Generated audio saves countless hours in manual work, and while it does a great job most of the time, getting perfect pronunciation is something of an art form and may require several iterations. This article addresses commons questions when generating text-to-speech audio files, including a video on how to make changes during the generation process.
Adjust Pronunciation of Generated Audio in a Specific Scene
Go to the Scenes tab in the navigation panel
Drill down into the project you want to focus on by clicking the arrow
Drill down into the group you want to focus on by clicking the arrow
Locate the scene that you want to regenerate the audio for, and click Edit
Click Update Audio
Make any alterations you would like to the Current guide card text
Select a name from the Choose a Voice Profile drop-down menu
Click Generate Audio
Allow your audio a few moments to generate, and once it appears, click Play
Make any changes to the text if you need to, and click Generate Audio again
If you have made changes, click Play again to hear them
Click Save Audio
Click Save Draft to save your changes, and click Publish if you want to publish them
Adjust Pronunciation of Specific Words or Phrases Across All Tour Content
If you have identified a word or phrase that has a unique pronunciation that will need to be manually re-generated across all generated audio, create a pronunciation record in your CMS.
Frequently Asked Questions ❓
I'm trying to generate audio for words that have unique spelling or pronunciation. What should I do?
Try spelling the word phonetically: by sounding it out, and spelling it the way it sounds rather than the way it should be spelled.
You can also try breaking down bigger words into smaller words that are easier for the AI to process.
For example: Alumni can be broken down into "A-lum-nie" or "Alumn eye"
Words that are originally from languages other than English may have to be broken down this way too.
Once you have determined the phonetic spelling to correct the AI pronunciation, create a pronunciation record in your CMS for future audio generation to use this spelling.
My title sounds out of place... I think I want my generated audio to say something different than what is written in my guide card text. What do I do?
You can opt to remove the title all together, if you think the rest of your text covers what is mentioned in your title succinctly.
If you want to keep your title, try finding a way to blend it in with your body paragraph so that it reads seamlessly.
For example: Your title could be something like "Sign up for our Newsletter!" and the body paragraph could be "Want to learn more about us? Click the link below." Change this in the Current guide card text box to say: "If you want to learn more about us, sign up for our newsletter by clicking the link below."
I don't like what syllable is being focused on. What should I do?
Accenting a different syllable of a word often helps to put emphasis on an alternate place in the word.
You can do this by separating the word into smaller words, and by making sure that the syllable you want the AI to focus on is isolated into its own word.
For example: "Neighbourhood" can become "Neigh boar hood"
Once you have determined the phonetic spelling to correct the AI emphasis on the specific syllable, create a pronunciation record in your CMS for future audio generation.
How do I get the AI to pronounce letters individually for abbreviations or initialisms?
The AI voice generator will often read abbreviations like CFL like "seeffell." If it is pronouncing that like a word, try adding a space or periods between each letter like "C F L" or "C.F.L."
A best practise for narrating acronyms is to change your text to one of the following:
"C F L, also known as Canadian Football League"
"Canadian Football League, or C F L for short"
If your acronym is pronounced as a word like LEED (Leadership in Energy and Environmental Design), try spelling it as "lead" for audio conversion.
Once you have determined the phonetic spelling, spacing, or punctuation to correct the AI emphasis on the specific syllable, create a pronunciation record in your CMS for future audio generation.
It seems to be reading the list too quickly, how do I get it to slow down?
Add a pause in the middle of a list with a comma or a period.
Take advantage of an oxford comma: it often does the trick!
It seems to ignore any text that is in brackets. What should I do?
Remove the brackets and replace them with commas.
The AI is designed to skip over (or ignore) converting text that is in brackets.
It doesn't seem to want to generate audio for my text at all. Why is this?
Audio generation is only possible to a maximum of 1000 characters. If your audio isn't generating, it could be that your text is too long.
Try to shorten your text to under 1000 characters. You can do this by removing the title, hyperlinks, or anything in brackets that you do not wish for it to read.
If that does not work, you may need to split the text into separate guide cards.
For example: separate the text from one guide card into two hotspots on your scene.
Generate and Regenerate Audio
Learning how to generate audio is the first step to mastering your generated audio guide. Generated audio can be easily updated, or regenerated, to reflect changes made to your virtual tour. Click the links below to get started! 🖱️
💡 Need more help?
Send us a message and we will be happy to assist you.