MIT_preparation.py

#!/usr/bin/env python3

# July 2018

# Description: 
# This code prepares MIT for the following speaker identification experiments.
# It normalizes the amplitude of each sentence.
 
# How to run it:
# python MIT_preparation.py $MIT_FOLDER $OUTPUT_FOLDER data_lists/MIT_all.scp

import shutil
import os
import soundfile as sf
import numpy as np
import sys
from os import listdir, makedirs, remove, rmdir
from os.path import isfile, exists

def ReadList(list_file):
 f=open(list_file,"r")
 lines=f.readlines()
 list_sig=[]
 for x in lines:
    list_sig.append(x.rstrip())
 f.close()
 return list_sig

def copy_folder(in_folder,out_folder):
 if not(os.path.isdir(out_folder)):
  shutil.copytree(in_folder, out_folder, ignore=ig_f)

def ig_f(dir, files):
 return [f for f in files if os.path.isfile(os.path.join(dir, f))]

def maybe_make_directory(dir_path):
  if not exists(dir_path):
   makedirs(dir_path)

in_folder=sys.argv[1]
out_folder=sys.argv[2]
list_file=sys.argv[3]

# Read List file
list_sig=ReadList(list_file)

# Replicate input folder structure to output folder
copy_folder(in_folder,out_folder)


# Speech Data Reverberation Loop
for i in range(len(list_sig)): 
 
 # Open the wav file
 wav_file=in_folder+'/'+list_sig[i]
 wav_file = wav_file

 [signal, fs] = sf.read(wav_file)
 signal=signal.astype(np.float64)

 # Signal normalization
 signal=signal/np.abs(np.max(signal))

 # Save normalized speech
 file_out=out_folder+'/'+list_sig[i]

 file_out = file_out.lower()
 final_folder = file_out.split('/')
 final_folder = '/'.join(final_folder[:len(final_folder)-1]) + '/'
 maybe_make_directory(final_folder)
 sf.write(file_out, signal, fs)
 
 print("Done %s" % (file_out))