Pandemic Shifts Voice UI from Feature to Necessity

Article By : Ofer Elyakim

In the wake of Covid-19, voice, as a user interface, will become a significant factor in architectural design of many consumer products.

There are moments in history that even as they unfold, it is clear that nothing will ever be the same again. The Great Depression forever changed approaches to personal financial security, and that fateful blue-skied Tuesday morning of 9/11 forever changed how nations approached national security. Even now we know that the Covid-19 pandemic has changed how we view health security.

With heightened awareness of how germs in general — not just the Covid-19 virus — move from host to host, how long will it take us to feel comfortable among tightly packed crowds of strangers? Will we ever be OK using high-touch kiosk or register checkout screens, elevator buttons, and door handles, especially when technology has already been primed to offer alternatives?

For now, it looks like mandatory shelter-in-place rules have worked to slow the spread of the virus and “flatten the curve” in many areas, but as economies open and people emerge, much of what we have learned and the habits we have developed, may affect our activities and actions, long term. For example, the Centers for Disease Control and Prevention in the U.S., and similar bodies across the globe, recommended we avoid touching surfaces in public places that have traditionally been high-touch, including the aforementioned elevator buttons, as well as light switches, and trash cans.

Likewise, at home, it’s smart to more frequently clean surfaces of shared devices, including remote controls, switches, alarm keypads, door handles, smartphones, and tablets, as well as light switches, and oven controls. Many of these devices are already using voice-user interfaces (VUIs), and the list of applications is growing (see below). This trend will only accelerate in the wake of the pandemic.

Voice as a user interface (UI) had already been gathering momentum in applications that include smart speakers, smart TVs, and light switches, but a reluctance to touch public surfaces has put VUI on trajectory to intersect with a global trend toward personal health security. (Source: DSP Group)

Analyses of how novel coronaviruses like Covid-19 spread have made the global population hyperaware of just how contaminated surfaces can be. Now, armed with comparative knowledge about the viability of the Covid-19 on glass and stainless steel (measured in days) versus cardboard (measured in hours), as well as the morbidity rate of the pandemic, we’ve all stepped up our personal hygiene habits. So much so, that there will inevitably be a long-term, underlying reluctance to touch any public surface. For many, it may be wise to be a bit obsessive, for others, finding unique ways to avoid touching public surfaces will simply be a new habit, formed over many weeks of brief excursions to stores, and reinforced by the nature compulsion to protect themselves and their loved ones.

These are wise habits to nurture, but as it turns out, technology, in the form of the voice as a user interface (UI), was already on a trajectory to intersect with this new approach to interacting with devices, from mobile phones and Internet of things (IoT) devices, to doorways and security systems.

The quiet pervasiveness of VUIs

Voice-based command and control using Google Home, Amazon’s Alexa, and Apple’s Siri, have always seemed like a “cool” and useful feature to interact with handsets, laptops, and computers, as well as to control everything from TVs, smart speakers, cameras, lights, window blinds, and home HVAC systems. Supported by accelerating advances in keyword recognition, always-on capability, far-field voice control using beamforming with multiple microphones, ultra-low power consumption for wearable technologies such as true wireless (TWS) headsets, and more recently, multiple simultaneous keyword recognition, VUIs have become more seamlessly integrated into both the home and work life, from the able-bodied to the infirm, from the Luddite to the tech-savvy.

Clearly, VUIs have become more useful than cool, but over the past few weeks, something has changed. By their nature, they greatly reduce the need to touch surfaces, this means that in the wake of Covid-19, they’ve evolved from a feature to a necessity. Applications are many in the home, but in public spaces, VUIs can open doors and trash cans, complete point-of-service (POS) transactions, and even assist in elevators to select a floor without compromising personal hygiene, a particularly important issue in hospitals.

As VUIs become more pervasive, the algorithms behind their intelligence will become more refined and the underlying hardware will become more advanced and efficient, further extending the application possibilities and seamlessness of VUIs in a new world where limited touching has become a cultural norm.

Approaches to VUI implementation
Advanced algorithms and efficient hardware are required for rapid, low-power buffering, filtering, and processing of voice and possibly other sensor inputs, such as cameras, for user authentication and context awareness for intent verification (e.g., a user walking toward a specific door). In many cases, this processing has been done in the cloud, but in the drive to reduce power, lower latency, and ensure end-user privacy, as much of the processing as possible needs to be done the edge.

This cloud-edge split is one high-level architectural approach, but even at the edge, the Always On capability and voice recognition and processing functions can be split further. Take a power-sensitive remote control (RC), for example. To date, these have relied on push-to-talk (PTT) instead of Always On listening in order to conserve power. However, Always On capability is still the end game, so system architects have come up with a way to subdivide the functions further. Specifically, a single microphone in the RC is used to detect a voice and, if a keyword is valid, the RC can work with a TV or set-top box mic and processor to further process the incoming commands, leveraging advances in natural language (NL) voice recognition for smoother interactions (see below).

An innovative architectural approach to VUIs at the edge splits the processing of voice between the remote control and the device being controlled, whether it is a TV or air conditioner. This greatly reduces power consumption and allows RCs to move from push-to-talk, to Always On capability. (Source: DSP Group)

There are other advances in and around VUI, such as active noise cancellation (ANC), low-power MEMS microphones, and the integration of wireless technologies such as Bluetooth and ultra low energy (ULE), the latter enabling reliable two-way voice communications. While these advances are important, they support an emerging but fundamental change in how we interact with devices and systems in the post–Covid-19 world. That VUIs happen to be on the rise at the time of its emergence is pure chance, but it doesn’t change the realization that now’s a good time to think about how to go about integrating them into next-generation designs such that they’re intuitive and friction-free. Let’s work to do it right. Users have been traumatized enough.

— Ofer Elyakim is CEO of DSP Group.

Leave a comment