Spaces:
Running
Running
File size: 17,068 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 | """
Audio Processing Module
This module provides comprehensive audio processing capabilities including
format conversion, quality enhancement, and preprocessing for the speech
translation system.
"""
import os
import logging
from typing import Optional, Union, Tuple, List
from pathlib import Path
import numpy as np
import librosa
import soundfile as sf
from pydub import AudioSegment
from scipy import signal
import torch
import torchaudio
from ..config import SAMPLE_RATE, MAX_AUDIO_DURATION, AUDIO_FORMATS
class AudioProcessor:
"""Handles audio file processing, conversion, and enhancement."""
def __init__(self, target_sample_rate: int = SAMPLE_RATE):
"""
Initialize the audio processor.
Args:
target_sample_rate: Target sample rate for processing
"""
self.target_sample_rate = target_sample_rate
self.max_duration = MAX_AUDIO_DURATION
self.supported_formats = AUDIO_FORMATS
self.logger = logging.getLogger(__name__)
def load_audio(
self,
audio_path: Union[str, Path],
normalize: bool = True,
mono: bool = True
) -> np.ndarray:
"""
Load audio file and convert to target format.
Args:
audio_path: Path to audio file
normalize: Whether to normalize audio amplitude
mono: Whether to convert to mono
Returns:
Audio data as numpy array
"""
audio_path = Path(audio_path)
if not audio_path.exists():
raise FileNotFoundError(f"Audio file not found: {audio_path}")
if audio_path.suffix.lower() not in self.supported_formats:
raise ValueError(f"Unsupported audio format: {audio_path.suffix}")
try:
self.logger.debug(f"Loading audio: {audio_path}")
# Load audio using librosa (handles most formats)
audio_data, sample_rate = librosa.load(
str(audio_path),
sr=self.target_sample_rate,
mono=mono,
dtype=np.float32
)
# Validate duration
duration = len(audio_data) / self.target_sample_rate
if duration > self.max_duration:
self.logger.warning(f"Audio duration ({duration:.1f}s) exceeds maximum "
f"({self.max_duration}s). Truncating.")
audio_data = audio_data[:int(self.max_duration * self.target_sample_rate)]
# Normalize amplitude if requested
if normalize:
audio_data = self.normalize_audio(audio_data)
self.logger.debug(f"Loaded audio: duration={duration:.2f}s, "
f"sample_rate={self.target_sample_rate}, shape={audio_data.shape}")
return audio_data
except Exception as e:
self.logger.error(f"Failed to load audio {audio_path}: {str(e)}")
raise RuntimeError(f"Audio loading failed: {str(e)}")
def save_audio(
self,
audio_data: np.ndarray,
output_path: Union[str, Path],
sample_rate: Optional[int] = None,
format: Optional[str] = None
) -> None:
"""
Save audio data to file.
Args:
audio_data: Audio data as numpy array
output_path: Output file path
sample_rate: Sample rate (uses target_sample_rate if None)
format: Audio format (inferred from extension if None)
"""
output_path = Path(output_path)
sample_rate = sample_rate or self.target_sample_rate
try:
# Create output directory if needed
output_path.parent.mkdir(parents=True, exist_ok=True)
# Determine format from extension if not specified
if format is None:
format = output_path.suffix.lower().lstrip('.')
# Ensure audio data is in correct range for format
if format in ['wav', 'flac']:
# For lossless formats, keep full precision
sf.write(str(output_path), audio_data, sample_rate, format=format.upper())
else:
# For compressed formats, use pydub
self._save_with_pydub(audio_data, output_path, sample_rate, format)
self.logger.debug(f"Saved audio to: {output_path}")
except Exception as e:
self.logger.error(f"Failed to save audio to {output_path}: {str(e)}")
raise RuntimeError(f"Audio saving failed: {str(e)}")
def _save_with_pydub(
self,
audio_data: np.ndarray,
output_path: Path,
sample_rate: int,
format: str
) -> None:
"""Save audio using pydub for compressed formats."""
