1 post • Page 1 of 1
Building a machine or computer program that understands human speech was a dream for a long time. Decades ago, scientists and engineers clung to phonetics research to tackle this problem, thinking that a very good understanding of the acoustic properties of speech and how they are mapped into perception would direct us to the solution. Fast forward--despite the time, manpower and dollars invested, such efforts bore no fruit. Instead, approaches largely ignoring phonetics and human perception lead to the creation of our new friends, Siri, Alexa, or Cortana, whom many of us cannot live without. We phoneticians lost. Miserably.