Audyssey eVR: Enhanced Far Field Voice Recognition

N/A (technology is licensed by hardware manufacturers and embedded in products later sold to integrators and consumers)

Audyssey eVR is a suite of audio technologies that enables far field voice recognition in CE products that rely on voice-based user interaction. Manufacturers license eVR to optimize far field voice recognition in their products, thus delivering a more enjoyable customer experience.

Audyssey eVR enables far field voice recognition in hardware products that rely on voice-based user interaction. The far field improvement that eVR provides can dramatically extend the effective listening distance of voice-controlled products and allow them to operate in noise-challenged environments.

Manufacturers license and embed Audyssey eVR’s suite of technologies to optimize far field voice recognition in their products, thus delivering a more enjoyable customer experience. Contemporary examples include thermostats like Honeywell’s WiFi Smart Thermostat, home hubs like Amazon’s Echo and robots like Softbank Robotics’ Pepper. But these mainstream examples are just the tip of the iceberg, as voice recognition is poised to permeate many other segments of the consumer electronics market.

Audyssey eVR is broken into three basic components – capture, clean-up and hot word detection (including barge-in when the hot word must be recognized during other content playback from the device). Although voice recognition engines typically reside in the cloud, these front-end functions must be performed quickly, and therefore they must reside in the processor within the manufacturer’s product. Audyssey developed highly optimized code for these functions so that the required processor footprint is low and can easily be accommodated in typical processors used today. Updates and improvements can easily be delivered because all of these products are connected to the internet.

Audyssey conducted extensive testing to quantify how Audyssey eVR extends the effective listening distance and improves barge-in performance:

• At distances of two, four and six meters in a reverberant room simulating different far field conditions, eVR outperformed existing devices by 45, 42 and 53 percent increases in word recognition, respectively.
• Barge-in performance testing hot word recognition revealed that eVR delivered 59-82% improvement in recognition, even when the speaker is playing music at high levels.

Approximately six billion products with microphones and voice interaction will reach consumers by the year 2020. Input voice quality will be a key enabler of this new user experience. With the first wave of voice-controlled hardware products now in consumers’ homes, it is increasingly clear that, while these products respond well when the user is within close proximity of the device, the user experience deteriorates rapidly when interacting with the device from greater distances. This “far field communication” is the Achilles heel of voice-controlled hardware devices.

Audyssey eVR solves the problem of input voice quality in voice-controlled hardware products. Leveraging the company’s 10+ years of experience with room acoustics, Audyssey researchers invested two additional years of R&D characterizing the problems of far field voice capture and developing new algorithms to capture and clean up voice signals so they match what the various voice recognition engines in the cloud are expecting to hear. These devices must also be alerted to start listening to user commands. Audyssey developed a unique method to create customizable “hot words” that can be used to wake up devices even when they are playing loud music.

With all this advanced processing running on the input voice signal, the benefit to homeowners is quite simple: eVR-enabled products “simply work.” When someone invests in new technology for the home, nothing is more frustrating than when those products don’t function as advertised. Voice control is the critical link to the smart devices of tomorrow, and eVR makes it possible for hardware manufacturers to make voice-controlled products that respond well in everyday household settings – even those with particularly challenging acoustical characteristics.

Installers are being increasingly called upon to implement “smart” solutions. Internet-connected devices that make up the so-called “internet of things” present great opportunities for integrators charged with installing smart solutions in homes, corporate environments and even public settings.

Voice recognition is an increasingly popular method for users to interact with these devices, yet significant barriers exist to voice recognition’s mass adoption. The most notable obstacle is how poorly voice recognition products perform in far field settings. Audyssey eVR makes it possible for hardware companies to eliminate this obstacle and deliver the performance customers expect in any room or listening environment.

The root of this problem is that hardware innovation is outpacing software engineering, consequently many voice-controlled products are falling short of what manufacturers promise their consumers. This presents a unique challenge to integrators, as it becomes difficult for them to piece together smart solutions that offer seamless voice-activated interaction. But now integrators can specify Audyssey eVR-enabled devices for their installations, secure in the knowledge that those voice-controlled devices will perform as promised in even the most acoustically challenging environments.