# Convert to 16-bit PCM for pydub
audio_16bit = (audio_data * 32767).astype(np.int16)
# Create AudioSegment
audio_segment = AudioSegment(
audio_16bit.tobytes(),
frame_rate=sample_rate,
sample_width=2,
channels=1
)
# Export with format-specific settings
export_params = {}
if format == 'mp3':
export_params['bitrate'] = '192k'
elif format == 'ogg':
export_params['codec'] = 'libvorbis'
audio_segment.export(str(output_path), format=format, **export_params)
def convert_format(
self,
input_path: Union[str, Path],
output_path: Union[str, Path],
target_format: str = 'wav'
) -> None:
"""
Convert audio file to different format.
Args:
input_path: Input audio file path
output_path: Output audio file path
target_format: Target audio format
"""
audio_data = self.load_audio(input_path)
# Update output path extension if needed
output_path = Path(output_path)
if output_path.suffix.lower() != f'.{target_format}':
output_path = output_path.with_suffix(f'.{target_format}')
self.save_audio(audio_data, output_path, format=target_format)
self.logger.info(f"Converted {input_path} to {output_path} ({target_format})")
def normalize_audio(self, audio_data: np.ndarray, target_db: float = -20.0) -> np.ndarray:
"""
Normalize audio amplitude.
Args:
audio_data: Input audio data
target_db: Target RMS level in dB
Returns:
Normalized audio data
"""
# Calculate RMS
rms = np.sqrt(np.mean(audio_data ** 2))
if rms > 0:
# Convert target dB to linear scale
target_linear = 10 ** (target_db / 20.0)
# Calculate scaling factor
scale_factor = target_linear / rms
# Apply scaling with clipping prevention
normalized = audio_data * scale_factor
normalized = np.clip(normalized, -0.95, 0.95)
return normalized
return audio_data
def remove_silence(
self,
audio_data: np.ndarray,
threshold_db: float = -40.0,
frame_length: int = 2048,
hop_length: int = 512
) -> np.ndarray:
"""
Remove silence from audio.
Args:
audio_data: Input audio data
threshold_db: Silence threshold in dB
frame_length: Frame length for analysis
hop_length: Hop length for analysis
Returns:
Audio data with silence removed
"""
# Calculate frame-wise energy
frames = librosa.util.frame(
audio_data,
frame_length=frame_length,
hop_length=hop_length
)
energy = np.sum(frames ** 2, axis=0)
# Convert to dB
energy_db = librosa.power_to_db(energy)
# Find non-silent frames
non_silent = energy_db > threshold_db
if not np.any(non_silent):
self.logger.warning("No non-silent frames found, returning original audio")
return audio_data
# Convert frame indices to sample indices
start_frame = np.argmax(non_silent)
end_frame = len(non_silent) - np.argmax(non_silent[::-1]) - 1
start_sample = start_frame * hop_length
end_sample = min(len(audio_data), (end_frame + 1) * hop_length + frame_length)
return audio_data[start_sample:end_sample]
def apply_noise_reduction(
self,
audio_data: np.ndarray,
noise_factor: float = 0.1
) -> np.ndarray:
"""
Apply basic noise reduction using spectral subtraction.
Args:
audio_data: Input audio data
noise_factor: Noise reduction factor (0.0 to 1.0)
Returns:
Noise-reduced audio data
"""
# Compute STFT
stft = librosa.stft(audio_data)
magnitude, phase = np.abs(stft), np.angle(stft)
# Estimate noise from first few frames (assume silence)
noise_frames = min(10, magnitude.shape[1] // 4)
noise_spectrum = np.mean(magnitude[:, :noise_frames], axis=1, keepdims=True)
# Apply spectral subtraction
magnitude_clean = magnitude - (noise_factor * noise_spectrum)
magnitude_clean = np.maximum(magnitude_clean, 0.1 * magnitude)
# Reconstruct signal
stft_clean = magnitude_clean * np.exp(1j * phase)
audio_clean = librosa.istft(stft_clean)
return audio_clean
def resample_audio(
self,
audio_data: np.ndarray,
original_sr: int,
target_sr: int
) -> np.ndarray:
"""
Resample audio to different sample rate.
Args:
audio_data: Input audio data
original_sr: Original sample rate
target_sr: Target sample rate
Returns:
Resampled audio data
"""
if original_sr == target_sr:
return audio_data
return librosa.resample(audio_data, orig_sr=original_sr, target_sr=target_sr)
def split_audio(
self,
audio_data: np.ndarray,
chunk_duration: float = 30.0,
overlap: float = 0.5
) -> List[np.ndarray]:
"""
Split audio into overlapping chunks.
