Updated: Feb 24
Founded by three former Stanford students in 2020, Sanas is an American technology start-up that aims to provide AI-powered solutions for real-time accent translation. With no noticeable lag and edge deployment, the patent-pending Sanas software intercepts audio and converts accents through a speech-to-speech approach, building a virtual bridge between the audio device and the computer, and then sending the new signal to communication application, like Zoom or Hangouts, according to the vendor. Internal studies from Sanas showed increased foreign language learner fluency, as well as decreased word error rate (15% on industry-leading automatic speech recognition devices). For the current application, the solution is tested by call centres who previously needed to provide costly and time-consuming accent reduction training for employees. Sanas' algorithm can convert English to and from American, Australian, British, Filipino, Indian and Spanish accents, and the team is planning to add more. They can add a new accent to the system by training a neural network with audio recordings from professional actors and other data, a process that takes several weeks. The American-accented results sound somewhat artificial and stilted, like virtual-assistant voices such as Siri and Alexa, and the team is working on improving the technology. With positive feedback from call centres, Sanas aims to foster seamless communication in all areas of business, education, telemedicine, entertainment and more.