With advancements in artificial intelligence, and neural engines, text to speech provides a more enjoyable listening experience.
Beyond Capabilities
What was in its infancy five years ago, is now going through adolescence. BeyondWords may be the most advanced text-to-speech option available. But it is not idle. Users receive new voices, better pronunciation, and new features on a regular basis.
There’s no way this article can break down everything possible with BeyondWords. Their snazzy website is the best source for current documentation and pricing plans.
Frictionless
What I can share are things I’ve learned to improve workflow. Mind you, the company touts frictionless text to speech. If you desire a single voice for the body of your articles and perhaps another for the headline, you can enjoy automated text conversion from an RSS feed.
But if you want to add voices to stories with multiple characters, you’ll want to disable content management system (CMS) updates for such articles.
Voice Selection
It would be nice if you could tap on desired paragraphs and then apply the desired voice from a single dropdown menu. Currently, you must use the voice menu next to each paragraph.
To save time, use the Settings tab to change the default voice to the one speaking the most. Available choices depend upon Language selection within the same tab. There are about three dozen for English (USA). Conversely, there is currently just one for English (Wales, UK).
When there are many options, choosing voices can be time consuming with a multi-person script. You can press the arrow next to a voice to hear a sample. It is also beneficial to make a table that categorizes voices in meaningful ways.
Male US English Voices
| Voice | Pitch | Pace | Age ≈ |
|---|---|---|---|
| Aaron | Mid-high | Med | 16–30 |
| Adam | Mid-low | Med | 20–40 |
| Brandon | Mid-high | Med | 16–40 |
| Brian | Mid-low | Med-slow | 30–60 |
| Carlos | Mid-low | Med-slow | 20–50 |
| Christopher | Mid-low | Med | 20–50 |
| Daniel | Mid range | Med-fast | 20–50 |
| Davis | Low | Med | 30–60 |
| Eric | Low | Med-slow | 30–50 |
| Guy | High (raspy) | Med | 20–40 |
| Ian | Mid range | Med | 20–40 |
| Jacob | Mid-low | Med-slow | 20–50 |
| Jason | Mid range | Med | 20–40 |
| Jimmy (Fox News) | Mid-high | Mid-fast | 20–40 |
| Joey | Mid-low | Slow | 20–40 |
| Jonny | Mid range | Med | 20–40 |
| Justin (child) | High | Slow | 5–13 |
| Kevin (child) | Mid-high | Med | 5–16 |
| Matthew | Mid-low | Med | 20–50 |
| Matt (News) | Mid range | Fast | 20–40 |
| Tony | Low (raspy) | Med-slow | 30–60 |
Note: Carlos has Hispanic accent. Joey could be a bad guy. Tony sounds like a villain.
Female US English Voices
| Voice | Pitch | Pace | Age ≈ |
|---|---|---|---|
| Amber | High | Med | 13–30 |
| Ana (child) | High | Med | 3–6 |
| Aria | Mid-high | Med | 16–40 |
| Ashley | Mid range | Slow | 16–40 |
| Clara | Mid range | Med | 20–50 |
| Cora | Mid-high | Med | 20–40 |
| Evan | Mid range | Med | 20–40 |
| Francis | Mid-high | Med-fast | 20–30 |
| Gabriela | Mid range | Med–slow | 20–50 |
| Ivy (child) | Mid-high | Med-slow | 3–6 |
| Jane | Mid-high | Slow | 20–40 |
| Jenny | Mid range | Med | 20–40 |
| Jenny (multi) | Mid range | Fast | 20–40 |
| Joanna | Mid range | Med-slow | 20–50 |
| Joanna (news) | Mid range | Med-fast | 20–40 |
| Jodi | Mid range | Med | 20–40 |
| Kate | Mid range | Med | 20–40 |
| Kendra | Low | Slow | 30–50 |
| Kimberly | Mid-low | Slow | 20–30 |
| Maria | Mid-high | Med | 20–30 |
| Michelle | Mid range | Med | 20–40 |
| Monica | Mid range | Med | 20–40 |
| Nancy | Mid-high | Slow | 20–40 |
| Salli | Mid-low | Slow | 20–30 |
Note: Highlighted voices have ideal cadence, though others may be appropriate for specific characters.
