Bring your AI Avatars to Life
Use the following tips to help with pronunciation and cadence, so your Avatars sound more natural and life-like.
You can sometimes fix word pronunciation by inserting hyphens or using their phonetic spelling. Here are a few examples:
Insert hyphens
Try inserting hyphens to make the word sound like you want. For example:
Content → con-tent
Contracter → con-tractor
Acronyms
Acronyms (or individual letters) may need to be spelled phonetically to get the right sound. Some acronyms are pronounced as a word (NASA), while others aren't (WHO).
Try spelling acronyms like they would sound. For example:
Examples |
|
Numbers
Try spelling numbers as words. For example:
Ten eighty-nine -> 10 89
Two five eight six -> 2 5 8 6
One hundred and forty-eight: -> 148
Use punctuation marks
If you have problems with sentence rhythm, try inserting commas, periods, or quotation marks to add different break styles in your script.
Commas will add shorter breaks than a period.
Periods will add a longer break and downward inflection.
"Quotes" will emphasize a specific word or part of the sentence.
Phonetic spellings
In most cases, you can adjust the pronunciation of a word by using phonetic spelling. Below is a handy table to help you replace letters with phonetic alternatives.
💡 TIP
When spelling phonetically, please ensure you use the Latin alphabet. Do not use phonetic symbols.
Letter | Phonetic Spelling | Alternatives |
A | ai | ah, ay, eight |
B | Bee | be, but |
C | See | kuh, suh, cee |
D | Dee | de |
E | Ee | eh, eei |
F | Ef | fuh |
G | Jee | guh |
H | eitch | huh |
I | eye | aiy, ah-ee |
J | Jay | jai |
K | kay | kai |
L | el | ell |
M | em | emm |
N | en | nuh |
O | o | ah, oh, uh, oo |
P | pee | puh |
Q | cue | kwuh |
R | ah-r | arr, are |
S | ess | zuh |
T | tee | tuh, tea |
U | yoo | yew |
V | vee | vea, vuh |
W | double-you | double-yew, Duboyoo |
X | eks | exx |
Y | wah-ee | wye, wyje |
Z | zee | zea, zuh |
Inflection of words
Sometimes, words are mispronounced and the above suggestions aren't helping...
Example: “Space the holes as evenly as you can”.
In this case, the AI voice is swallowing the S in space.
Solutions
Sentence Restructuring: Occasionally, restructuring a sentence can help the AI understand the context and pronunciation better.
In this case, restructuring the sentence to “Go ahead and space the holes as evenly as you can.” solves the issue. ✅
Choose a Different Voice: Using a different voice can fix pronunciation issues. Each voice has its own diction database and may handle certain words or phrases better.
Changing punctuation: Playing with punctuation can help with inflections. You can either use them directly in your script or use the diction function. See the table below for tips and tricks:
More Tips!
Pre-read your script out loud
Read your script out loud to see if it sounds like how someone would naturally speak.
Add word-fillers and/or pauses
To make your avatar sound more human, consider adding typical word-fillers such as “like…” “so,....” “well…” etc..
Adding pauses
Adjust the cadence of the script by inserting pauses using (copy these code blocks)
<break time="2s" /><break time="50ms" />Update the time as needed e.g., 3 seconds = 3s; 30 milliseconds = 30ms
Break up the text + be mindful of the length
If you have a long script, break it down into smaller paragraphs. Also, try breaking up long sentences into shorter sentences.
For video length, a best practice is a maximum of 150 words (~1 minute).
Stick to one language
Avoid mixing languages. For example, using English and Spanish in the same script.
Add Gestures
Make your avatar gesture by adding the following to the script (copy these code blocks):
Head nod:
<gesture label=\"nod\"/>Yes:
<gesture label=\"head yes\"/>No:
<gesture label=\"head no\"/>Raise eyebrows:
<gesture label=\"eyebrows up\"/>
