path}"
)
if file_validation['sample_rate'] < 16000:
validation_result['warnings'].append(
f"Low sample rate ({file_validation['sample_rate']} Hz): {sample_path}"
)
except Exception as e:
validation_result['errors'].append(f"Invalid sample {sample_path}: {str(e)}")
if total_duration < VOICE_CLONE_DURATION_MIN:
validation_result['errors'].append(
f"Total duration ({total_duration:.1f}s) below minimum ({VOICE_CLONE_DURATION_MIN}s)"
)
validation_result['valid'] = False
validation_result['info'] = {
'total_samples': len(voice_samples),
'valid_samples': valid_samples,
'total_duration': total_duration
}
return validation_result
def _generate_speaker_embedding(self, speaker_name: str) -> None:
"""Generate speaker embedding for XTTS models."""
if self.tts is None:
self.load_model()
try:
voice_data = self.voice_samples[speaker_name]
# Concatenate all samples for embedding generation
combined_audio = []
for sample in voice_data['samples']:
combined_audio.extend(sample['audio_data'])
# Convert to tensor and generate embedding
audio_tensor = torch.FloatTensor(combined_audio).unsqueeze(0)
# This is a placeholder - actual implementation depends on TTS model
# For XTTS, you might use the model's speaker encoder
self.logger.info(f"Generated speaker embedding for {speaker_name}")
except Exception as e:
self.logger.warning(f"Failed to generate speaker embedding: {str(e)}")
def clone_voice(
self,
text: str,
speaker_name: str,
language: str = "en",
output_path: Optional[Union[str, Path]] = None,
**kwargs
) -> Dict[str, Any]:
"""
Generate speech using cloned voice.
Args:
text: Text to synthesize
speaker_name: Registered speaker name
language: Target language
output_path: Output file path (optional)
**kwargs: Additional TTS parameters
Returns:
Dictionary with synthesis results
"""
if self.tts is None:
self.load_model()
if speaker_name not in self.voice_samples:
raise ValueError(f"Speaker '{speaker_name}' not registered")
try:
self.logger.info(f"Generating speech for '{speaker_name}': {text[:50]}...")
# Get voice samples for the speaker
voice_data = self.voice_samples[speaker_name]
# Use first sample as reference (could be improved by selecting best sample)
reference_audio_path = voice_data['samples'][0]['path']
# Generate speech
if "xtts" in self.model_name.lower():
# XTTS-specific generation
audio = self._generate_xtts(text, reference_audio_path, language, **kwargs)
else:
# Generic TTS generation
audio = self._generate_generic_tts(text, reference_audio_path, language, **kwargs)
# Save audio if output path provided
if output_path:
output_path = Path(output_path)
self.audio_processor.save_audio(audio, output_path)
self.logger.info(f"Saved generated audio to: {output_path}")
result = {
'text': text,
'speaker_name': speaker_name,
'language': language,
'audio_data': audio,
'sample_rate': SAMPLE_RATE,
'duration': len(audio) / SAMPLE_RATE,
'output_path': str(output_path) if output_path else None,
'model_used': self.model_name
}
self.logger.info(f"Voice cloning completed: {result['duration']:.1f}s audio generated")
return result
except Exception as e:
self.logger.error(f"Voice cloning failed: {str(e)}")
raise RuntimeError(f"Voice cloning failed: {str(e)}")
def _generate_xtts(
self,
text: str,
reference_audio_path: str,
language: str,
**kwargs
) -> np.ndarray:
"""Generate speech using XTTS model."""
try:
# XTTS generation
audio = self.tts.tts(
text=text,
speaker_wav=reference_audio_path,
language=language,
**kwargs
)
return np.array(audio, dtype=np.float32)
except Exception as e:
self.logger.error(f"XTTS generation failed: {str(e)}")
raise RuntimeError(f"XTTS generation failed: {str(e)}")
def _generate_generic_tts(
self,
text: str,
reference_audio_path: str,
language: str,
**kwargs
) -> np.ndarray:
"""Generate speech using generic TTS model."""
try:
# Generic TTS generation
audio = self.tts.tts(
text=text,
speaker_wav=reference_audio_path,
**kwargs
)
return np.array(audio, dtype=np.float32)
except Exception as e:
self.logger.error(f"Generic TTS generation failed: {str(e)}")
raise RuntimeError(f"Generic TTS generation failed: {str(e)}")
def get_registered_speakers(self) -> List[str]:
"""Get list of registered speakers."""
return list(self.voice_samples.keys())
def get_speaker_info(self, speaker_name: str) -> Dict[str, Any]:
"""Get information about a registered speaker."""
if speaker_name not in self.voice_samples:
raise ValueError(f"Speaker '{speaker_name}' not found")
voice_data = self.voice_samples[speaker_name]
return {
'speaker_name': speaker_name,
'num_samples': voice_data['num_samples'],
'total_duration': voice_data['total_duration'],
'registered_at': voice_data['registered_at'],
'samples': [sample['path'] for sample in voice_data['samples']]
}
def remove_speaker(self, speaker_name: str) -> bool:
"""Remove a registered speaker."""
