X

Auto-Narrated Audiobooks With Ryan Dingler From Google Play Books

    Categories: Futurist

What is auto-narration of audiobooks and how can it benefit authors and rights-holders as well as listeners? What are some of the common objections to auto-narration and how can we keep a positive attitude to embracing change? Ryan Dingler from Google Play Books goes into detail on these questions and more.

You can also listen to my recent round-up of AI narration options across multiple platforms in episode 623.

Ryan Dingler is a product manager at Google and also writes about the intersection of technology and business.

You can listen above or on your favorite podcast app or read the notes and links below. Here are the highlights and the full transcript is below. 

Show Notes

  • What is auto-narration? Who can use it and what languages and accents are available?
  • Tackling common objections to auto-narration — “the voices aren't good enough;” and narration should be “kept for humans”
  • Stratification of audio rights and how both human and AI narration fit into the audio ecosystem
  • Multi-cast audiobooks
  • The potential for growth in non-English-speaking markets
  • Pricing an AI narrated audiobook
  • How will the market change in the coming year?

You can find Google Play Books at play.google.com/store/books and you can publish at Play.google.com/books/publish/u/0/

g.co/play/autonarrated – Publisher website for auto-narrated audiobooks

g.co/play/narrator-library – Narrator library

g.co/play/autonarrated-help – Help center

Transcript of Interview with Ryan Dingler

Joanna: Ryan Dingler is a product manager at Google and also writes about the intersection of technology and business. Welcome back to the show, Ryan.

Ryan: Thanks for having me on.

Joanna: It's good to talk to you again. You were on the show in April 2021, talking about publishing on Google Play Books in general. And we just mentioned auto-narration for audio, which was in beta at the time, but we're going to go into that in detail today.

What is auto-narration for audiobooks? Give us an update on where the program is now.

Ryan: We've come a long way since I last spoke with you both in terms of the beta and in terms of what we've done with the product.

But just to start off, what are auto-narrated audiobooks, it's very simply, instead of being read by a person, auto-narrated audiobooks are read using Google's text-to-speech technology.

We have a whole tool set and framework around building these auto-narrated audiobooks. And this all came about a few years ago, actually, where we noticed just a massive gap between eBooks and audiobooks.

95% of eBooks do not have an accompanying audiobook, which in our book catalog is millions and millions of books.

And it's not that we looked into it, some of these eBooks would make sense as an audiobook, it's just that, as most everyone knows, audiobooks are very expensive to create and take a lot of time.

So that's when we came up with the idea of auto-narrated audiobooks. We have been progressing it forward since then.

Joanna: And It was in beta. Who can get into it now? Has it come out of beta completely?

Ryan: Beta can be somewhat of a confusing word; it is still in beta. Google is known for keeping our products in beta for a long time.

Available in 8 countries and 2 languages and multiple accents and genders (as of mid-2022)

We are just beginning our process, but it is actually available generally in eight countries today. Those countries are the United States, the UK, Canada, Australia, New Zealand, Mexico, Spain, and Argentina.

If you are a publisher in one of those countries, and you have an account with us in our partner Centre you can go in and try it out. It's in two languages, English and Spanish, as you might be able to guess, by the countries.

And all you really need is three things, you need an eBook that's in English or Spanish, and it needs to be an ePub format. And the book does need to be live on Google Play.

We have, as I said, before, a whole editor around this process to help walk you through it and make sure that the auto-narrated narrator is actually pronouncing things the way that you want.

And that your eBook that is now going to become an audiobook has all the things that an audiobook would have and doesn't have, which of course, it's like the Table of Contents, copyright, and things like that.

Joanna: I've got a quite a few questions coming from that then. I have tried this. I have a couple of auto-narrated books.

You've got English and Spanish but so just using English because that's what I was doing. You have different accents as well, don't you? Because of course you and I both speak English, but we both have, well, different genders but also different accents. What are the range of accents available in those languages?

Ryan: We have, I think about six different accents for English language today. Off the top of my head, I think we have American, we have British, we have Indian, Australian, and a few other ones. We have really tried to provide publishers a full-range of accents.

Obviously, we have nowhere near covered the wide variety that there is. Obviously, there's even accents inside of the UK, inside of the American, that would be further.

But our goal is to continue adding these accents so that publishers can choose the right narrator for their book, whether they're trying to represent the author or represent a unique character that would be speaking it.

So we have seven or so today by language and we also have those accents in Spanish as well.

But we're continuing to add more and one word on languages, so we are in English and Spanish only today but we are looking to add German and French and that we're expecting to have those made available by later this year, and in Portuguese as well might be following just after that

Joanna: That's brilliant. I narrate my own as me, but I had a American woman, an African American woman that the voice sounded like and it's lovely. I love it.

Also what I like is that you can actually speed it up slightly. You can do like I think like a 1.1 if you want the speed to be a bit different. I'm right on that, aren't I?

You can change the kind of speed and that kind of thing?

Ryan: Yes, exactly. You can change the default narrator speed. I think we have it It can go from roughly 0.5, which is very slow, to 1.5 from the default narrator speed. Each narrator speaks at their own cadence.

We try to give a sense of this by telling them the speed, which is really just the amount of words per minute that they would speak normally. Everyone speaks at different paces. If you like a narrator, but they speak a little bit too slowly, you can speed them up or slow them down either way.

Joanna: I do like the editor, I think the editor process is brilliant. I think it's much better than a lot of other things that I've seen in the AI space. It's very user-friendly.

You mentioned that you can change the pronunciation of things. Explain how that happens.

Ryan: This came about as we were looking at our own public domain audiobooks. So public domain books are books that anyone in general society can use and create. So we started creating some of our own books, think of books like ‘Frankenstein', that have been on the market for 50 to 70, I think is the cutoff for public domain, 70 plus years or more.

What we found is that just some words, especially words that were from older English, because these books are quite old, were just pronounced in a way that we would have otherwise wanted different.

So we have two ways to do it. One is if it's just pronounced wrong, we try to fix that in our system. There's also just personal preference by the publisher or the author. I created this name. I want it spoken this specific way.

If that is the case, you can go into the editor, you can click on a word, and click a right click, hit Edit Pronunciation.

And this whole panel on the right-hand side opens up to either change the spelling of a pronunciation or try to tweak the way that it's pronounced.

We can do that through a few ways. One, if it's like a homograph it can be pronounced in a few different ways. For instance, I always think of tomato/tomato, or if you're British, the way that you pronounce water is very different from an American would pronounce water. So we would have those alternative pronunciations, especially if you do it with a British narrator versus an American narrator.

We've also found that sometimes it's just really hard to phonetically type a pronunciation. So we do have a feature that allows you to speak the pronunciation into your mic if you just only speak that specific word. We try to convert that into our own kind of pronunciation language and capture the meaning or the way that you are trying to pronounce the word. I always like to think of when Harry Potter came out, everyone was pronouncing Hermione as Hermon-e,

Joanna: Or Ho-mione.

Ryan: Or Her-mione. A ton of different variations. And if she wanted, J.K. Rowling could just go in there and say, ‘Oh, no, it's Hermione.' And then it would be fixed throughout the entire audiobook.

Joanna: This is kind of magic. I've been narrating audiobooks now for a few years, and also working with narrators. And there's this horrible moment if you're working with a human narrator and you realize you haven't told them how you wanted a name to be pronounced.

I have an example, I had a character called Gest, G-E-S-T. And they pronounced it ‘jest,' as if it had a J at the beginning. And I just couldn't see how that would have been…like a hard G rather than a soft G.

But with the AI, you could just type it once, right? And then it changes the whole audiobook rather than having to re-record anything, which is frankly, amazing.

