Charisma AI: How to tell powerful stories using Synthetic Voices
Recall your favourite movie. The one with the intense dialogues and character development that makes you jump out of the seat and leave you feeling an emotional connection to the personas that developed right in front of your eyes. Writers who deliver brilliant stories give meaning to the words by weaving them together to create these sorts of experiences and then find the right voice to bring those words to life. Iconic characters like Mickey Mouse, Yoda, Rocky Balboa, Bruce Wayne, or the Joker are ingrained in our heads. We know what they sound like without re-watching them on the big screen.
Often, we find ourselves relating to the characters and personas that we watch, and wish that we had the ability to interact with them more than just having a one-sided conversation where they just speak to us. This is where Guy Gadney, CEO and Founder of Charisma AI comes in. We asked Guy a few questions about storytelling and how Resemble AI helped build an experience that gave character development a whole new meaning.
What does Charisma do?
Charisma.ai is an interactive storytelling engine that using Artificial Intelligence to bring characters to life, enabling audiences to talk to them using natural language, and feel totally immersive in the storyworld.
We’ve been pioneering this for a while now, and recently launched projects with the BBC and Sky in the UK, and globally with our interactive graphic novel app, Charisma.
What makes a good interactive story?
The magic occurs when the user’s agency in the story is balanced correctly with the tension and pace of the story itself. Achieving this creates proper emotional engagement between the audience and the story, and plays to the strengths of new digital platforms.
What’s the most important shift you see in storytelling?
The most important shift is actually not visible. It’s not HD, 3D, VR or even 5G. The most important shift is from linear to interactive. This is what creates the immersion, but also requires a re-think in how stories themselves are written. We have centuries of experience writing linear stories for books, television and film, but interactive storytelling requires writers to go back to the DNA of storytelling and build it back up with the recognition that there is a new character present called the audience.
Where does synthetic voice fit into the arc?
To have true interactivity, a significant part of the story needs to be created dynamically. Whether it’s mentioning the player’s name in a sentence, requoting the player, or bringing dynamic data like football scores into the story, these are media elements that cannot be pre-rendered. A good synthetic voice solves this problem, allowing true personal storytelling, and maintaining immersion.
How has Resemble helped you?
Synthetic voices are not easy to get right. It is tempting to think that a top quality voice can be created by speaking a bunch of sentences into a microphone, but I feel it’s closer to the video editing process, where a base voice is captured, and then optimized, mixed and remixed until it’s absolutely fit for purpose. Resemble helped us realize this, and then worked hard to make our serial killer’s voice resonate in exactly the way we wanted him to. Automation is a great thing, but human creativity is always needed!
What have you been most impressed with?
Our characters are all about emotional performance. Siri, Alexa and other customer call centre voices are monotonous, but Charisma characters come to life, get happy, sad, angry. Resemble’s capabilities in this regard are awesome and their markup language gave us the flexibility we needed to achieve our goals. That said, I know that the Bulletproof project only hinted at the huge impact that AI will have on storytelling, and the ideas we have to push the limits of synthetic voices in entertainment productions are ambitious and game-changing. It’s certainly an exciting time to be forging these bridges between technology and creativity.