Our main product is an AI voice-over service called Typecast. All you need to do is type the scripts, and you can cast different voice actors by clicking characters and listening to their voice samples. In a world where CG replaces a real human, people are still striving to find the best way to produce a real human-like voice. Although there were voice compositing technologies before, controlling diverse vocalizations was difficult. We’ve created a technology that can recognize the surrounding situation and conversation content, and control emotions, which are applied to Typecast to control the timing of reading and pausing rhythms and even sophisticated emotion.
In terms of technology, we provide hyper-realistic speech just like a professional voice actor’s recording, and method to control emotion and prosody for the speech. Furthermore, we plan to expand to a service that generates a video of AI actors by just typing a script. In the traditional virtual avatar technology, a video is created by recording an actual performer’s voice and capturing his/her motion, which significantly limits efficiency and scalability in creating content. We’ve developed realistic human-sounding speech synthesis, which converts text to speech.