File size: 3,864 Bytes
a361db3 | 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 | """
Compute TDOA-based DoA for original 4-channel audio files
and update results.json with actual spatial estimates
"""
import json
from pathlib import Path
import soundfile as sf
from .doa_tdoa import compute_tdoa_based_doa_for_mixture
def analyze_4ch_doa(input_4ch_file, output_dir):
"""
Analyze a 4-channel audio file for Direction of Arrival
Args:
input_4ch_file: Path to 4-channel WAV
output_dir: Output directory where results.json exists
"""
# Load 4-channel audio (downsample for faster TDOA computation)
audio_4ch, sr = sf.read(str(input_4ch_file))
# Downsample to 16 kHz for faster processing while maintaining accuracy
downsample_factor = sr // 16000
if downsample_factor > 1:
audio_4ch = audio_4ch[::downsample_factor]
sr_analysis = 16000
else:
sr_analysis = sr
print(f"\nAnalyzing {input_4ch_file}")
print(f" Original shape: {audio_4ch.shape}")
print(f" Analysis sample rate: {sr_analysis} Hz (downsampled by {downsample_factor}x)")
# Compute TDOA-based DoA
print("\nComputing TDOA-based Direction of Arrival...")
doa_result = compute_tdoa_based_doa_for_mixture(audio_4ch, sr_analysis)
print(f" Left-Right angle: {doa_result['lr_angle']:.1f}° (xcorr: {doa_result['lr_xcorr']:.3f})")
print(f" Front-Rear angle: {doa_result['fb_angle']:.1f}° (xcorr: {doa_result['fb_xcorr']:.3f})")
# Update results.json with DoA information
results_file = Path(output_dir) / 'results.json'
if results_file.exists():
with open(results_file) as f:
results = json.load(f)
# Add TDOA DoA analysis
results['tdoa_analysis'] = {
'input_4ch_file': str(input_4ch_file),
'method': 'TDOA (Time Difference of Arrival) cross-correlation',
'microphone_pairs': {
'front_pair_lr': {
'channels': [0, 2],
'labels': ['LF', 'RF'],
'lr_angle_degrees': float(doa_result['lr_angle']),
'xcorr_confidence': float(doa_result['lr_xcorr']),
'interpretation': 'Negative = left, Positive = right'
},
'front_rear_pair': {
'channels': [0, 1],
'labels': ['LF', 'LR'],
'fb_angle_degrees': float(doa_result['fb_angle']),
'xcorr_confidence': float(doa_result['fb_xcorr']),
'interpretation': 'Negative = front, Positive = rear'
}
},
'compass_angles': {
'0_degrees': 'Front (center)',
'90_degrees': 'Right',
'180_degrees': 'Rear',
'270_degrees': 'Left',
'note': 'Sound source direction relative to listener'
}
}
# Update direction_of_arrival field with actual result
results['processing_methods']['direction_of_arrival'] = 'TDOA cross-correlation (4-channel analysis)'
# Save updated results
with open(results_file, 'w') as f:
json.dump(results, f, indent=2)
print(f"\n✓ Updated {results_file}")
print(" Added TDOA analysis with spatial estimates")
return doa_result
else:
print(f"✗ {results_file} not found")
return None
if __name__ == '__main__':
# Analyze both files
files = [
('data/example_mixture.wav', 'output/analysis_example_frankenstein'),
('data/mixture.wav', 'output/analysis_frankenstein'),
]
print("=" * 60)
print("TDOA-Based Direction of Arrival Analysis")
print("=" * 60)
for input_file, output_dir in files:
analyze_4ch_doa(input_file, output_dir)
print("\n" + "=" * 60)
print("✓ TDOA analysis complete")
print("=" * 60)
|