Text this: Time Domain Representation of Speech Sounds