And we were saying just before we started, we were trying to get the room tone right. Again, as a narrator, I'm like, ‘I have to edit out all these clicks and pops and noises of my jaw.‘ Sometimes if I'm dehydrated, I have to edit those out or lip smacks, or tummy gurgles, and all these things and they just don't happen.

So there are a lot of speeds, as you say. And that's humans are humans, they make human noises. But it is interesting how I just love that universal change. I think that that is a killer app, basically.

Ryan: One thing that I always like to point out too is if you have a mistake in a traditional audiobook and you realize it post-publishing, you're probably not going to fix it, it takes a lot of time. It depends on the severity of what you would see as the mistake.

With an auto-narrated audiobook, all you need to do is go back and change the one word or the same that was pronounced in a way that you would have wanted differently. And you can publish it again, download the files, and it takes a matter of minutes to fix those types of errors.

Joanna: But of course, there are some objections to auto-narration. So let's go into some of those.

Objection 1: The voices are not good enough. They sound like robots, or they don't sound like humans. They don't express enough emotion. They are just not good enough.

What are your thoughts on that one?

Ryan: I think first and foremost is we also believe that auto-narration cannot capture the complete nuances that human narrators can.

Some titles do require a lot of emotion and a very deep understanding of the books' contexts like emotionally charged dramas or romances.

So, if you as a publisher, an author can afford to invest in a full audiobook production with a professional narrator, we would highly encourage you to do so and of course encourage you to sell that audiobook on our platform.

But that being said, we do believe that our auto-narration is quite high quality.

Many people's experience with text-to-speech comes from the earliest versions of text to speech that they heard, which did sound very robotic and were quite hard to listen to for long periods of time.

But text-to-speech quality does vary significantly in terms of quality. It also has improved a lot over the intervening years since it was first introduced. Google has been working on text-to-speech for 15 plus years and seen lots of progress over that time.

Google has been particularly invested in text-to-speech because it's in so many of our products with Google Assistant across a wide range of products, You can speak and hear back from the device that you're speaking into. So Google has invested significantly, and we do think that the quality has improved drastically. And Google we do think stands out quite a bit.

I think the most important thing is if you're curious is just to listen to it yourself. You can always go to our partner center, we have a ton of samples, sort of different narrators, quality does vary by narrator, but we found it's really a matter of personal preference. Some people prefer one narrator, others prefer another. But just go listen to yourself if you're curious about it.

[Here's the narrator library where you can listen to samples.]

Joanna: Absolutely. I just had a follow-up question on the voices.

Can you use multiple voices within one audiobook?

Ryan: That is a good question. Right now it's one. You have a default narrator.

But we've been working for a while on adding multiple narrators to an audiobook. It's definitely one of the biggest points of feedback we've had from publishers, if you have either, let's say like a romance book with two points of view, a male and female, you want to be able to have both, one by chapter, you could do that.

Once we do this functionality, it actually will go down to the word level. You could have a conversation back and forth between two characters. And it doesn't stop at two, you'll have the ability to have as many characters as we have narrators.

I think we have roughly 15 to 20, English narrators. You could have 15 to 20 characters in your book, and you just go through and tag it.

And this is something that you could do kind of as I was mentioning before; if you already have it published, you can just go back and say, ‘I want to add characters to improve the quality and improve understanding,' you'll have the ability to do that. And we expect it to come out sometime in the next few months.

Joanna: Okay, great. Hopefully within 2022. That multi-cast would be awesome, because I feel like that's also one of the most expensive and complicated audio projects to do as independents who don't have their own studios. If you have to not just hire all the different actors, but also then edit that together, that is a big project. I could see that as a sort of creative.

That's something I would love to do. But I definitely could not afford to do that. I would love to play with that within the tool where it's easy enough to change things. So that will be really interesting when that comes out.

Let's talk about the one that comes up a lot as well.

Objection 2: It is ethically wrong to use AI for narration. We need to keep jobs for narrators.

And this has gone so far as in the UK we've got Equity with a campaign to stop AI stealing narrators' jobs basically.

What do you think about that one?

