QNX Acoustics for Voice 3.0 targets high-end hands-free equipment and voice recognition systems, offering improved echo cancellation and noise reduction, adaptive equalisation, and automatic volume control. In addition, the software supports the Wideband Plus voice recognition standard and thus meets the requirements regarding sound quality as defined by the latest smartphone connectivity protocols for telephony and VoIP services. According to QNX, the software has been developed with the intention to filter out noise generated by the vehicle's wheels, engine, air conditioning and similar sources.
The software features an improved voice recognition algorithm for Wideband Plus that causes less CPU load. As a result, more CPU power is available for other acoustic and infotainment applications. A comprehensive diagnosis tool set facilitates the fine-tuning of the parameters and thus enables users to reduce the time-to-market.
Voice recognition can be handled quite flexible: Depending on the user's requirement, this function can be implemented in the cloud or embedded in the car. It supports a broad range of operating systems, processors and DSPs, enabling customers to migrate their implementations across model families.
The standard library supports narrow-band as well as broadband voice recognition with sampling rate of 8 and 16kHz. The premium library, available as an option, adds the support for the larger frequency range of Wideband Plus with 24kHz sampling rate. Thus, it is possible to connect the latest smartphone generation.
The software has been introduced at the Detroit Telematics Conference; it will be available in Q3 2014.
Related links and articles: