BeyondWords Blog

BeyondWords Blog

By BeyondWordsBusinessTechnology
Download on the App Store

BeyondWords Blog episodes

  • The importance of text preprocessing in TTS
    Text-to-speech (TTS) is the artificial production of human speech based on input text. TTS can be viewed as a sequence-to-sequence mapping where text gets transformed into one of its many possible speech forms.
    However, this process is not as straightforward as it might seem. Any user of early sat nav systems should be able to recall numerous instances when the technology mishandled pronunciations, often resulting in an embarrassing slip. Nuances in pronunciation become particularly important when dealing with delicate news material or financial research.
    The main challenge of perfecting TTS lies in the nature of text itself, as it's inherently ambiguous and under-specified with regards to many aspects of speech.
    In this post, we outline the key challenges presented in processing input text and describe how a text preprocessor can help.
    Ambiguity in written language
    We can roughly categorize text into natural language and non-standard words (NSWs). Loosely speaking, natural language is any text that can be read out loud immediately, whereas NSWs have to be verbalized first.
    For example "April first" can be read out loud as is, whereas "1st April" has to be converted into its natural language form before it can be read out loud.

    There are many types of NSWs that need conversion, such as dates ("01/01/2001" into "January the first two thousand and one"), currencies ("$100" into "one hundred dollars") or units ("10m" into "10 meters" or "10 million"), just to name a few.
    But even text without NSWs might require some form of disambiguation. For example, abbreviations need to be expanded into their natural form, like "St." into "Street" or "Saint". Another example are acronyms and letter sequences which can be read as words (such as "NASA") or by pronouncing each letter separately (such as "FBI").
    The process of disambiguating and expanding natural language and NSWs is commonly referred to as text normalization. The output of which can be interpreted as a sequence of graphemes - letters that represent sounds in a written language.
    Ambiguity with respect to pronunciation
    The process of normalizing text into a sequence of graphemes removes some of the ambiguity, albeit not fully.
    Take heteronyms - words that are spelled the same but pronounced differently depending on the context they appear in. For example, "lead" can refer to being in charge, in which case it's pronounced like "leed". However, "lead" can also refer to the metal, in which case it is pronounced like "led".
    Pronunciation can also vary based on regional or personal preferences. For instance, the word "either" can be pronounced as "ee-thur" or "eye-thur". Or consider the brand name "Nike", which in the US is commonly pronounced "nai-kee", whereas in other parts of the world people might say "nyke".
    Furthermore, unknown words can appear over time, such as "gif" a few decades ago. Some people use a hard "g" (like "gift"), while others use a soft "g" (like "giraffe"). Sometimes these new words deviate from existing orthographic rules, which makes inferring their pronunciation challenging.
    The examples above motivate the need to better specify the intended pronunciation of input text. This is commonly achieved with phonemic transcription, a process which transcribes the grapheme sequence into a sequence of phonemes.
    Phonemes are the smallest units of sound that can distinguish one word from another, and are represented by special alphabets such as the International Phonetic Alphabet (IPA). With hundreds of symbols it allows for more granular control than mere grapheme-based input.
    Ambiguity with respect to various aspects of speech
    Even after text normalization and phonemic transcription, a reader may often have little information about many of the aspects of speech, such as timbre, which characterizes who the speaker is (e.g. their age) or where they come from (e.g. their regional accent or dialect).
    It is often unclear how a given text should be read out loud in terms...
    7 min
  • Audio Engagement Report 2023
    We've rounded up the most important audio engagement statistics from 2022, including AI audio narration benchmarks from our own analytics.
    We're listening more than ever
    The Infinite Dial surveys found that online audio engagement grew considerably in 2022. The percentage of the population listening weekly reached:
    71% in Australia (up five percentage points)
    67% in the US (up five percentage points)
    64% in Canada (up three percentage points)

    40% of listeners plan their day and activities around audio content.
    Spoken-word is most popular format of daily listeners
    The Spoken Word Audio Report showed that spoken-word audio continued to displace music in the US. Americans listened to an average of 1 hour and 4 minutes every day, which represented 29% of their total listening time (an increase of one percentage point).
    The effect is particularly pronounced among daily spoken-word audio listeners, who represent 46% of the population. For the first time, this group spent more time with spoken word than with music (51% vs 49%). On average, they listened to the likes of audio articles, audiobooks, and podcasts for 2 hours and 19 minutes a day—13 minutes longer (+10%) than in 2021.

