Audience Releases earSmart eS320 for Mobile Computing
Making its first foray into the PC space, Audience, a provider of advanced voice and audio processing for mobile devices, has launched earSmart eS320, the world's first advanced voice processor created specifically for tablets, ultrabooks, and hybrid all-in-one PCs.
The earSmart eS320 features advanced voice processing based on the science of human hearing, the company said, enabling computing devices to hear in a way similar to how humans do. Features such as 360 degrees voice, super wideband noise suppression, and PC-optimized Automatic Speech Recognition (ASR) Assist provide PC manufacturers with the ability to create the new convertible and ultrabook form factors for consumers.
The 360 degrees voice feature allows the user to move freely around his or her device while preserving voice quality. Unlike legacy solutions that require three or more microphones to achieve performance, Audience's positional robustness is delivered with just two microphones, based on the ability of the earSmart eS320 to automatically localize and enhance voice sources. As a result, manufacturers have greater flexibility with microphone placement, which can lower production costs and reduce limitations on a device's design.
"We have commercialized a field of research known as auditory neuroscience, and we're looking at mimicking true digital signal processing the way that humans hear sound using binaural processing," says Bart DeCanne, vice president, product management and marketing, at Audience. "By doing the same [thing] using digital processing and getting signals from multiple microphones and locating and classifying sounds between sounds and undesired sounds, we can focus on a particular [sound] source. This field of research is also called computational auditory scene analysis."
This technology, used by Audience in its earSmart eS320, which features Super Wideband noise suppression, enables users to make and take clear VoIP calls in almost any environment. In addition, acoustic echo cancellation and de-reverb allows users to drastically reduce double-talk and echo effects in larger, resonant spaces.
Another earSmart eS320 feature is ASR Assist technology, which addresses challenges faced by speech recognition applications such as voice search and speech to text. Audience's custom hardware-accelerated algorithms help isolate the speaker's voice from surrounding environmental noise, which can help with word recognition accuracy and task improvement.
DeCanne explained that with Audience's background with speech recognition, "we have some benefits over [other] companies. A lot of competitors just do noise suppression for voice calls. We have a lot of knowledge of the front end of the speech recognition engine." In addition, ASR Assist is also speech engine–agnostic.
"ASR Assist mode it works well whether you use Google, Bing or [IBM's] Watson speech engine," says DeCanne. "Whichever background speech engine you use we can demonstrate significant improvement performance when you use our ASR Assist."
Consumer devices equipped with the earSmart eS320 are expected to be available in the fourth quarter of 2013.
Related Articles
Dell's Vostro models to feature advanced voice processor.
06 Jun 2013
Companies have also partnered on the design of voice quality testing laboratories.
05 Jun 2013
Offering delivers HD voice and optimized speech recognition.
26 Feb 2013