Renesas recognizes that voice recognition applications are growing rapidly worldwide, driven by the explosion of connected devices, particularly for consumer applications, boosted also by the COVID-19 pandemic. Cyberon’s proven voice-recognition technology has been adopted in millions of devices worldwide and Renesas decided to make a Voice Reference Hardware Platform available to enable rapid prototyping and development of voice interfaces.
Cyberon Corporation is headquartered in New Taipei City, Taiwan, and was established in 2000 to work in speech algorithm and application development. Its speech recognition and text-to-speech technologies have been widely adopted for all sorts of applications, from smart toys to automotive equipment, and enterprise customers. Today, Cyberon provides a full range of voice solutions for embedded MCU/DSP, OS platforms, and server-based services.
The new Renesas hardware platform enables local voice recognition without a network connection using Cyberon’s DSpotter solution. DSpotter’s phoneme-based modeling approach allows quick command customization with only text input, eliminating the need for massive voice data collection and thereby reducing development costs and time. DSpotter supports over 40 global languages, empowering brands and manufacturers to adopt voice technology on embedded applications worldwide.
As Renesas highlights, this combined Voice User Interface (VUI) solution demonstrates how easy it is to control a system with a simple voice command interface without extensive coding experience or in-house expertise.
"Customers in multiple segments are looking for help to quickly add reliable voice user commands to their products," says Mohammed Dogar, Vice President of Global Business Development and Ecosystem in Renesas’ IoT and Infrastructure Business Unit. "Our partnership with Cyberon gives them access to proven voice technology across our entire line of RA MCUs, from entry-line to full-featured devices."
The Renesas Voice User Interface reference solution, which includes complimentary access to the DSpotter Toolchain, allows Voice Activity Detection (VAD) to reduce power consumption by more than 30%, allowing power-constrained end-point applications to run in always-on mode. Voice Audio Playback (VAP) enables voice input and voice responses and the solution offers support for up to two digital microphones and two analog microphones, with external QSPI flash to provide the ability to store voice samples and library data. Multiple communications interfaces include a Pmod interface (type-2A/3A/6A), USB type-C and micro-B, and on-board j-link Segger debugger/programmer.
"We are honored to collaborate with Renesas to simplify the development of embedded voice recognition functions," adds Alex Liou, Vice President of Cyberon's Embedded Solution BU. "Cyberon's DSpotter algorithm meets the market's diverse requirements for both costs and performance. The combination of DSpotter and Renesas’ leading MCU technology enables customers to develop high-performance, reliable and cost-efficient products in a timely manner."
More information on Renesas’ VUI solutions, including a link for qualified companies and developers to request the new kit and the VUI solution package, is available at www.renesas.com/voice-recognition.
www.cyberon.com.tw
www.renesas.com
