Cyberon Optimizes Voice Recognition Software for Cadence Processors
Cyberon has optimized its CSpotter voice recognition software for the Cadence Tensilica Fusion and HiFi digital signal processors (DSPs). CSpotter is a voice recognition engine that supports low-power always listening with keyword spotting.
Cyberon's CSpotter software is specifically designed for always-on mobile and Internet-connected devices, listening to ambient speech to detect and respond to wake words and/or trigger commands. CSpotter supports 32 languages.
The Tensilica HiFi DSP audio/voice/speech DSP family supports more than 150 proven audio/voice software packages and more than 70 software partners.
The Tensilica Fusion DSP, announced in April, is geared toward applications requiring merged controller plus DSP computation, ultra-low energy, and a small footprint.
"Our goal is to make it quick and easy to add voice commands to a wide range of applications, from smartphones, smart toys, wearables, automotive, to IoT," said Tai-Hsuan Ho, founder and CEO of Cyberon, in a statement. "We feel three factors are very important for our success. The first is that we are running on the most popular platforms, which is why we picked the Tensilica DSPs. Secondly, CSpotter is one of the most accurate, noise-robust ,and small-footprint voice recognition engines used as a wake word trigger in a voice-activation application. Finally, we offer multilanguage flexibility with support for 32 languages, enabling developers to create voice-enabled devices that can be shipped around the world."
"The Cyberon software is a valuable addition to our huge audio/voice/speech software partner ecosystem, offering advanced features including the ability for users to create their own personal and speaker-dependent trigger words," said Larry Przywara, group director of audio/voice IP marketing at Cadence, in a statement.
DSP's Audio Weaver development environment has been optimized for Cadence's Tensilica HiFi Audiuo/Voice processors.