Conexant AudioSmart voice technologies enable GiGA Genie AI-infused virtual assistant media hub from KT to provide excellent near- and far-field performance for voice interface and video conferencing
IRVINE, Calif., February 21, 2017 – Conexant Systems, Inc., a leading provider of audio and voice technology solutions that enable a more natural user experience, today announced that KT, a leading Korean telecom corporation, is utilizing its AudioSmart™ CX20921 Far-Field Voice Input Processor to help the new GiGA Genie voice-controlled virtual assistant media hub to hear users’ requests accurately from a considerable distance away from the device, as well as in environments with ambient noise. Conexant’s AudioSmart Far-Field Voice Input Processor offers advanced voice processing techniques that filter out ambient noise, which provides KT’s artificial intelligence service with more intelligible speech requests. This processed speech boosts KT’s AI service’s speech recognition and natural language interpretation accuracy, as well as provides improved voice intelligibility for users during video conference calls.
Conexant AudioSmart technology was developed to revolutionize how CE manufacturers design voice-enabled products, allowing them to quickly bring to market new applications that boast exceptional voice interface performance and far-field capabilities.
“GiGA Genie is a first-of-its-kind product that combines the convenience of having an advanced AI-infused personal assistant, media hub and video conferencing system, all in one device, offering customers an interactive voice experience unlike anything else,” said Gyutae Baek, SVP, Service Lab. “Because of this, we wanted to ensure that GiGA Genie delivers excellent speech recognition accuracy, regardless of the distance the user may be from it. To accomplish this, we worked with Conexant to utilize their two-microphone voice input processing solution, allowing us to achieve the unique form factor of GiGA Genie, while also providing excellent near- and far-field voice performance, and impressive robustness to noise.”
Getting new voice enabled applications to work at peak performance in all kinds of environments poses significant engineering challenges related to echo cancellation, background noise, microphone and speaker positioning, user proximity and more. Conexant AudioSmart solution for voice applications separates speech signals from background noise and processes only the speech signal for improved speech recognition engine accuracy. This provides a better end-user experience for artificial intelligence platforms, even in noisy, real-world conditions, and/or when users are several meters away from the device. Conexant’s voice solutions can also improve the performance of teleconferencing applications by utilizing different voice processing techniques optimized for how people hear speech. These specialized tunings provide a crystal-clear telephony experience on both the near- and far-end of a call. Conexant’s Far-Field Voice Input Processor was designed to significantly reduce the engineering time required for OEMs to develop powerful new voice enabled products, like GiGA Genie, that offer an exceptional user experience.
“CES 2017 serves as proof that the voice revolution is at full stride, with voice as an interface being one of the hottest trends in tech this year,” said Saleel Awsare, President of Conexant. “With so many new voice-enabled products featuring Conexant technologies coming to market, our AudioSmart voice input processing solutions have seemingly become the industry’s de facto standard for voice performance. We are pleased to be working with companies like KT, that understand the value associated with voice and speech, and designing their products with Conexant’s industry-leading voice input solutions to offer an unparalleled end user experience.”