instagram pinterest linkedin facebook twitter goodreads

A Psychologist's Thoughts on Clinical Practice, Behavior, and Life

Using Artificial Intelligence (AI) to Create Text-to-Speech (TTS)

Many great books have never been created as an audio edition so what I've been playing with is listening to them using the TTS of Amazon and Google with an old Kindle E-book (version 2) and a modern Android (8.1) smartphone. The speech using this AI is obviously robotic and not terribly enjoyable but I've listened while doing chores. Human-read books are vastly better. Amazon has a Polly AI which is commercial and cheap and I listened to a sample. It is much improved over the others but still not perfect. There's a human emotion fluctuation after a comma and a period which it doesn't get quite right. I found this interesting and wondered if, sooner than later, AI produced audio books will put human readers out of business since it is so much cheaper. Or if it is already being used by publishers. There are already very human-like news readers in China with an increasing ability for news stories to be AI created. All-in-all, it's a fascinating development. But I don't believe that AI will ever push serious writers out of existence. For them and most people, creativity is what counts.

Be the first to comment