Thoughts and Theory

Automatic Speech Recognition: Breaking Down Components of Speech

Understanding the basic units of speech to train acoustic models

Kat He
Towards Data Science
10 min readMar 17, 2021

--

Table of Contents

  1. Overview
  2. Difficulties in ASR: noise, elocution, ambiguities in language, lack of context

--

--

Software engineer who enjoys learning about anything new, whether that be about machine learning, swe, or miscellaneous facts about the universe.