Bad Apple!! but its rendered as short time fourier transform

Each frame is converted into audio in such way that when showed on spectrogram its a frame back again. Uses mainly numpy, librosa and opencv with Python (X axes are wrong but whatever) Github: original video: Enjoy!