    Podcast data also demonstrates an uptick in spoken-word audio consumption. 34% of people surveyed across 20 international markets said they had listened to a podcast in the previous month, up from 30% in 2021. Growth was strongest in:
    Sweden, up seven percentage points to 44%
    Ireland, up five percentage points to 46%
    Norway, up five percentage points to 42%
    Argentina, up five percentage points to 34%
    Singapore, up five percentage points to 33%
    All this increased listening is translating into increased revenue. In January 2022, 25% of listeners in the US said they were very or somewhat likely to pay money for a podcast over the next 12 months. Just 16% said the same in March 2021.
    Meanwhile, the IAB forecast US podcast advertising revenues to grow by 43% last year, to $2 billion.
    Publishers increase their audio investment
    With listening habits strengthening, 80% of news leaders said they would put more resources into podcasts and digital audio in 2022. As a result, audio articles and podcasts became a more integral part of publishers' revenue strategies.
    Aftenposten, Norway's largest news publication, achieved huge success with its podcast operation, reaching more than 160,000 paying subscribers. They consequently partnered with BeyondWords to add automated narration to their articles, using a voice clone of their podcast host.
    Director of Product, Karl Oskar Teien, said: "How do you break through in the attention economy? We think it starts by adapting to the fact that, for many users, the audio-first world is already here."

    The Irish Times, the most popular digital news subscription in Ireland, expanded its audio offering as part of a site and app relaunch. This helps them to tackle "unread guilt factor", a leading cause of subscription churn.
    "Recognizing that our audience was often on the move or busy with other things in life, while at home, we wanted to provide an alternative or complementary way for them to consume our journalism. We very much believe that what is most important is the journalism itself and the means of delivering it, the platform, should be influenced heavily by the audience preference," said digital editor Paddy Logue.

    But it's not just the news industry taking advantage of the audio boom.
    In financial services, for example, analysts are using AI audio to help clients extract more value from financial research. Top firms are also using podcasts to better reach and engage their target audience.
    AI audio narration benchmarks
    To see how listeners interacted with AI audio narration in 2022, we analyzed more than 17 million audio plays through the BeyondWords Player. This is the audio player that some of our publishers use when embedding their audio versions or sharing them via URL.
    We found an average listen rate of 4.19%. That me...
    6 min
  • AI audio and the future of financial research
    You package financial research into crisp and concise reports because your clients' time is precious. For the same reason, you should make your insights available in audio.
    Listening enables engagement and multitasking where reading doesn't. Clients can consume audio briefings while they're driving to work or grabbing lunch, for example. They'll have more opportunity to consume your valuable research and get the most from your products.
    It's also the practical option when they're away from their computer. Who wants to navigate a PDF on their phone, with awkward zooming and endless scrolling, when they can put in their earbuds instead? They'll no doubt appreciate the much-needed screen break.
    Convenience is particularly important when you consider the limited timespan of most financial research. If clients don't get round to your daily briefing today, they're unlikely to revisit it tomorrow. So, you really need to capitalize on every engagement opportunity.
    Audio can also improve accessibility, and many simply prefer listening or audio-assisted reading. Approximately 30% of Americans are auditory learners, and 60% of listeners say they process audio information more efficiently.
    So, as well as being more willing and able to engage in the first place, listeners may extract more value from your research. That will translate into better retention and conversion.
    The financial audio landscape
    Many financial services firms are already tapping into the demand for audio insights, using the format to demonstrate their digital-first approach and better engage target audiences.
    Since 2018, JPMorgan has delivered its annual letter to shareholders in audio. Not only is the additional format user-friendly, but its Chairman and CEO's voice lends the message personality and authenticity. The company also publishes audio commentaries from Global Liquidity Portfolio Managers on its website.
    UBS, Morgan Stanley, and Goldman Sachs are just some of the major players reaching listeners through podcast platforms like Spotify and Apple Podcasts. And their presence is attracting even more finance professionals to these apps, to the potential benefit of the entire finance industry.
    Meanwhile, publications like The Wall Street Journal and Bloomberg are using voice AI to deliver audio articles at scale on their websites and apps. Bain and Coronation are two of the financial services firms using BeyondWords text-to-speech to make their expert analysis and insights audible.
    The practical approach