Ryan: Auto-narration is about affordably creating audiobooks.

As I said before, we don't think that auto-narration can really match the nuances that the human voice can. The human voice is still by far and away the superior storyteller, but because of significant costs and time to create a traditional audiobook, that's why we have seen this large gap between eBooks and audiobooks.

That gap would continue into the foreseeable future. And for many publishers, the choice really isn't between human narration and auto-narration, it's between auto-narration and no audiobook at all.

Publishers can try out for their first time audiobooks through auto-narration, and they can really use it as a tool to assess audiobook demand for their titles before making an investment in human narration.

We see them as going hand-in-hand. Both are trying to increase the audiobook catalog, grow the demand for audiobooks in general. And it's really just a testing ground for what you want to do further in audiobooks. It's kind of a first step.

Joanna: I'm with you. I don't think the two are the same thing.

And in fact, I've been advocating for a change in publishing contracts to allow for stratification of audio rights.

At the moment, people just sign away audio rights to publishers, whereas I think there should be more of a stratification of audio rights because many authors will sign away those rights and then it will never happen.

It's almost like there needs to be a difference between a human-narrated project and an AI-narrated project. Also, I like having both.

For example, my short story trilogy A Thousand Fiendish Angels there's me narrating it, so a British female, and then I actually have a male AI narrating it as well.

Some people like listening to a different voice, right? Some people would rather have a different voice. I actually think having multiple editions with multiple different voices is interesting as well.

Is that possible? Or is it just a one-on-one link within Google?

Ryan: It is a one-on-one link. But it's only that if the publisher chooses it. So when they see a book, that book will only have one narrator, or as we talked about, it could have many narrators as part of it, but it'll only have one default narrator.

But a publisher can choose to do many narrators for an auto-narrated audiobook.

If you wanted to do one that's different by geography, for instance, you wanted to have an American female for the U.S. market and a British male for the UK market, you can do that. It's really up to you as a publisher.

Right now, we don't have the ability for consumers to choose which narrator they want to speak the audiobook. So they'd actually have to be different books. In time, we probably will get to allowing for more consumer choice that way, if you're in the U.S. and you want a different narrator, and you can choose from a handful ones, you'd have the ability to do that.

Right now we're constrained a little bit by running these in kind of, not in real-time, so we can get the highest quality. But as text-to-speech quality improves, it'll really come down to we can do these live, and they can choose which narrative they want.

Joanna: That's fantastic. I do see, as you said, we're still in early days. What's funny is I've been talking about it for years, you've been talking about it for years. But it feels like it's actually starting to happen now.

I guess you mentioned their demand, and what have you seen in terms of the adoption, both by creators, but also by listeners.

Are downloads and sales increasing of these auto-narrated books? Or are they just sitting there?

Ryan: We've seen significant downloads. It all starts with supply, though. And we've been pretty impressed with the level of adoption that we've seen and how quickly it has happened since we rolled out the beta.

Now on our platform, thousands of publishers have not only created but are publishing an auto-narrated audiobook.

And for a lot of small publishers that we hear from particularly a lot of self-publishers, being able to produce an audiobook has always felt completely out of reach, and having access to this technology is allowing them to publish their audiobook for the first time.

In some ways, we've heard publishers share stories about connecting with readers for the very first time who've never been able to access their content because they only listen to audio format, whether for personal preference, or because they're in the blind and low vision community or they have challenges with reading. So we've seen a lot of publishers try it out, continue to grow adoption.

We have been quite impressed on the consumption side as well. People are buying these books, listening to them, engaging with them, leaving reviews.

And the reviews are quite good, especially when the publisher spends a lot of time to make sure that the audiobook has all the things that an audiobook would typically have.

We've been particularly impressed with our progress with Spanish language.

We only introduced it a few months ago and we have a lot of listeners who are very eager to consume the Spanish language auto-narrated audiobooks.

We think that's probably because our Spanish language catalog gap between eBooks and audiobooks is even wider than our English gap. If you have either a book that's natively in Spanish or translated to Spanish, we've seen a lot there as well.

