This script maps an audio file, speech f.ex., (mp3 until now only) with the respective transcript. After mapping it in a json file, it splits up the original file into small mp3 files for each sentence.
These files can then be used for annotation and for training of a neural network for example.