From Publishers Weekly:
In a post today in the Kindle Direct Publishing community forum, the self-publishing giant announced that it has begun a beta test on technology allowing KDP authors to produce audiobook versions of their e-books using virtual voice narration. The ability to create an audiobook using synthetic speech technology is likely to result in a boom in the number of audiobooks produced by KDP authors. According to an Amazon spokesperson, currently only 4% of titles self-published through KDP have an audiobook available.
Under the new initiative, authors can choose one of their eligible e-books already on the KDP platform, then sample voices, preview the work, and customize the audiobook. After publication, audiobooks will be live within 72 hours, and will distributed wherever Audible titles are sold. Prices can be set between $3.99 and $14.99 and authors will receive a 40% royalty. All audiobooks created by virtual voice, the post says, will be clearly labeled and, as with any audiobook, customers can listen to samples.
“We are excited to introduce a new option for customers and authors,” said Amazon spokesperson Lindsay Hamilton. “Virtual voice gives authors more choices to create audiobooks and will deliver greater selection to customers.”
The new virtual voice option complements Audible’s existing Audiobook Creation Exchange (ACX), which matches audio rightsholders (authors, agents, publishers) with audio producers (narrators and studio professionals). KDP said that it plans to grow the virtual voice beta test over time, and will share updates in the coming months.
Link to the rest at Publishers Weekly
PG says it was only a matter of time.
Lots of very nice people earn a living or supplement their income from other sources by working as readers of audiobooks. As with all human endeavors, some do a better job than others. Regular audiobook listeners will sometimes choose an audiobook because they have heard the voice artist before and like his/her voice and style.
That said, the voice artist is a means to an end – turning a written story/account into a spoken version of the original.
PG hasn’t listened to any of the KDP books created using virtual voice narration, but he’s likely to try one out fairly soon.
Assuming that the quality is reasonable, there are some benefits the indie author can reap from audiobooks.
There is no delay between the roll-out of the printed book and ebook and the release of the audiobook. Authors need not delay KDP print and digital to wait for the creation of the audiobook if they wish to announce audio and print/electronic at the same time.
To be certain that a human reading a book aloud has not made any embarrassing blunders in pronunciations of non-standard or foreign words, an audio proofreader may be hired to check the audiobook. Presumably, at least after working through early-product glitches, the virtual voices will produce highly-predictable results.
Finally, of course, there is the cost of an audiobook with a human narrator vs. the cost of an audiobook produced with digital narration. Absent some sort of promotional pricing, human-performed audiobooks typically cost more than the paperback versions and much more than the ebook versions.
As mentioned in previous posts, PG’s exposure to audiobooks is quite limited because, boorish Neanderthal that he is, he consumes ebooks at a much faster pace than he is able to injest any other format.
So, he would be interested in the opinions of serious ebook readers about virtual narrators vs. narrators who inhale from time to time.
Jim Frangione’s Chet and Bernie audio books are a performance. He brings emotions to his voice and uses multiple character voices.
That is a high bar for AI.
I do not know if this is true but it seems to me that a book written for reading would need revising to fit the spoken word. Perhaps AI can be of use here.
I’m curious to see how/if it handles non-english/made-up names and words in SF&F.
Will it allow an author exception guide for such?
FYI, Kristine Kathryn Rusch tested a bunch of tools a month or two ago. Enough so that she could offer an audio version of her newsletter that isn’t horrible, but it is far from a natural rhythm voice. Eventually, it will get there, but based on her informed triage, it doesn’t seem like it is there yet. I’ll be interested to see people’s experiences with it…Amazon doesn’t blunder on releases of features that often…if they think it’s ready…
I confess my interest is a bit sideways. I would love the ability to create different voices for videos…think of a comic strip, for example, where each character would have a distinctly different sounding voice, but NOT requiring me to do the voices or it sounding too computer-y. Narrated audio with some variability would get you close to that for dialogue.
I don’t see a price to make the audio book. If this is a freebie, and the voice is half decent, what’s the downside?
Agreed on all, PG. One point of clarification. AI – Narration companies are inflating what they say are costs for human narration so that their price falls well below that. I am an audiobook consultant for authors who don’t want to learn all the ins and outs. I find quality narrators all the time that cost the same as AI. If an author pays a listener/proofreader, then the AI version could cost more. That said, it was only a matter of time before the behemoth that is Amazon made this service free, presumably.
Many non-fiction books will sound great with AI, once pronunciations are fixed. However, the art of using different voices for fiction will be missed.
Thanks for sharing your inside expertise, Darren.