Text to Speech
Quality voice output, closely matched to its audience, can make a big difference to the ease and naturalness of conversation with our AI, especially where an avatar is present. From a big range of the highest-quality voices, in many languages and from multiple vendors, we find the optimal voice for each project.
Evie on mobile uses advanced speech recognition also known as speech to text, is a quickly-improving art. It cannot be used in all circumstances, and can still make mistakes. Where it does, we simply learn and predict those errors. Thanks to context, what the user said is understood.
Our software integrates, and controls the voices, introducing emotional expression wherever possible, and creating lip synch information for display - even while avatars themselves express emotion. Voices are delivered seamlessly to browsers as mp3 files in real time, with no client download required.