Absof*ckinglutely!
Make you cringe? Love to say it? Either way, this is just one example of "expletive infixation," which is a fairly common speech pattern where you stick a curse word in the middle of another word.
But did you know, most TTS models really struggle to pronounce words like this, or any compound words really? The most common mistake—if the model can pronounce the word at all—is stressing the wrong syllable. The name for this is "misaccentuation," and when you hear it, something just sounds off.
But humans don't have the same struggle. While expletive infixation and compound words break the by-the-book rules of English, almost everyone intuitively knows the right way to say them. Real speech is more than a list of strict rules—it’s messy, expressive, and full of creativity.
Rime gets this. We train our models on proprietary in-studio data from regular people in addition to voice actors, so our voices just sound natural and can handle words that trip up other models.
We also give you tools to easily fine-tune pronunciation, so you'll never mispronounce brand names, invented words, or whatever else comes up. We're the only next-gen TTS model to do this. Pretty fricking useful, right?
Ok so let's go back for a second. Now how would you describe the word "absofrickinglutley"? Well let's break it down, starting with the word "frick." When you swap a similar sounding word for actual expletive, it's called a "minced oath" or "euphemistic phonetic alteration." So combining terms, you'd get a "euphemistic phonetic alteration infixation." What a mouthful!
Head spinning? Just be honest—what’s your favorite cheesy exclamation?!