    The case for audio hasn't only been strengthened by widespread adoption of smartphones and digital listening habits: advancements in voice AI have significantly improved the ease, cost, and speed of implementation.
    Using our API, WordPress, or Eidosmedia integration in combination with our player SDKs, financial services firms can make insights on their website or app listenable automatically. Audio is typically embedded within just a few minutes of publishing and dynamically updated.
    Audio reports and briefings can be shared via URL, too - ideal for emails and PDFs. We also recommend that you auto-distribute via podcast feed to tap into existing audio habits. Subscribers on platforms like Spotify and Apple Podcasts will receive the latest episodes direct to their app, with notifications prompting engagement. You can even build custom feeds or playlists to group audio by category.
    Research teams also have the option to create or edit audio in our Text-to-Speech Editor. Intros, outros, and sound effects are easily inserted to enhance the listening experience and strengthen sonic branding.
    Firms get access to our AI voice library, but may prefer to create and use voice clones of their financial analysts or other stakeholders. This can lend the briefings more credibility and assist with relationship-building, without additional demand on already-busy staff.
    All voices are supported by our customizable natural language processing algorithms, which ensure accurate ...
    5 min
  • The Irish Times: Driving subscriptions with AI audio
    The Irish Times is the most popular digital news subscription in Ireland, with 39% of all news subscribers in the country saying they're signed up to the publication.
    One of the many features contributing to its success? Audio.
    We spoke to Paddy Logue, Digital Editor at The Irish Times, about the publication's adoption of voice AI and its role in a wider news audio strategy.
    The move into audio
    The Irish Times began offering audio articles with BeyondWords in 2019, when listening habits were really picking up pace in the country. Earlier that year, Reuters said that podcasts "appear to be reaching critical mass" and found that 37% of Irish people listened to podcasts monthly.
    Podcasts remain extremely popular in Ireland. Last year, the country led the pack among 20 international markets, with 46% of the online population listening monthly.
    "We can see, through the growth in podcasts, audio becomes more of a typical activity for news audiences who don't necessarily have the time or inclination to sit down and read," said Logue.
    "Recognizing that our audience was often on the move or busy with other things in life, while at home, we wanted to provide an alternative or complementary way for them to consume our journalism. We very much believe that what is most important is the journalism itself and the means of delivering it, the platform, should be influenced heavily by the audience preference."
    At the same time, voice AI had reached a tipping point. With the development of lifelike AI voices and automated audio publishing tools, there was finally a viable alternative to human voice-over. The Irish Times knew they needed to take control of their audio UX and sonic branding.
    "We were very impressed with the quality of the AI voices, and the customer service and support during implementation was seamless from the BeyondWords team," said Logue.
    Using audio to drive subscriptions

