Convert MP4 to Spectogram

convert_spectogram.md

Convert files in MP4 folder to FLAC

Scipt to convert

import os
from glob import glob
import subprocess

MP4_PATH = '~/Desktop/notebooks/mp4/*.m4a' # path to all mp4 files
OUTPUT_PATH = '~/Desktop/notebooks/flac/' # any path to save as flac file
OUTPUT_SPECTOGRAM_PATH = '~/Desktop/notebooks/spectogram/'

mp4_paths = glob(MP4_PATH)
if not os.path.isdir(OUTPUT_PATH):
    os.mkdir(OUTPUT_PATH)
if not os.path.isdir(OUTPUT_SPECTOGRAM_PATH):
    os.mkdir(OUTPUT_SPECTOGRAM_PATH)

for mp4_path in mp4_paths:
    file_name = os.path.splitext(os.path.basename(mp4_path))[0]
    file_name_flac = file_name + '.flac'
    file_name_png = file_name + '.png'
    subprocess.call(['ffmpeg', '-i', mp4_path, '-acodec', 
                     'flac', '-ab', '128k', '-ar', '44100', 
                     '-ac', '1', '-f', 'flac', '-t', '60', 
                     os.path.join(OUTPUT_PATH, file_name_flac)])
    subprocess.call(['sox', os.path.join(OUTPUT_PATH, file_name_flac), 
                     '-n', 'spectrogram', 
                     '-Y', '200', '-X', '50', '-m', 
                     '-r', '-o', os.path.join(OUTPUT_SPECTOGRAM_PATH, file_name_png)])

See explanation on -X and -Y here. Basically, 50 means 50 pixels/second (20 ms bin size) and 200 means it will select 128 as frequency bin since it's the highest 2^n that before 200.