Spaces:
Sleeping
Sleeping
File size: 2,537 Bytes
1c19314 |
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 |
import librosa
import numpy as np
def extract_features(file_path):
try:
print(f"🟢 Processing file: {file_path}")
# Load audio with fixed sample rate
y, sr = librosa.load(file_path, sr=16000)
# Trim or pad audio to exactly 5 seconds
target_length = sr * 5
if len(y) > target_length:
start_sample = np.random.randint(0, len(y) - target_length)
y = y[start_sample:start_sample + target_length]
elif len(y) < target_length:
y = np.pad(y, (0, target_length - len(y)), mode='constant')
print("✅ Audio loaded and standardized (5s, 16kHz)")
# Extract features
mfcc = librosa.feature.mfcc(y=y, sr=sr, n_mfcc=13)
chroma = librosa.feature.chroma_stft(y=y, sr=sr)
spec_contrast = librosa.feature.spectral_contrast(y=y, sr=sr)
zcr = librosa.feature.zero_crossing_rate(y)
rms = librosa.feature.rms(y=y)
centroid = librosa.feature.spectral_centroid(y=y, sr=sr)
bandwidth = librosa.feature.spectral_bandwidth(y=y, sr=sr)
rolloff = librosa.feature.spectral_rolloff(y=y, sr=sr)
hnr = librosa.effects.harmonic(y)
pitches, _ = librosa.piptrack(y=y, sr=sr)
# Aggregate features (mean + std for each feature)
features = {
"mfcc": np.concatenate([np.mean(mfcc, axis=1), np.std(mfcc, axis=1)]),
"chroma": np.concatenate([np.mean(chroma, axis=1), np.std(chroma, axis=1)]),
"spectral_contrast": np.concatenate([np.mean(spec_contrast, axis=1), np.std(spec_contrast, axis=1)]),
"zcr": [np.mean(zcr), np.std(zcr)],
"rms": [np.mean(rms), np.std(rms)],
"centroid": [np.mean(centroid), np.std(centroid)],
"bandwidth": [np.mean(bandwidth), np.std(bandwidth)],
"rolloff": [np.mean(rolloff), np.std(rolloff)],
"hnr": [np.mean(hnr), np.std(hnr)],
"pitch": [np.mean(pitches), np.std(pitches)]
}
# Flatten all feature arrays into a single list
feature_vector = []
for value in features.values():
feature_vector.extend(value)
print(f"✅ Features successfully extracted and formatted (Total: {len(feature_vector)} features)")
return feature_vector
except Exception as e:
print(f"❌ Error extracting features from {file_path}: {e}")
return None |