    In Ireland, 16% of the population pays for online news. Subscriber-only features, like the 'Listen' audio service, are helping The Irish Times to capture the largest share of this market.
    The 'Listen' brand brings together various podcasts and audio versions of articles. And while a selection of opinion columns and feature articles are author-read, with the audio hosted on BeyondWords, voice AI is crucial to delivering the bulk of news audio on time and at scale. Integration with their Arc XP CMS, alongside our player SDK, means audio versions are created and published automatically.
    "We now have AI audio on almost all of our articles and we have decided to make this a subscriber benefit. When a user clicks the icon to listen they are prompted to log in or subscribe," said Logue. "We see it as a major part of a more holistic audio offering."
    And it's not just about enticing people with listening needs and preferences to sign up. Another key benefit of audio articles is reduced churn.
    When subscribers have the option to listen-not just read-they're better able to fit news into their day. That means they're less likely to experience "unread guilt factor", the reason behind 13% of digital news subscription cancellations. Plus, Northwestern University research shows that frequency of consumption is the biggest predictor of retention in digital news.
    A localized voice for a localized audience
    The Irish Times saw positive results with other voices, but it was the adoption of Irish-accented voices that really cemented their commitment to automated audio. "The availability of an Irish accent was a big determining factor in our expansion in the use of AI on our newly designed site," says Logue.
    "Many listeners don't realize it's not a human that's speaking."
    Not only does the localized voice allow them to better connect with their localized listeners, but it delivers more accurate pronunciations of regional words. This works in conjunction with voice support from BeyondWords.
    For example, thanks to our customizable NLP algorithms, we're able to add pronunciation...
    5 min
  • 15 news publishers using audio articles to drive revenue
    Audio articles are on the brink of ubiquity. There was "rapid development" in the format last year, according to Reuters Institute, and 80% of media leaders are planning to invest more in digital audio throughout 2022.
    "In our conversations around trends and predictions, it is clear that many publishers believe that audio offers better opportunities for both engagement and monetization than they can get through similar investments in text or video", wrote Nic Newman.
    Let's take a look at some of the leading news publications making their stories listenable:
    The New York Times

    With 7.6 million subscribers as of September 2021, The New York Times is the world leader in digital news subscriptions. It's also a major force in audio journalism, reaching around 20 million listeners a month.
    The publication is currently testing New York Times Audio, an app that brings together "audio journalism and storytelling, as well as narrated articles, podcasts and audio content from a slate of premier publishers".
    The Daily, its flagship podcast and audio newsletter, reaches around 4 million listeners every day - that's almost twice the print edition's peak circulation. Chief executive Meredith Kopit Levien said: "It's helpful in driving affinity to the brand. It's harder to track directly how it drops people into the core news subscription funnel, but we have every reason to believe it does, given how our results have improved.
    "What I would say on audio is that's a place where we think there's going to be real demand at high CPM for some time to come, and our product set is expanding," said Kopit Levien. "I think audio will be something to watch in our ad business for some time to come."
    The Washington Post

    With 3 million subscribers as of November 2020, The Washington Post is the second biggest news subscription service worldwide.
    In 2020, readers who listened to audio articles on The Washington Post's apps engaged more than three times longer with their content. This persuaded the publication to adopt voice AI, so that they could deliver audio articles at a larger scale.
    "We conducted user research and learned that users want to stay informed but are busy, so they appreciate an option to get up to speed on the latest news developments while cooking dinner, running errands or exercising," said Emily Chow, director of site product.
    "What we've learned from users is that they listen to the news while doing other things, and are consuming far more content than they would normally," said product manager Leila Siddique.
    The Irish Times

    The Irish Times is the most popular digital news subscription in Ireland, with 39% of all subscribers in the country saying they're signed up. The publication's "Listen" brand plays an important role in the publication's revenue strategy, helping them to increase conversions and reduce churn.
    Using a BeyondWords integration with Arc XP, The Irish Times began offering automated audio articles in 2019. They now use an Irish-accented voice from our library to deliver engaging narration on their website and app, as well as distributing their human-read audio through our platform.
    "We now have AI audio on almost all of our articles and we have decided to make this a subscriber benefit. When a user clicks the icon to listen they are prompted to log in or subscribe," said digital editor Paddy Logue. "We see it as a major part of a more holistic audio offering."
    The Wall Street Journal

    With 2.8 million subscribers as of September 2021, The Wall Street Journal is the third-biggest news subscription service globally. The publication has been pretty quiet regarding its audio strategy and results, but it delivers audio versions of most articles using AI voices.
    The Economist

