This blog post was originally published at Synaptics’ website. It is reprinted here with the permission of Synaptics.
Remote controls have tended to be practical, utilitarian devices with a well-deserved reputation for being the hardest to find when needed most. It doesn’t have to be this way in an era when users are using voice to control devices at home, on the go, in the car, or at work. Enabled by always-on voice (AOV) capability, remote controls can instead become a powerful command center to search, select and control the content, execute commands with minimal latency, and enable voice biometrics for purchase authentication and parental control seamlessly and intuitively.
While AOV holds much promise, it is challenging to implement in battery-powered devices, particularly when operating in a noisy environment, for example, with a TV, music, or multiple people talking. Noise increases power consumption as more processing is required to prevent performance degradation. However, with appropriate design and implementation, AOV can be enabled in battery-powered devices with a lifetime of over a year on just two AAA batteries.
Accomplishing this requires innovation in several areas. For battery life and performance, an ultra-low-power wake-word detector is needed, ideally with a dedicated neural network (NN) processor. This must be matched with proven algorithms for noise suppression, beamforming, and voice activity detection.
To reduce latency, adding local commands is critical to eliminate the delays associated with cloud-based approaches. Local commands also improves reliability, ensuring devices will respond without being affected by connectivity issues: users can trust that their TV or music player will “Mute” upon command should a phone call come in.
The addition of voice biometrics enables functions such as voice authentication for content purchases, parental controls, and content profiles for each user. And finally, an AOV-based “Find my remote” feature can be readily added.
Synaptics’ remote control solution, based on the DBM10L low-power edge AI SoC, brings together all the elements needed for an optimal AOV product design, while its combination of high-performance, low-latency, reliable AOV, and value-added features ensure an exceptional end-user experience. It also meets service providers’ and system developers’ ongoing push for ever smaller and smarter devices and greater customer attachment.
Find out more about how to implement AOV that addresses these requirements on the DBM10L low-power processor that combines an NN core, a DSP engine, and abundant memory by emailing [email protected] to book a time slot in our IBC Balcony Suite BS17.
Shay Kamin Braun
Director of Product Marketing, Low Power AI, Synaptics