Joanna: I wonder whether we're going to see the growth in AI narration audiobooks in markets other than English, mainly because of that reason is that this market is hungry for content.

Whereas the English audiobook market is so mature that there is some enough content to a point, but there's also a whole network of narrators.

Whereas I remember going to Frankfurt Book Fair and heard a lady from Ghana saying that across the whole of Africa, and India, and Asia, people don't have enough content in their own language, or their own dialect.

And we just can't replicate audiobooks in every single language and every single dialect without AI. I almost feel like even if in English, it's not adopted so much, although I think it will be.

I feel like these other languages might have incredibly rapid growth [because they don't have much existing content].

Ryan: We see the same. That's why we're adding more languages. We are starting with the ones that have a little bit more robust of a audiobook market. But as you said, no language has as robust a market as English. So the gap is always wider in other languages.

One thing, just more broadly, is as we do more other languages for auto-narrated audiobooks, we also look at how can we bridge the gap between English books and books in other languages to help with that process there?

I do think we'll see significant growth not only in auto-narrated audiobooks, generally, but specifically by language, because there's just much more demand for them because they don't have access to a lot of audiobooks in their language.

Joanna: Absolutely. I think that's always my ethical response to people is, well, how can we not have AI audiobooks when 99% of the world pretty much is not able to access them in either an accent, or a voice, or a dialect, or a language, that they want to for an affordable cost?

Let's talk about money.

How much does it cost (at the moment) to use the Google Play version of the auto-narration?

Ryan: Currently, it is entirely free.

It is free right now because the product is in our beta program. And during the beta program, we have said that we will always have it free. It's free for creation and publishing for an unlimited number of auto-narrated audiobooks, of course, as long as your eBook is live on Google Play.

Once you create it and publish it, you have the ability to download it and to distribute it to any other retailer or distributor that you so choose. We don't have a specific end date in mind for the beta. We've learned a lot through this beta program and are expecting to continue it for a while.

With that said, we do, when we end the beta program, expect to have a very modest fee after the beta program. It'll still be significantly, significantly cheaper than a traditional audiobook. But that is not now. And we're still in the beta program. So we'd continue to encourage publishers to try it out as much as they can.

Joanna: Yes. So as of mid-2022, it is free. However, I do want to say that obviously Google is a very big company that uses data to train AI. And I feel like we are helping Google train a text-to-speech AI with what we're doing with some of the fine-tuning.

I absolutely think it's worth it. But I do think that it is data that we are helping with.

It is free financially. But I do feel that Google is getting a benefit as well. What would you say to that?

Ryan: We do use the inputs, especially in the pronunciation correction tool, which is our internal name for it, to try to improve it globally for all publishers.

If you find a mistake with one of our pronunciations, and we think, ‘Oh, that's actually like a universal mistake. It's not just a personal preference,' we do update it globally, so that not only the other publisher across the way would be able to have that improvement.

But you specifically, if you came back and were doing another book a few months later, let's say, you wouldn't have to fix that again. So we are taking the inputs that publishers provide to try to improve it more broadly. And I think that's part of the program for us.

We understand that this goes hand-in-hand, benefits go both ways between us and publishers. And that's why we've started off with it free. Even when we do remove the beta program, I think publishers will be very surprised at how modest the fee will likely be.

Joanna: In terms of selling the audiobook, what I don't like about general digital stuff is that we're in this race to the bottom, everything should be free on some subscription model, which devalues a lot of the content.

But then I hear people saying, ‘AI narration or auto-narration should be free or cheap because it doesn't cost anything to create.' But it's still the creators' intellectual property, it's still an experience for the listener.

What is your guidance on pricing an auto-narrated audiobook?

Ryan: We don't offer specific guidance for a specific audiobook, but I can say broadly, I really think about what is the additional cost of production for this audiobook when you already have the eBook created. eBooks in of themselves takes a lot of time to write a book, takes a lot of time to write a book that people want to buy, and that should be monetized.

