foldertomatlab.py

from pydub import AudioSegment
import numpy as np
from pathlib import Path
from scipy.io import savemat
from scipy.signal import medfilt

duration = 10 # 10 seconds opening of any audio files

def folder_to_matlab(input_folder, output_folder=None):
    try:
        input_folder = Path(input_folder)
        if not input_folder.exists() or not input_folder.is_dir():
            raise FileNotFoundError(f"Input folder not found: {input_folder}")
        
        if output_folder is None:
            output_folder = input_folder / "output"
        output_folder = Path(output_folder)
        output_folder.mkdir(parents=True, exist_ok=True)

        files = sorted(input_folder.glob("*.mp3")) + sorted(input_folder.glob("*.m4a")) # can adjust file 
        if len(files) == 0:
            raise ValueError(f"No MP3 files found in {input_folder}")

        target_rate = 44100 #CHANGED FREQUECNY SO IT IS CNSOSTNET

        for idx, file in enumerate(files, start=1):
            try:
                audio = AudioSegment.from_file(file, format="mp3")
                wav_data = np.array(audio.get_array_of_samples())
                # Strip leading zeros from wav_data
                # wav_data = np.trim_zeros(wav_data, 'f')
                wav_rate = audio.frame_rate

                # Determine the number of samples for 10 seconds
                max_samples = wav_rate * duration  # 10 seconds
                if len(wav_data) < max_samples:
                    print(f"Skipping {file.name}: Audio shorter than 10 seconds.")
                    continue
                
                # slice for the first 10 seconds
                wav_data = wav_data[:max_samples]

                # Apply chunk-based median filter
                chunk_size = 45  # 0.1 second chunks for 44100 Hz sample rate
                filtered_data = []
                for i in range(0, len(wav_data), chunk_size):
                    chunk = wav_data[i:i + chunk_size]
                    median_value = np.median(chunk)
                    filtered_data.extend([median_value] * len(chunk))
                wav_data = np.array(filtered_data)
                wav_data = wav_data[::chunk_size]
                time_axis = np.linspace(0, duration, num=max_samples)  # time axis 

                # save to MATLAB .mat file
                output_filename = f"{file.stem}.mat"
                output_path = output_folder / output_filename
                data = {
                    "time": time_axis,
                    "amplitude": wav_data
                }
                savemat(output_path, data)
                print(f"Processed {file.name} -> {output_filename}")
            
            except Exception as e:
                print(f"Error processing {file.name}: {e}")
    
    except FileNotFoundError as fnf_error:
        print(f"File not found: {fnf_error}")
    except ValueError as value_error:
        print(f"Error: {value_error}")
    except Exception as e:
        print(f"Error: {e}")

# Example usage
input_folder = "./"  # replace with your relative folder path
output_folder = "./"  # current directory for output
folder_to_matlab(input_folder, output_folder)