    The first major publication to invest in an audio edition, The Economist has long flown the flag for listenable journalism. It all began when the publication identified a problem it called "unread guilt factor": subscribers were canceling because t...
    12 min
  • Unread guilt factor: The silent killer of news subscriptions
    In 2011, The Onion published an article entitled "The Economist' To Halt Production For Month To Let Readers Catch Up". It was satirical, but it probably touched a nerve at The Economist HQ. After all, a lack of reading time was the number-one reason for subscription churn.
    Five years later, the publication publicly acknowledged the article - and the problem.
    "The Economist can be overwhelming," wrote Product Manager Richard Holden. "It can be difficult for even a motivated reader to keep up."
    When subscribers watch print editions pile up, see headlines stay un-clicked, they feel guilty. How can they justify paying a monthly fee for something they're barely using? Who needs a constant reminder that they're falling behind on their reading goals? And so, they cancel.
    It's a problem that The Economist's then-Head of Strategic Development, Denise Law, referred to as "unread guilt factor".
    She wrote: "Many of our former subscribers found it difficult to stay on top of The Economist edition each week and simply gave up. We spoke to some of them to validate that the problem actually existed. It did."
    Subscription guilt isn't unique to the news industry, but it is particularly prone to it. With such a huge volume of content published every day, it's easy for readers to get overwhelmed. Nieman Lab found that 13% of news subscription churn results primarily from people having too much to read in too little time.
    So, how are publications tackling subscription guilt? With audio.
    Tackling unread guilt factor with audio
    Giving subscribers the option to listen - not just read - means there are more opportunities to stay updated. They can consume your content while they're driving, exercising, or cooking, for example.
    "The idea is just to leave it up to the reader to decide what the most convenient form of consuming the content is, and in many situations that will be audio," said The Economist's then-Digital Editor Tom Standage. "We get a lot of feedback from people saying that it is how they stay on top of the information coming their way."
    In fact, many people go out of their way to multitask with audio: 25% of podcast listeners listen to "fill empty time". While readable content is somewhat of a burden, listenable content is actively sought-after - something to keep minds occupied and entertained during mundane or physical tasks.
    "What we've learned from users is that they listen to the news while doing other things, and are consuming far more content than they would normally," said Leila Siddique, the Washington Post's Senior Product Manager.
    The New York Times identifies these extra opportunities for engagement in the "Audio Day." Engage subscribers at these times, and they're far less likely to feel guilty and cancel.

