The Honking Goose

My vision is 20/800. I still love reading, but paper books with their notoriously small font are logistically out of reach, so falling in love with audiobooks became part of my vision loss journey. Today virtually every eBook platform is compatible with screen reader technology. That means you can listen to the book being read in the screen reader’s voice. There are plenty to choose from, some more robotic, some surprisingly human, and which one you use is entirely personal preference. Or you can output the book to a digital braille display, a tablet-like device that connects to your smartphone or laptop and renders braille instead of audio through your headphones. With so many options, one might ask: why would a blind person want a human-narrated audiobook at all?

Imagine listening to the driest university professor you’ve ever had, all day long, paying close attention to every word and trying to make sense of it by evening. That is the reality of being dependent on a screen reader. I’ve talked about it a bit in therapy, admitting that I talk to Siri Voice 1 more than I do to my mom, whom I live with. Checking out an audiobook from a library means hearing a story told by a new voice, even if only for a few hours. It humanizes something that can otherwise feel impersonal.

Between the printing press, eBooks, and self-publishing platforms, hundreds of millions of books are published every year. Even before AI and large language models made writing faster, that implies tens of millions of unique authors, and that’s a modest estimate. Now imagine if there were only one writer in the world. Everyone would eventually grow tired of reading books written by the same person. If publishers stop producing human-narrated audiobooks, that is exactly what listening becomes. Human narrators add the same variety to audiobooks that thousands of authors add to the printed page.

Last week my mom and I were driving home from a medical appointment, and she turned on an audiobook on Audible. The book was most likely human-narrated, but the narrator sounded like the Bookshare Reader had been handed a new voice and told to keep the same delivery. The tone and the pauses between different kinds of punctuation were flawlessly identical. The AI voices eBook platforms keep adding sound closer and closer to human narration every year, and this narrator sounded like a drill sergeant had spent months drilling her on every line. Or maybe she just sounded that way, and listening to so much TTS has changed how I hear human voices. I was so focused on how she sounded that I remember almost nothing about the book itself. Either way, it gave me a clear picture of what a world without human narrators could feel like, and I think that is what bothers me the most.

Accessibility is more than mere utility. Tools need to solve the problem at hand, but they also need to account for side effects, and the difference between a usable tool and a good one usually lives in those details. Take a small example: if you use a screen reader and are also listening to a podcast or music on Spotify, you cannot listen to both at full volume at the same time. The solution borrows from music itself, a technique called audio ducking. It’s the same trick a DJ uses when they speak to the crowd and the music drops underneath them. When the screen reader starts speaking, everything else quiets down until it’s finished. Audio ducking is one of countless small problems screen readers have had to solve to fit into the rest of the world, and the care that went into solving it is exactly the kind of care human-narrated audiobooks represent.

Text-to-speech is still extraordinarily useful. It’s the primary tool I use to interact with technology, and for books without an audiobook (virtually every college textbook, for example) it can be the difference between reading on my own and needing someone to read to me. It should not be seen as the only solution, or as a reason to stop producing human-narrated audiobooks.

Small, meaningful actions add up. Next time you’re buying an eBook, consider the audiobook instead. If you have a library card, use it (and if you don’t, please get one, they’re free). Many libraries partner with apps like Libby that let you borrow eBooks and audiobooks right to your phone or tablet. Libraries track those checkouts and use them to decide what to license next. Publishers track audiobook revenue and use it to decide what to commission. The market follows the demand.

We as consumers decide the future of audiobook production. Will you settle for speech synthesis, or stick with human-made audio?