LONDON – STMicroelectronics NV has introduced a digital audio processor IC that is intended for use in multi-microphone applications.
Multiple microphones can be used to locate sound sources in space and to enhance noise cancellation methods. With silicon-membrane MEMS microphones it is possible to produce arrays of small microphones or strips containing many silicon microphone die to improve the sound quality of mobile phones, tablet computers, gaming devices and video security systems.
The Smart Voice audio processors enable features such as acoustic echo cancellation, noise suppression, beam-forming. Multi-microphone systems also address next-generation audio applications, such as 3-D sound processing, sound-source localization, virtual microphones, and audio zooming.
ST’s newest digital audio processor is based on an acoustic processing engine that offloads intensive computing tasks from the main application processor. It handles up to 6 digital microphone inputs with dynamic array re-configuration and dedicated channel processing. The device integrates a high-quality scalable acoustic processing core with a tunable 10-band equalizer, a peak limiter, and gain and volume controls.
The Smart Voice processors come bundled with AP Workbench, a programming tool.
The STA321MPL, the first Smart Voice processor, is sampling with mass production scheduled for 2Q12. The unit pricing is $4 per unit for amounts of about 1,000 pieces.
EE Times Confidential is working on a MEMS Sector Profile and Database report and if you would be interested in being advised when copies are available, please send an email to email@example.com
As MEMS microphones get cheaper, it should be increasingly common to see microphone arrays in smartphones and tablets. There are already some Bluetooth headsets offering microphone arrays for noise cancellation.
I guess the large scale use of mic arrays was pioneered for aircraft detection until radar arrived, but later progressed for Cold-War espionage 'bugging'.
This is a great application for multiple processors cooperating in real-time, you could absorb as much processing power as you can buy. The trick is to achive the processing with minimal group delay, whih means processing in parallel.
Hearing aids also stand to benefit, as mic directionality is very helpful for discriminating voices in a crowd. Coupling this with assist from a camera could greatly
benefit hearing-aid users (and spooks of course).
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.