You know, I didn't believe them when Draft2Digital said that it could take up to two months to produce an audiobook for Apple, since Google could deliver an editable version of one in two hours. But I was wrong. It looks like it can, and will, take two months, or more, to spit out the audiobook versions of all my novels. I signed on to Apple Audio books via Draft2Digital on the first of January 2024. The Secret of the Tzaritsa Moon appeared on the 18th of January, two more appeared on the 8th of February, two more 16th of February, and here we are, on the 21st of February and they are still not all out. So, yeah, they weren't just under promising.
Beggars can't be choosers, so I'm not exactly complaining. I'm just noting the glaring differences between the two companies, in terms of products delivered.
There are many other differences as well. Let's start with Google.
First, you have to upload your own homemade ePub file to Google. And once they have your ebook, you can then select the audiobook option. Within a couple of hours you have the text, or more correctly, the audiobook script, available to you prior to release which you can edit before releasing it.
Google offers about a dozen different English narrators in both male and female voices to choose from; two each with different English accents, including American, British, Australian, and Indian. You can also alter the narrators speed to further customize the voices.
You can review the text, and highlight words to hear how the AI pronounces them, which is very handy in the case of invented words; names, places, etc., and you can alter the spelling in the audiobook script to get these words to sound the way that you envisioned them. You can also insert pauses where you think they're needed, but not inserted automatically.
You can also choose to assign different voices to different characters, though you would have to go through the script highlighting each instance of that character's dialog, perhaps a hour or two's worth of work. If the character's voice is distinctive enough, you might also be able to eliminate dialog tags ("he/she said.") as you go. You get distinctive voices rather than one narrator "doing voices," which honestly, I find sort of annoying, though I've not heard the popular voice actors doing it, so it might be better than the small sample I've listened to. I haven't taken advantage of this feature because I picture the story being told by a single narrator. However, from some of the samples of my stories I've listened to, I can see the advantage of this feature, since without the visual clues of the text, it can be unclear who's doing the talking. But if you do one character, would you need to do all the major ones as well? A slippery slope...
When you're happy with how things sound, you just select a price, and hit "publish." It will take a few hours for it to be live. You can go back and edit it, whenever you like.
Turning to Apple, you do Apple's auto-generated audiobooks through Draft2Digital's page for the book you want to offer as an audiobook. Click on the audiobook option, and then on the Apple option.
After that, the audiobook is produced from the ebook version published by D2D. All you have to do is select the genre you want to offer the book in and its price. Apple offers only two male and two female voice, and they select which one will be used based on your book's genre. Click publish and wait. You can not modify the audiobook text for free after that, and have to keep it up for at least six months. It's very simple, but you have very limited options.
Both services require a square cover; Google a 1Kx1K pixel one, Apple a2Kx2K one, though D2D will create one for you from your ebook cover. I used my own.
So what do they sound like? The first link is to the sample page for The Mysteries of Valsummer House on Apple Books. Hit Listen. The second is that book's audiobook page on Google, once again just hit play, though it starts at the beginning instead of a minute or two into the text like the Apple sample.
I'm not sure what voice Apple used.
The Google voice is "Archie," a British English voice.
Whose AI do you think does a better job at creating a human narrator? And if you're a regular audiobook reader, how far from human narration are these books?
No comments:
Post a Comment