Fun with Voice AI
I came across this website, 11ElevenLabs, and haven’t had time to tell you about it until now because I also wanted to demonstrate it. The company has a lot of suggestions for how to professionally use their software, but I’m attaching a video where I demonstrate how to use it to just experiment and also for learning to control intonation.
I’ve only created a free account and that’s all I’m suggesting for you to do and you can also hear how different artificial voices sound with the sample text they provide or whatever you want to type into the box on the homepage without creating an account.
After creating my own voice sample on the Voice Labs page using the “cloning” option, I can use it to have the computer use my voice to speak any of the text I type into the box on the Voice Synthesis page. That’s where I start with the demonstration.
To find the options that I’m demonstrating on the website, open the “Voice Settings” box to change the settings for “Stability.”
I recorded examples at different levels of stability from 100% variable to 0% variable (“more stable”).
I downloaded the recordings and converted them from mp3’s to wav’s so I could upload them to the WASP program to see what the pitch changes looked like. That’s what I’m demonstrating in the video:
Be Your Own Example
I think imitation of speakers that are examples of speaking styles that you like is a great way to practice, not just for American English but for learners of any language.
I have always had the idea that if people could modify several options that would change the pitch range, pace, intonation, vowel duration, basically all the aspects of prosody, of their own voice, they could create a model of how they would sound differently but still hear their own voice qualities.
The ability to hear differences with your own voice sample with the 11ElevenLabs voice AI is a step in that direction. Their website lists examples of how to apply their technology to things like audiobooks and voiceovers, but I see a useful application to teaching and learning speech skills.
I see it as a way to be your own example of what you could sound like if you wanted to. If you like the way the computer generated examples sound and you practice controlling your voice to imitate that speech, you can achieve it. It’s like learning from your future self, in your future voice.
Even if you don’t think it sounds completely like your own voice, just experimenting by changing the one option of “Stability” is a good way to practice your listening skills to notice the changes and imitating it is a good way to practice controlling your voice to do something different.
Happy Anniversary Adastra Speech! Congratulations, Dr. Christi! In these 8 years you have improved people's lives all over the world with your personal tutoring, meet-ups,and newsletters!