Try to choose a consistent narration voice for your website that aligns with your brand and resonates with your audience. For lengthy text, you may prefer a medium-fast voice. Complex subject matter may benefit from a medium-slow voice.
With CMS updates enabled (and a URL to the article), audio is reprocessed within a few minutes of editing the article text. This update uses the default CMS voices you have selected (replacing any custom selections). In some cases, this can be advantageous.
You may wish to change the voice for each paragraph from male to female. Change desired voices within Settings. Select “Reprocess audio”. After the audio updates, make additional text edits to audio file if necessary.
Phonetics
BeyondWords voices use AI to resolve many homographs contextually. Sometimes the sentence structure does not have enough words to predict context. Words like “content,” “read,” “lives,’ and “resume” are challenging. Some voices may pronounce them correctly. But perhaps the one you choose may not. As an international app, most voices pronounce “sake” like “sockee.”
To force correct pronunciation, you may need to edit text within the BeyondWords dashboard. Spell out the word phonetically. For example, “rezumay,” “reed,” or “lyves.”
| Homograph | Phonic | Alternate |
|---|---|---|
| apparatus(es) | <!--aperadtess--> | aperadtesses |
| bow | <!--bo--> | bough |
| content | <!--cuntint--> | content |
| lead | <!--led--> | leed |
| live(s) | <!--lyve--> | lyves |
| perfect | <!--per-fekt--> | perfects |
| present(s) | <!--preezent--> | preezents |
| read | <!--reed--> | red |
| résumé | <!--rezumay--> | resume |
| resume(s) | <!--rezoom--> | rezooms |
| sake | <!--sayk--> | sahkey |
| suspect(s) | <!--sespec--> | sespecs |
| suspect | <!--sesspect--> | sesspects |
| tear(s) | <!--teer--> | teers |
| tear(s) | <!--tare--> | tares |
BeyondWords now includes Alias. Highlight a word and type the phonetic spelling, which can apply throughout the current article or entire project. It’s very helpful. But if you use resume and résumé within the same article, phonetic spelling is necessary for distinction.
Sound Effects
With a paid plan, you can upload audio that immerse listeners within your story. When limited to text, you must describe music or narrate doorbells. Consider creating short MP3 sound effects for discretionary insertion.
Using an application like GarageBand or Adobe Audition, you can edit and combine audio clips. For example, using text-to-speech, you may type the words individual members of a crowd. Then overlay them with ambient background shouts to simulate an angry mob. You can even record you own voice.
Each clip within a project must have a unique file name. If, for example, you want listeners to hear footsteps on a squeaky floor several times within a story, duplicate and append the same file with sequential numbers.
Pricing Plans
You can do quite a bit with the free plan but 30,000 characters per month goes quickly. That’s about 1,500 words. This may be a consideration if you add a few new articles or podcasts per month. It is also an option to maintain an existing library of audio with occasional minor edits.
For access to custom voices, you need at least the Creator plan for playlists, analytics (by audio), self-serve ads, audio upload (sound effects), chat support, custom rules, and chat support.
The Pro plan is primarily for multiple users, more characters, five projects, and multiple playlists with custom rules. The Enterprise plan is customized to a corporation’s needs.
Download Audio
You need a paid plan in order to download audio. This is useful for more than just a backup. It allows you to pause and scrub forward or backward while listening for errors.
Mobility
You can do most things on a mobile device as you can on your desktop within the BeyondWords website dashboard. An exception is the inability to edit settings. This is because the Save Changes button is far to the right. As the browser window shrinks, the button is truncated out of view. Therefore, you cannot save a new voice within settings.
Advanced Options
You can filter a custom RSS feed with filters for optimum text-to-speech results. BeyondWords also lets you create an automatic podcast feed to podcast directories like Apple Podcasts, Google Podcasts, and Spotify. It is possible to receive notifications when audio assets are created, updated, or deleted via webhooks.
Did you know that you can have your own voice (or that of a voiceover artist) cloned? Such options are beyond the scope of this article, but it establishes the breadth of features available. For example, audio for many health articles here converts automatically.
Visitor Adaptation
Quality audio is ideal for visually impaired visitors as well as people who do not read well or prefer to listen. As the average time on pages decreases below 20 seconds, listening to audio may increase visitor engagement.
An international audience of visitors rely on dynamic text translation. Therefore, you might discover that a small percentage of them actually listen to your preferred language articles in whole or in part.