That is a lot of effort both to create it, but also to go through the process of publishing it. So what we advise is to think about, okay, I have an eBook for sale at a certain price. I could create a traditional audiobook that'd be quite expensive for me to do so.

But if I'd create an auto-narrated audiobook, costs are very minimal. So we would encourage publishers to think a little bit more in terms of their eBook price. We're not advising publishers to do anything to say, like, this should be free, because the cost of creation is still about the eBook. You still needed to write the whole book, and you want to monetize it in the best way possible.

We encourage them to think more like your eBook prices, a little bit less like a traditional audiobook, because it doesn't have that huge upfront cost they need to recoup. We do see a lot of success with publishers monetizing these very similar to eBooks.

And if you have a series, you maybe would put the first book in the series free, and try it out that way. There really isn't too much differences that we would see in terms of behavior besides that you now have an audiobook that is a little bit cheaper than a typical audio-book.

Joanna: I agree. Definitely cheaper than a human-narrated, but still not free, unless it's, as you say, a first in series or something that you are doing a promotion on but just not in general.

As part of an increase in the use of AI tools, I've worked with the Alliance of Independent Authors to create an ethical framework for use, which I'll link in the show notes. And that includes labeling of AI narrated audiobooks, or auto-narrated, as you call it.

Imy audiobooks, left narrated by me, right digitally narrated

I've added on my covers, I've added this round circle, which says digitally narrated, and I put it very clearly in my title as well. But it's not a rule. It's just something that I feel is important.

What are your thoughts on this kind of labeling or anything around making sure that customers know an audiobook is auto-narrated?

Ryan: First and foremost, I think it's very important to communicate to customers what they're purchasing. If you told someone that your book was a romance novel, and it was actually a horror novel, that would be quite a frustrating experience for your customer.

We don't really see any differences for auto-narrated audiobooks, which is why we've embedded two areas where we tell our customers that this is not a narrated audio-book. One is on the book detail page.

If you go to an auto-narrated book in any of our surfaces across iOS, Android, or our web experience, we do have a badge on every single book detail page that says, ‘Hey, this is an auto-narrated audiobook. Here's a link to a help center to help you understand what that means You can learn more about it.' So that's the first thing.

The second is, for every single auto-narrated audiobook, on our platform, we have a one sentence intro that just says, ‘Hey, this is an auto-narrated audiobook'.

And it just is there to inform users that they are playing a sample, they know right away what they're listening to. And it's actually not in the one that gets downloaded. So publishers don't have to worry about that.

We only want it for the ones that are on our platform, because it does mention Google in the sentence.

But beyond that, it's really up to the publisher, as you said, to figure out how they want a message that this is not an audiobook you can do it with the cover, the description, or the price, as we just talked about.

Joanna: I want to encourage people listening, I want to remove any stigma or shame that people seem to be feeling around using AI narration. It's almost like, people think, well, we've got to hide it somehow. And that's not the point.

I feel almost this comes back to thinking that the voices aren't good enough. It's like, well, there's lots of problems with humans, breathing noise and, as I said, mouth noises.

I call them the ‘voice artifacts' of humans. And then there are the ‘voice artifacts of auto-narration.'

I feel like there's something that we can probably embrace as distinctive, and a reason why you know that it's auto-narrated.

We're not trying to hide it through making it error-free. We're labeling it.

We're embracing the technology and proudly doing this, as you say, to try and bridge the gap between eBooks and audiobooks.

I want to change the feeling around the whole thing. Obviously, you're very positive about it.

What do you think about that?

Ryan: I think that's exactly the way that we're thinking about it as well. You said it earlier. But it's really about a whole new category of audiobooks.

You have a traditional human-narrated audiobook and an auto-narrated audiobook and they're just different categories.

People should know which category they're looking at, and decide based off that and all the other factors that they have for the book whether they want to purchase it and listen to it. It should be transparent in what's going on.

