computer-generated audiobooks?

A reader emailed to ask if I’m planning to use computer-generated voices to turn any of my books into audiobooks. (I think the question was inspired by this article from Wired.)

No, I’m not, for two reasons.

1.) The technology isn’t as good as its proponents think that it is. It just isn’t. It can do a very excellent facsimile of a human voice, but it’s not quite there. It’s a little too controlled, too modulated, to be recognizably human. Like a really good interactive phone menu. Ironically, the technical perfection of the computer-generated voice may cause the brain to realize that it’s fake, since nothing human is truly perfect. Additionally, computer-generated voices can’t do emotion very well, so reaching a scene that’s a passionate declaration of love might be a bit strange when the synthetic voice reads it in a calm monotone that would also be suitable to Alexa giving the weather forecast.

Also, I’m a fantasy author, and getting the correct pronunciation of names like Agrimnalazur and Delaxsicoria through any sort of automated process would be a huge pain.

2.) A more practical reason – Amazon and Findaway Voices both ban the sale of computer-read audiobooks, and when computer-read audiobooks slip through Quality Control, they’re deleted as soon as someone reports them. Since Amazon’s ACX and Findaway Voices are the two main ways indie authors can get audiobooks into the stores, that kind of makes computer-generated audiobooks a non-starter.

And any computer-generated audiobooks that elude Quality Control will get reported, because a very large subset of listeners hate computer-read audiobooks with the burning nuclear heat of the fiery cores of a thousand suns. I know some authors have produced computer-read audiobooks and are selling them through their own stores (powered by Shopify and Payhip and the like), but I’m curious what the reaction is. I’m willing to wager that it’s more on the negative side.

A good example: my absolute favorite science fiction novel of all time is Timothy Zahn’s THE ICARUS HUNT from 2000. It’s never had an audiobook available in the modern audiobook era (there was one on cassette that disappeared into unavailability), but while writing this post I was delighted to discover that it’s getting an audiobook narrated by Marc Thompson (of STAR WARS audiobook fame) that comes out on February 1st. I will buy it, immediately, the instant it is available, because it’s my favorite book and Marc Thompson is an excellent narrator.

But if it was narrated by a computer-generated voice, I wouldn’t be mad about it…but I wouldn’t buy it, either.

So, two reasons why I won’t use synthetic voices for my audiobooks.

That said, I can definitely see a use for the technology in the domain of accessibility. I think the ideal use for it is to get integrated into the ereader apps for visually-impaired readers. A while back I had an angry email from a reader claiming that I was biased against the visually impaired because I disabled Text-To-Speech in my books. But I didn’t! It’s totally enabled in every place that has it. If you have a Kindle Fire, it will read any Kindle book aloud to you. (Unfortunately, that feature is not available in the Kindle Android and iOS apps.) Google Play Books includes a pretty good Text-To-Speech feature, and I think Apple Books has one on iOS and iPadOS that’s integrated with Siri.

The unfortunate fact is that audiobooks are expensive and difficult to produce, and the majority of book titles in every language will probably not have audiobooks attached to them. So I definitely think that it’s a good idea to integrate computer-generated voices into ebook readers. However, I simply don’t see the sale of computer-generated audiobooks becoming common, unless the technology gets a lot better at expressing emotion, which seems unlikely, or if audiobook listeners get less picky, which is about as likely as those thousand burning suns spontaneously turning to ice. 🙂

-JM