Args:
audio_data: Input audio data
chunk_duration: Duration of each chunk in seconds
overlap: Overlap between chunks (0.0 to 1.0)
Returns:
List of audio chunks
"""
chunk_samples = int(chunk_duration * self.target_sample_rate)
overlap_samples = int(chunk_samples * overlap)
step_samples = chunk_samples - overlap_samples
chunks = []
start = 0
while start < len(audio_data):
end = min(start + chunk_samples, len(audio_data))
chunk = audio_data[start:end]
# Pad last chunk if needed
if len(chunk) < chunk_samples:
chunk = np.pad(chunk, (0, chunk_samples - len(chunk)))
chunks.append(chunk)
if end >= len(audio_data):
break
start += step_samples
return chunks
def get_audio_info(self, audio_path: Union[str, Path]) -> dict:
"""
Get audio file information.
Args:
audio_path: Path to audio file
Returns:
Dictionary with audio information
"""
try:
# Use librosa for detailed info
audio_data, sample_rate = librosa.load(str(audio_path), sr=None)
duration = len(audio_data) / sample_rate
# Get file size
file_size = Path(audio_path).stat().st_size
info = {
'path': str(audio_path),
'duration': duration,
'sample_rate': sample_rate,
'channels': 1 if audio_data.ndim == 1 else audio_data.shape[0],
'samples': len(audio_data),
'file_size': file_size,
'format': Path(audio_path).suffix.lower(),
'bit_depth': 'float32', # librosa loads as float32
'rms_level': float(np.sqrt(np.mean(audio_data ** 2))),
'max_level': float(np.max(np.abs(audio_data)))
}
return info
except Exception as e:
self.logger.error(f"Failed to get audio info for {audio_path}: {str(e)}")
raise RuntimeError(f"Audio info extraction failed: {str(e)}")
class AudioValidator:
"""Validates audio files and data."""
def __init__(self, processor: AudioProcessor):
"""
Initialize audio validator.
Args:
processor: AudioProcessor instance
"""
self.processor = processor
self.logger = logging.getLogger(__name__)
def validate_audio_file(self, audio_path: Union[str, Path]) -> dict:
"""
Validate audio file.
Args:
audio_path: Path to audio file
Returns:
Dictionary with validation results
"""
validation_result = {
'valid': False,
'errors': [],
'warnings': [],
'info': {}
}
try:
# Check if file exists
audio_path = Path(audio_path)
if not audio_path.exists():
validation_result['errors'].append(f"File does not exist: {audio_path}")
return validation_result
# Check file format
if audio_path.suffix.lower() not in self.processor.supported_formats:
validation_result['errors'].append(
f"Unsupported format: {audio_path.suffix}"
)
return validation_result
# Get audio info
info = self.processor.get_audio_info(audio_path)
validation_result['info'] = info
# Check duration
if info['duration'] > self.processor.max_duration:
validation_result['warnings'].append(
f"Duration ({info['duration']:.1f}s) exceeds maximum "
f"({self.processor.max_duration}s)"
)
# Check sample rate
if info['sample_rate'] < 8000:
validation_result['warnings'].append(
f"Low sample rate ({info['sample_rate']} Hz) may affect quality"
)
# Check audio level
if info['max_level'] < 0.01:
validation_result['warnings'].append("Audio level is very low")
elif info['max_level'] > 0.99:
validation_result['warnings'].append("Audio may be clipped")
# If we get here, file is valid
validation_result['valid'] = True
except Exception as e:
validation_result['errors'].append(str(e))
return validation_result
def validate_batch(self, audio_files: List[Union[str, Path]]) -> dict:
"""
Validate multiple audio files.
Args:
audio_files: List of audio file paths
Returns:
Dictionary with batch validation results
"""
results = {}
valid_count = 0
for audio_file in audio_files:
result = self.validate_audio_file(audio_file)
results[str(audio_file)] = result
if result['valid']:
valid_count += 1
return {
'total_files': len(audio_files),
'valid_files': valid_count,
'invalid_files': len(audio_files) - valid_count,
'results': results
} |