Ultimately, I think it's better for the entire ecosystem to have this new category so that we can really try to bridge that gap and get more and more people listening to audiobooks.

Joanna: Absolutely. As an audiobook listener myself, I listen a lot of the time at like 1.5x speed anyway. So even humans sound like a robot.

So, okay, you said that we can download those files. And it's really easy, you just download it. And then for example, I'm selling through my Shopify store, which is creativepennbooks.com. And I'm using BookFunnel to deliver those files.

People can buy them on Google Play, Google Play Books, they can also buy them direct.

But the reality is that some of the biggest companies in the audiobook market do not allow the sale of audiobooks that are not human-narrated, it's the conditions on their site.

But I wonder, we've had recently Spotify acquired Findaway and an AI narration company, Sonantic recently, that's still going through. But it feels like mainstream adoption is on its way. And I would say that that would benefit everyone. If Spotify did AI narration, it would benefit Google's AI generation and vice versa.

What do you think in terms of how these audiobook incumbents will adopt AI going forward?

And that's your opinion, obviously.

Ryan: Certainly I can't speak to the specific decisions that a retailer makes. Every retailer makes their own decision. But there are a few factors that I think are going to be poignant when retailers do make those decisions.

One is just what we've been talking about this whole time is the quality of the auto-narration, and we believe our narration is high quality, and that they are a valuable contribution to the audiobook catalog. If the quality is high, that's an important factor.

Another is just supply. Are there a lot of these auto-narrated audiobooks there?

And also looking at languages, as we talked about before. If you have a large gap in your Spanish language content and there's a large supply of auto-narrated, Spanish content, that's, again, something important to consider, as these retailers are making their choices.

And of course, the last one really is demand for the book in audio format. If you have a lot of users that are hoping to consume a book through an audiobook or through a listening experience, that is certainly very important to consider, overall.

But ultimately, we want our publishers to be able to sell their content in whatever storefront they choose. That's why we don't have any exclusivity requirements and we are supportive of wide distribution across all retailers.

I also think that it'll really increase more broadly as it goes outside of just the books industry as well.

In podcasting, as we're talking now, a lot of publishers are looking at, ‘I have a lot of blogs and posts that I do every single week, how can I turn those easily into podcasts?'

I think that's where we'll also see growth, and that will go hand-in-hand with auto-narrated audiobooks as well.

Joanna: Yes, as people are more used to it, I feel like right now, people are quite used to voices through their various devices. And they know that that's not human. You get used to hearing the different voices.

As you say, I think the adoption is happening. And I think it's only going to speed up and, yes, the voices are getting better and better.

Would Google be looking at any point to help people like me who are narrators to train our own voice clones?

Ryan: This is kind of an important thing that we've looked at over the last few years. I think the first important principle that we have is that we always want to make sure that it's very transparent that to the author that they're like signing a contract to create their own text-to-speech voice model.

We do see a lot of difficulties with doing that on a scaled basis. So that if we had it on a website, anyone could just submit some audio files of someone speaking and say, ‘Hey, I want to create a text-to-speech version of this.'

There's not a lot of proof that we can say that this is the specific individual. So we do have challenges with rolling that out on a scale basis.

Every single one of our narrators that we have today, our teams, mainly with Google Assistant, has met in person, they recorded in studio, and they signed a very specific agreement to say, ‘Hey, we can use these samples to record and improve a text-to-speech model.'

I don't think we have a good solution to do this on, as I would call like a scaled basis. But we are looking at it and trying to figure out if there's a way that we could do it with select people, but it's quite far away, I would say and I think we really haven't seen the industry move there just because of all the potential bad actor scenarios that we would have to deal with.

Joanna: This whole area is a nightmare. But as a narrator myself, I feel like I need to license my voice before someone else does it since I've been podcasting since 2009. There's so much of my voice out there. And as you know, you don't even need that much anymore to train a model with a voice.