    Plus, Northwestern University research shows frequency of consumption is the biggest predictor of retention in digital news.
    "Our evidence suggests that the audio edition is a very effective retention tool," said Standage. "Once you come to rely on it, you won't unsubscribe."
    And let's not ignore the many other benefits of offering articles in audio. It's not just time-poor readers who are turning to audio articles, audio newsletters, and podcasts. There are all kinds of reasons that people listen and many ways to capitalize on audio habits.
    AI brings audio to the masses
    At first, The Economist's success was hard to replicate. The publication employed a large team to produce its audio editions, and they had an afternoon a week in which to do so. Those with fewer resources at their disposal, and shorter deadlines, found audio publishing infeasible.
    Our text-to-speech platform changes the game. We give publishers access to lifelike AI voices and a voice cloning service, and offer tools that allow for seamless integration into the publishing workflow. Engaging audio versions can be distributed effortlessly, affordably, and at scale. There is finally a viable alternative to human-read audio.
    In their journalism predictions...
    5 min
  • Why audio content is booming
    Modern audiences are engaging with audio content like never before.
    Spoken-word audio's share of digital listening increased by 40% from 2014 to 2021. Almost half of over-13s in the US are listening for 2 hours and 6 minutes every day. Podcasts, audiobooks, and narrated content are more prevalent and popular than ever.
    Listening needs and preferences are nothing new, so what's behind this audio boom? It's something we call the audio flywheel.
    What is the audio flywheel?
    The flywheel effect is a business metaphor developed by Jim Collins. It suggests that multiple small efforts are required to build momentum, and that momentum will then perpetuate itself.
    We can apply this concept to audio content. Various factors have contributed to the rise of audio and, now that it's built momentum, its growth is showing no signs of slowing.
    Consumers, creators, advertisers, and technology companies all form part of the audio flywheel. Developments in one area have a knock-on effect in others, forming a positive feedback loop.
    Getting the wheel turning
    One of the main contributing factors to the rise of audio was increased accessibility of audio publishing and broadcasting. While the internet made it possible for anyone to share digital audio, hardware and software developments made production more affordable. This empowered more creators to explore the format's potential and attract listeners with new types of content, such as podcasts.
    Recently, advancements in text-to-speech have reduced the barriers even further. Our lifelike AI voices and sophisticated tools provide a practical alternative to human voiceover, allowing creators to produce engaging spoken-word audio quickly, easily, and affordably. This has opened the doors to previously unviable formats, such as audio articles.
    Meanwhile, ownership of audio-enabled hardware has rocketed. Smartphones and AirPods make it easy to listen anywhere, while bluetooth speakers and smart speakers promote listening in the home. Improved in-car connectivity means listening to your own audio while driving is just as convenient as switching on the radio. There's also been widespread adoption of audio apps like Apple Podcasts and Spotify, which are constantly developing new ways to keep listeners engaged.
    Changing media habits
    Consumers began adopting new media habits, too. Most importantly, a proclivity for multitasking. Technology gives us 24/7 access to information, and our brains deliver dopamine when we learn something new or feel productive. So, we're prone to filling every minute of spare time with content consumption.
    Audio is the most multitasker-friendly form of media: we can listen while we're driving, exercising, cooking, cleaning, getting ready…the list is endless. Combine this with the increased availability and accessibility of audio content, and it's little wonder it's become the format of choice for busy people. 71% of monthly spoken-word audio listeners say the ability to multitask plays a role in their habit.
    Concerns around screen time are also starting to play more of a role in audio consumption. More than half of Americans are on their phone more than they'd like, and around a third are taking steps to curb their habit. As a result, more are turning to audio formats, which fulfill their desire for information and entertainment without gluing their eyes to a screen.
    Audio goes commercial
    As audiences spent more time with digital audio, advertisers came along with them. And they saw impressive results - one Nielsen study found audio ads delivered a 24% higher recall than display ads and were twice as likely to drive purchase intent.
    US podcast advertising revenues exceeded $1.4 billion in 2021, equivalent to 72% year-on-year growth - twice that of the total internet advertising market. The IAB said this was fueled by "a continually expanding user base consuming a growing library of engaging and diverse content".
    The subscription economy has also been picking up pace. The proporti...
    6 min
  • Changelog: Aliases, pending review, ad positioning & more
    Welcome to the BeyondWords changelog for July 2022.
    For the most recent updates, visit our dynamic changelog.
    Create custom text-to-speech rules with aliases
    You can now add custom text-to-speech rules using aliases. This is helpful for expanding abbreviations. For example, you can decide that "MSFT" is always pronounced as "Microsoft" or "Sen." is always pronounced as "Senator".
    You can also provide pronunciations for sensational spellings. For example, music publications can ensure the band name stylized as "Chvrches" is correctly read aloud as "Churches".
    Don't want the alias to apply every time? Open the Text-to-Speech Editor to add, edit, or delete aliases in a particular audio.
    Let's say you're usually happy with it being pronounced "NBA" but on this occasion you're writing about the National Boxing Association rather than basketball. You can set an audio alias to prevent any ambiguity (and without having to make extensive edits to your text).
    Review audio before distributing with "Pending Review"
    Want to review your audio before it's distributed? Our Pending Review feature makes it easy.
    When Pending Review is enabled, processed audio goes into an unpublished state by default. This means that you can proof-listen and make any edits before publishing your audio.
    Change the position of audio ads
    You can now decide whether your audio ads are played pre-roll, mid-roll, or post-roll. This can help you to boost engagement with your sponsored messages and audio.
    Just head to the "Ad positioning" section in project settings to get started.
    Add categories for Google Podcasts
    If you've read or listened to our guide to optimizing your podcast feed, you'll know that choosing the right podcast categories can help your show to appear in relevant categories, charts, personalized recommendations, editorially curated collections, and search results.
    And we've now made it easy to optimize for Apple Podcasts, Spotify, and Google Podcasts. Just choose the relevant categories when setting up your feed.
