|
import os |
|
import librosa |
|
import numpy as np |
|
import soundfile as sf |
|
from tqdm import tqdm |
|
from concurrent.futures import ProcessPoolExecutor |
|
|
|
|
|
audio_dir = 'audio' |
|
output_dir = 'processed_audio' |
|
|
|
|
|
os.makedirs(output_dir, exist_ok=True) |
|
|
|
|
|
def load_audio(file_path): |
|
audio, sr = librosa.load(file_path, sr=None) |
|
return audio, sr |
|
|
|
|
|
def normalize_audio(audio): |
|
return audio / np.max(np.abs(audio)) |
|
|
|
|
|
def save_audio(output_path, audio, sr): |
|
|
|
sf.write(output_path.replace('.wav', '.mp3'), audio, sr, format='MP3') |
|
|
|
|
|
def process_audio(file_path): |
|
audio, sr = load_audio(file_path) |
|
audio = normalize_audio(audio) |
|
output_path = os.path.join(output_dir, os.path.basename(file_path).replace('.mp3', '.wav')) |
|
save_audio(output_path, audio, sr) |
|
return os.path.basename(file_path) |
|
|
|
if __name__ == '__main__': |
|
|
|
audio_files = [os.path.join(audio_dir, filename) for filename in os.listdir(audio_dir) if filename.endswith('.mp3')] |
|
|
|
|
|
with ProcessPoolExecutor() as executor: |
|
for result in tqdm(executor.map(process_audio, audio_files), total=len(audio_files), desc="Processing files", unit="file"): |
|
print(f'Processed {result}') |