Looking to address the still-chronic problem of voice quality and voice recognition accuracy in today’s smartphone and tablets, Cirrus Logic has just released a new line of voice processors for mobile applications.
According to Carl Alberty, director of marketing for Audio, Cirrus Logic, the ultra-low power Cirrus Logic CS48LV12/13 voice processors with the company’s integrated SoundClear technology are designed to elevate the user experience and improve voice recognition accuracy in smartphones, tablets and wearables by improving voice quality, eliminating background noise and delivering clear communications in any environment.
Included in the offering, he said, is technology for single/dual microphone noise suppression and acoustic echo cancelation for both narrow and wideband audio content, providing manufacturers with a complete hardware and embedded software product capable of delivering unmatched consumer experiences with voice recognition applications and phone conversations in handset or speakerphone mode.
“Speakerphone use cases for both productivity and entertainment are becoming increasingly important,” said Alberty. “Voice-enabled interface features will continue to fuel growth in mobile devices, yet will require superior signal processing technology to give consumers an exceptional user experience.”
To facilitate faster time-to-market for mobile applications with fundamental voice processing requirements, the CS48LV12 offers a fixed function, turnkey solution with the latest SoundClear voice techniques for noise suppression and echo cancelation.
The configurable CS48LV13 adds Cirrus Logic’s SoundClear ASR Enhance dual-microphone noise reduction technology to improve the reliability and accuracy of automatic speech recognition (ASR) engines operating in harsh, noisy environments.
The SoundClear processing engine makes it possible, said Alberty, to offer high definition voice/wideband and narrow-band support with a wide range of capabilities including: Tx and Rx noise elimination, AEC plus non-linear residual echo suppression for superior full-duplex echo-free speakerphone operation, ambient aware parametric EQ for both Tx and Rx voice streams, automatic level control for both Tx and Rx voice streams, automatic volume control for Rx output, spectrally matched comfort noise generator for Tx stream, and comfort noise generator for the Rx stream.
To enable “always on” voice wakeup of mobile devices, the CS48LV13 supports a low power implementation of Sensory Inc.’s TrulyHandsfree technology. By combining TrulyHandsfree technology with SoundClear ASR Enhance, as well as ultra-low power voice detection and optimized hardware and intelligent system control, the CS48LV13 provides developers with a means by which to use voice as an interface to wake up a device and the ability to always be active and on, without eroding the mobile device user’s experience due to observed battery life reduction.
Implemented and optimized on what the company claims is the industry’s smallest, lowest-power audio DSP, the 32-bit CS48LV12/13 is complemented with advanced multimedia software to enable a variety of audio and voice enhancements. Included with the offering is the SoundClear RAPID2 (Real-time Audio Perfection and Integration Diagnostics) PC-based acoustic tuning tool that allows developers to easily monitor and tune programmable parameters in real-time via a simple Graphical User Interface (GUI). It includes an intuitive block diagram view, real-time control and diagnostics, and system-level measurements and statistics.