    Remember, you can set up a podcast feed for an entire project or any custom playlist.
    Try new languages and voices
    Over the past three months, we've added a variety of new voices and languages to our AI voice library, so you might just find a voice that better resonates with your audience:
    Albanian (Albania): Ilir, Anila
    Arabic (Lebanon): Aysha, Abdullah
    Arabic (Oman): Rami, Layla
    Azerbaijani (Azerbaijan): Banu, Babek
    Bosnian (Bosnia): Vesna, Goran
    English (Great Britain, UK): Arthur
    English (USA): Roger, Tony, Jason, Jane, Davis, Nancy
    French (Canada): Liam
    Georgian (Georgia): Giorgi, Eka
    German (Germany): Daniel
    Italian (Italy): Rinaldo, Pierina (Child), Palmira, Lisandro, Irma, Imelda, Gianni, Fiamma, Fabiola, Cataldo, Calimero, Benigno
    Mandarin (China): Yunze, Yunxia (Child), Yunhao, Yunfeng
    Mandarin (China, Liaoning): Xiaobei
    Mandarin (China, Sichuan): YunxiSichuan
    Mongolian (Mongolia): Yesui, Bataa
    Nepali (Nepal): Sagar, Hemkala
    Portuguese (Brazil): Yara, Valerio, Nicolau, Manuela, Leticia, Leila, Julio, Humberto, Giovanna, Fabio, Elza, Donato, Brenda
    Spanish (Mexico): Yago, Renata, Pelayo, Nuria, Marina (Child), Luciano, Liberto, Larissa, Gerardo, Cecilio, Carlota, Candela, Beatriz
    Spanish (US): Pedro
    Tamil (Malaysia): Surya, Kani
    Fixes and improvements
    Support for Ghost 5.0
    Faster reprocessing speeds
    Improved Playlist SDK
    Improved docs and guides
    4 min
  • How AI audio fits into your publishing workflow
    An efficient workflow is essential in an increasingly competitive media industry. 67% of publishers say automation is of increased importance this year, and 81% of media leaders say AI will be crucial for newsroom automation.
    It's little wonder, then, that those looking to capitalize on the benefits of audio publishing are turning to voice AI over human voiceover. There's no need for costly recording sessions or complex logistics - the technology can handle all the hard work for you.
    But using voice AI doesn't have to mean handing over the reins completely. A little human input can elevate audio to the next level, allowing you to strengthen your sonic branding and boost listener engagement.
    The stages of an audio publishing workflow
    1. Create audio
    You can create audio automatically by connecting your CMS to BeyondWords. If you use WordPress or Ghost, our plugins make this very simple. Otherwise, you will need to use our RSS Feed Importer or API - this will require some development knowhow.
    Once setup is complete, newly published content will be auto-imported to BeyondWords, where it will be converted into speech using your default voices. You can control which elements are converted into audio, and how they are pronounced, with custom text-to-speech rules.
    By default, your audio will be automatically updated in line with changes to the source text. This means that the narration will always remain accurate. However, if you wish to differentiate the audio version from the text version, you can manually edit it anytime.
    You can create audio manually using our Text-to-Speech Editor. This allows you to turn any text into speech using any combination of voices. You can also insert your own audio files.
    Whether you're repurposing written content or writing audio from scratch, you have the opportunity to optimize your text for an audio format. You can also use different voices and sounds to enhance the listening experience and strengthen your sonic branding.
    2. Review and edit audio (optional)
    If you would like to review your audio before it's distributed, you can enable "Pending review". This means that audio goes into an "Unpublished" state by default, and must be manually "published" before it can be shared with your audience via an audio player or podcast feed.
    While this adds an extra step to your publishing workflow, it does mean that you can perfect your audio before anyone else hears it. This keeps listeners listening for longer - and coming back for more.
    After downloading and "proof-listening" to your audio, you can fix any problems by reporting a voice issue or adding a custom text-to-speech rule. You can also make edits in the Text-to-Speech Editor - edit your wording, change the voice associated with any paragraph, or insert your own audio files.
    (If your audio was created automatically, you may wish to disable automatic updates, to prevent any changes from being overwritten via your CMS.)
    Once you're happy with the audio, you can change its state to "Published" via the dashboard or API. It will then be ready for automatic or manual distribution.
    3. Distribute audio
    Automatic distributions help you to streamline your audio publishing workflow. They mean you can get your audio in front of your audience with no extra effort.
    If you are adding narration to web or app content - especially at scale - we recommend automatic embedding.
    This means that, when an audio version of a page is available, it's automatically embedded via an audio player. Our customizable Small Player UI, specially designed for on-page narration, is used by default, but you can build your own player UI using our JavaScript, iOS, or Android SDKs.
    You can also auto-distribute audio via playlists. Just create a playlist with custom rules, and qualifying audios will be added automatically. You can distribute these playlists via URL, embed code, or even podcast feed. Once set up, there's no extra work required.
    Manual distributions offer even more ways to ...
    5 min
  • Take control of your audio UX before someone else does
    Modern audiences want content narration. And if you don't provide it, third-party tools will.
    It's already happening. The Read Aloud Chrome extension alone has over 4 million users. And as listening habits become more entrenched, we're likely to see more sophisticated screen-reading solutions come along - and higher uptake of the technology.
    This might sound like a good thing for publishers, like third-party tools handling all the hard work for you. But there's a big problem: losing control of your audio user experience (audio UX).
    You don't get to decide what your audio sounds like.
    You can't optimize the listening experience.
    You don't get ownership over the audio content.
    This means missing out on valuable sonic branding, engagement, and monetization opportunities.
    Keeping creative control
    Whether they use text-to-speech or human narration, publishers who handle audio production themselves have full creative control over their audio UX.
    Unlike publishers who rely on user-side solutions, they can increase audio engagement by:
    building custom players for their website or app,
    distributing their audio through other channels, and
    creating different types of audio content.
    They can also boost listener retention and strengthen their sonic branding by:
    using the best voice(s) for their content and target audience,
    ensuring their audio content is narrated accurately, and
    elevating audio versions with sound effects and other features.
    They can even measure their success through audio analytics or directly monetize their listenership through audio ads and subscriptions.
    All of these efforts are about maximizing the benefits of audio publishing. But they also play a part in a wider branding and marketing strategy, helping publishers to foster a positive and pervasive image - and stand apart from their competitors.

    Berlingske, for example, has built a custom audio article player using our JavaScript SDK. This prompts users to sign up in order to listen, helping the publisher to drive new subscriptions and reduce churn.
    The team at Aftenposten worked with us to create a voice clone of their most popular podcast host, which means their audio articles are read by a tried-and-trusted voice. They have also collaborated with our engineers to improve how the AI voice pronounces local and out-of-vocabulary words, ensuring the intricacies of the Norwegian language are handled accurately. This has led to a 75% increase in listening duration.
    Protecting your sonic brand
    In 2009, the Author's Guild successfully fought against default text-to-speech on the Kindle to protect authors' and publishers' rights to control - and profit from - the audio versions of their books.
    Digital publishers need to follow in their footsteps. To seize ownership of their audio publishing rights.
    It's only a matter of time until audio narration becomes ubiquitous online. Publishers who act now can build their audio UX on their terms - start growing and capitalizing on their listenership. This puts them in a stronger position in today's media landscape and helps them future-proof for tomorrow's.
    BeyondWords is the practical solution. With our AI voices and CMS, you can publish branded audio content without undue effort or expense. Request a demo with our team.
    4 min

About BeyondWords Blog

From the publisher's feed

Stories, insights, and announcements from the BeyondWords blog—read aloud by AI. Learn how publishers are using audio and video to shape the future of digital storytelling.