TL; DR: Amazon just expanded the pipeline dumping bot-voice audiobooks into Audible.
Let me start by saying I have ethical objections to bot-voice audiobooks. A) Labor solidarity with actors, musicians, visual artists, and everybody else ripped off and told they’re obsolete because now Johnny NoSkills can push a button and “do” their job. B) Where do people get off charging customers for a shitty product that was no cost to generate? Readers can get literally the same thing, for free, by having their device’s text-to-speech read the ebook to them and they can choose a voice they like. Putting a price tag on an inferior version is a straight ripoff.
So while there’s no chance I’ll release bot-voiced audiobooks for public consumption, when a distributor throws a “Try our new bot-voice audiobook generator!” banner across the top of my dashboard, I’m compelled to test it, if only so I can call bullshit on its users who claim it sounds great, really and for truly.
Years ago, Google offered such a service. Their promo page extolling the virtues said it was best for nonfiction… and then the sample audio they offered was Dracula. You know, that dull little tract famously devoid of atmosphere and angst. Inside the actual conversion process, the voice samples recited a snippet of The Wizard of Oz. (And absolutely butchered “Auntie Em.” Good thing that wouldn’t be a fishhook in the ear a million times, am I right?) So while they made the minimum effort to warn that the robot could most successfully drone dull, dry information, all the examples they chose encouraged feeding fiction into it.
Personally, I couldn’t get through the first page of one of my own stories lurching through the headphones like a Silent Hill nurse. Better no audio than that garbage.
But that was years ago! So when Amazon recently threw their “Try our new bot-voice audiobook generator!” banner across the top of my dashboard, I was excited to hear how much the technology has advanced.
Or not. These bazillion-dollar tech companies have had years to improve this technology they’re so eager to thrust in everyone’s face, yet there has been zero change. Whether they’re uninterested in making it sound decent or this is the limit of what it will ever be able to achieve (because, for the millionth time, machines can’t do context), what they’ve released is unfit for purpose.
While Amazon’s fidget offers an option to input phonetic pronunciation corrections, it doesn’t fix the frequent weird emphasis. (Which varies by voice. Some handle Cheetos in a normal fashion, while others bray Chee-TOES!!) You can add pauses, but half the time there’s an artifact on the end of the word preceding the pause (“now-th pause the”). Why? This isn’t a recording in which a human mouth ran sounds together. It’s pulling the sounds of “now” and “the” from a table. There is no reason you shouldn’t be able to fully separate those generated sounds.
You can assign different voices per chapter, but you can’t assign different voices to speakers within chapters. There is no distinction between the inflection-free narrative and the inflection-free dialogue, so unless the book was written with “Kim said,” “Mervin shouted,” “Enid whispered snidely,” and so on before every single line of speech, you can’t tell someone is speaking until you process that these toneless words are a contextual shift from the previous toneless words. That means constantly slamming on the mental brakes, backing up, and reprocessing, taking on a lot of cognitive load that’s unnecessary when a human being makes even minimal effort to create verbal distinctions for the listener.
“Fuck you, Officer Beaver” delivered by Humorless Microwave #3 doesn’t land right. Now imagine a fight scene, a sex scene, a soul-baring scene, any moment designed to pluck at readers’ feelings, plunked down one flat word after another, blah blah blah, compelling as a toaster oven user manual. It’s the kind of noise you tune out as an irritant, for which you have the privilege of paying a minimum list price of $3.99.
Even if the ethical issues magically disappeared, I’d be ashamed to be associated with a product of such appalling quality.
Fortunately for Amazon, which keeps 60% of the sale price through Amazon and Audible, there’s no shortage of people eager to jump on a quick, effortless buck at customers’ expense. Sucks to be that customer forced to pick through the landfill in search of suppliers committed to quality, though.