if speaker_name in self.voice_samples:
del self.voice_samples[speaker_name]
if speaker_name in self.speaker_embeddings:
del self.speaker_embeddings[speaker_name]
self.logger.info(f"Removed speaker: {speaker_name}")
return True
return False
def save_speaker_data(self, output_dir: Union[str, Path]) -> None:
"""Save speaker data to disk."""
output_dir = Path(output_dir)
output_dir.mkdir(parents=True, exist_ok=True)
# Save voice sample metadata
metadata_file = output_dir / "speakers_metadata.json"
metadata = {}
for speaker_name, voice_data in self.voice_samples.items():
metadata[speaker_name] = {
'num_samples': voice_data['num_samples'],
'total_duration': voice_data['total_duration'],
'registered_at': voice_data['registered_at'],
'sample_paths': [sample['path'] for sample in voice_data['samples']]
}
with open(metadata_file, 'w') as f:
json.dump(metadata, f, indent=2)
self.logger.info(f"Saved speaker metadata to: {metadata_file}")
def load_speaker_data(self, input_dir: Union[str, Path]) -> None:
"""Load speaker data from disk."""
input_dir = Path(input_dir)
metadata_file = input_dir / "speakers_metadata.json"
if not metadata_file.exists():
self.logger.warning(f"Speaker metadata not found: {metadata_file}")
return
try:
with open(metadata_file, 'r') as f:
metadata = json.load(f)
for speaker_name, speaker_data in metadata.items():
# Re-register speaker with existing samples
sample_paths = speaker_data['sample_paths']
# Validate that samples still exist
valid_samples = [path for path in sample_paths if Path(path).exists()]
if valid_samples:
self.register_voice(speaker_name, valid_samples, validate=False)
self.logger.info(f"Loaded speaker: {speaker_name}")
else:
self.logger.warning(f"No valid samples found for speaker: {speaker_name}")
except Exception as e:
self.logger.error(f"Failed to load speaker data: {str(e)}")
def _get_timestamp(self) -> str:
"""Get current timestamp."""
import datetime
return datetime.datetime.now().isoformat()
def get_model_info(self) -> Dict[str, Any]:
"""Get information about the loaded model."""
return {
'model_name': self.model_name,
'device': self.device,
'model_loaded': self.tts is not None,
'num_registered_speakers': len(self.voice_samples),
'cuda_available': torch.cuda.is_available()
}
class BatchVoiceCloner:
"""Batch processing for voice cloning tasks."""
def __init__(self, voice_cloner: VoiceCloner):
"""
Initialize batch voice cloner.
Args:
voice_cloner: VoiceCloner instance
"""
self.voice_cloner = voice_cloner
self.logger = logging.getLogger(__name__)
def clone_batch(
self,
texts: List[str],
speaker_name: str,
language: str = "en",
output_dir: Optional[Union[str, Path]] = None,
**kwargs
) -> Dict[str, Any]:
"""
Generate speech for multiple texts using the same voice.
Args:
texts: List of texts to synthesize
speaker_name: Registered speaker name
language: Target language
output_dir: Directory to save output files
**kwargs: Additional TTS parameters
Returns:
Dictionary with batch processing results
"""
results = []
failed_texts = []
if output_dir:
output_dir = Path(output_dir)
output_dir.mkdir(parents=True, exist_ok=True)
self.logger.info(f"Starting batch voice cloning: {len(texts)} texts")
for i, text in enumerate(texts, 1):
try:
self.logger.info(f"Processing text {i}/{len(texts)}")
# Generate output path if directory provided
output_path = None
if output_dir:
output_path = output_dir / f"speech_{i:04d}.wav"
result = self.voice_cloner.clone_voice(
text=text,
speaker_name=speaker_name,
language=language,
output_path=output_path,
**kwargs
)
results.append(result)
except Exception as e:
self.logger.error(f"Failed to process text {i}: {str(e)}")
failed_texts.append({'index': i, 'text': text, 'error': str(e)})
batch_result = {
'total_texts': len(texts),
'successful': len(results),
'failed': len(failed_texts),
'results': results,
'failed_texts': failed_texts,
'speaker_name': speaker_name,
'language': language
}
self.logger.info(f"Batch voice cloning completed. "
f"Success: {batch_result['successful']}, "
f"Failed: {batch_result['failed']}")
return batch_result
# Utility functions
def create_voice_cloner(
model_name: str = TTS_MODEL,
device: str = "auto"
) -> VoiceCloner:
"""Create and initialize voice cloner."""
cloner = VoiceCloner(model_name=model_name, device=device)
cloner.load_model()
return cloner
def quick_voice_clone(
text: str,
voice_sample_path: str,
output_path: str,
language: str = "en"
) -> str:
"""Quick voice cloning for simple use cases."""
cloner = create_voice_cloner()
# Register temporary speaker
temp_speaker = "temp_speaker"
cloner.register_voice(temp_speaker, [voice_sample_path])
# Generate speech
result = cloner.clone_voice(
text=text,
speaker_name=temp_speaker,
language=language,
output_path=output_path
)
return str(result['output_path']) |