January 17th, 2022
Last October I wrote an article for Publishers Weekly, “AI Comes to Audiobooks.” With the growth of audiobooks continuing at scale, it’s not surprising that some enterprising vendors would start to explore where AI might enable the audiobook production process.
It comes down to using synthetic voices to supply the narration. Synthetic voices have come a long way since Siri, and they’re getting close, oh so close, to sounding human. They do a great job on short-form text. For full-length books, it’s possibly not quite there yet. You can never underestimate just how much cadence, intonation, inflection and emotion goes into the spoken voice — we tend to take it for granted.
As a result, nonfiction looks like the sweet spot for this technology, though, as you’ll see in the article, more than one vendor swears it can handle fiction as well.
None of the vendors claim that they’re offering a replacement for the top bestsellers. But what about the backlist? There are so many backlist titles that just don’t justify the cost of a full-blown audiobook production effort (plus many new titles with modest sales expectations.) If, instead, a “good enough” audiobook can be created for, let’s say, $1,000 total, how many more audiobooks would start to make economic sense? Answer: many.
Also included at the end of the PW article are links to a number of samples so that you can judge for yourself.
Two more recommendations:
- I was interviewed for this week’s Velocity of Content podcast with Christopher Kenneally. You can find it here.
- Joanna Penn has been following this topic closely, and offers a podcast with voice samples here. Joanna points out that another big opportunity for this technology is foreign language editions.
Subsequently:
Wired has just published an article, Synthetic Voices Want to Take Over Audiobooks, not dissimilar to mine, interviewing some users, and with more on the challenge of getting these books onto the Audible platform, still a major hurdle.
However, the article also links to Google’s new “Auto-narrated audiobooks:
Creating a high-quality audiobook is simple and affordable with Google Play Books.” You can choose from over 20 narrators, from a variety of gender, age, and accent combinations.
During the beta testing period, there’s no program fee to generate and publish an auto-narrated audiobook, and the audiobook can be downloaded and distributed on other platforms. Publishers will receive a 52% revenue share. Included is “advanced editing technology: Fine-tune the narration with the audio file editor. Choose from multiple word pronunciations or suggest your own.” Google says that auto-narration “works best on nonfiction titles.”
This is a potential game-changer for the startups in this space.
January 31, 2022: Apparently publisher Richard Charkin has been part of beta program for Google’s service since early last year, as reported in this Publishing Perspectives article. It includes links to book-length samples. (Thanks to Joanna Penn for the heads-up.)