I feel like anyone can do this stuff. But as you say, we're all responsible people and company, so we need to do it properly. But yes, these are interesting times for sure, but we are out of time.

Where can people find Google auto-narration? And where can they get help if they need it?

Ryan: We have a dedicated website for auto-narrated audiobooks. We have a link, I can maybe give it to put in the show notes. But if you who are looking for it, all you need to do is type in auto-narrated audiobooks into Google search, and it'll come up.

On that page, you'll find information about the entire program, listen to samples. There's also help center articles that tell you exactly how to get started with auto-narrated audiobooks. And there's actually one that provides a full list of all of the narrators that we have in our programs.

You don't need to be logged in to our partner center. You can send it over to a friend or a colleague without having to go into our partner center and listen to samples. I think that's always the first step that we tell people to get started with. Just try it out yourself and give it a listen.

Joanna: Fantastic. Well, thanks so much for your time, Ryan. That was great.

Ryan: Thank you. It's been great.

Joanna Penn:

View Comments (8)

  • This podcast edition couldn't have come at a better time for me.

    I have been vaguely aware of google auto narration since the last time you brought it up, but never thought I would have a need for it.

    I voiced my first book (Cobra Flight) and fully intended to do the sequel and others. I used to be a television and radio journalist and producer, and my voice was one of my prime journalistic tools.

    But, about a month and a half ago, the gods laughed in my face, and I developed laryngitis/ bronchitis, effectively losing my voice.

    Despite treatment and the best of Canadian publically funded health care tests, my voice is nowhere near returning to normal.

    My second book, (Shark Flight) is stalled because I cannot record the final draft, a step I consider vital for proofreading, pacing, and general understandability.

    It also obliterated my plans for launching a podcast related to the books.

    Fortunately, I listened to this week's conversation with Ryan Dingler and the heavy depression overcast on me for so long started to break.

    Since my books are wide and I am not exclusive to Audible or FindAway for the first one, I intend to make use of Google's technology to help with making sure that a more than decent version of Shark Flight will be available during the general launch.

    I also think that Google can help with the podcast scripts. When my voice returns to full use I will record a human voice version for people such as FindAway who do not accept machine aided editions.

    A most informative and useful podcast. Thank you.

  • Dear Google: Please, PLEASE make this available to those of us publishing from South Africa! I keep checking my Google partner account every few months to see if this has changed, and I'm disappointed every time!

    P.S. I do understand there are reasons for rolling out features territory by territory, but I have legitimate FOMO not being able to test out the auto narration tool for myself! 😂

  • Google Play autonarration is amazingly good. BUT there is a huge problem for many self-pubbed authors wanting to use this affordable option: Google Play requires you to sell your ebook in their store (which Kindle Unlimited authors can't do, of course) *or* have a 20% (min.) preview of your ebook in their store. But KU oly allows 10% of your book to be available elsewhere. So, KU authors cannot use Google Play autonarration.

    I do wish Google would allow a 10% preview so all those millions of KU authors could use their autonarration service. When I contacted Google Play, they insisted it was a "Google Partner" requirement and wouldn't even pass along my suggestion. Joanna, for the sake of all KU authors wanting to use Google Play autonarration, please could you talk to Ryan about this? Maybe he's unaware of the problem.

    • I think the problem is KU's requirement for exclusivity, NOT Google Play!
      If you want to publish wide, then publish wide.

  • One thing I was hoping you'd ask but didn't was if we can upload AI narrated books we made elsewhere. I actually spent a good deal of 2021 making AI narrated books using Amazon's Polly. I figured out how to upload, code, and tag everything so it would come out sounding really good. But now I have these files and I'm hoping to sell them other places than Direct. I'm going to look into uploading them to Kobo now, since that was mentioned, but I was also hoping to upload them to Google Play. I do plan on using GP for making more audiobooks, too.

  • This was excellent, engaging, and informative. Between this and AI art, it's a whole new world right now, especially for those with limited or no budget. Thanks for this!

Related Post