Building a Reliable Home Hub: Insights from Sensory’s Latest Guest Webinar

Sensory’s Jeff Rogers, VP of Sales and Marketing, recently presented a webinar with the Audio Product Development Alliance (APDA) on the topic of building reliable home hubs using local speech technologies. Check out the highlights below, or watch the full webinar here! 

Impact of Major Players in the Speech Technology Market

Two significant events drove the adoption of speech technologies: the launch of Apple’s Siri and Amazon’s Echo. Siri, despite its initial limitations, popularized the concept of voice assistants, while the Echo normalized the idea of having always-on, listening devices in homes. These developments led to an explosion in voice technology adoption across various industries.

Challenges with Cloud-Based Solutions

Jeff addressed the challenges associated with cloud-based speech solutions, particularly due to recent changes in the cloud ecosystem. Amazon’s decision to restrict third-party access to Alexa and Google’s scaled-back support and requirement for specific hardware has left many companies seeking alternative solutions. Additionally, the cost implications of cloud-based services, such as per-minute usage charges, can quickly become expensive for businesses with lots of users. These changes have been a source of frustration in the market, and have opened the door for local speech recognition solutions like those offered by Sensory.

The Case for On-Device Solutions

Jeff emphasized the benefits of on-device speech recognition, including:

• Lower Cost of Ownership: No cloud infrastructure is required, reducing operational costs.
• Complete Control: Auto OEMs have full control over the technology without relying on third-party services that can change their plans.
• Always-On Availability: Devices remain functional even without an internet connection.
• Complete Data Privacy: All data processing occurs locally, eliminating concerns about data privacy and false triggers common in cloud-based systems.
• Most Flexible & Efficient AI Performance: Supports CPU and deeply embedded architectures, with the option for hybrid approaches combining local and cloud AI if desired. 

Real-World Challenges with Smart Devices

Jeff shared his personal experience and frustrations of setting up smart home devices, highlighting the complexity and frequent disruptions that are all too common. This scenario emphasizes the need for simpler, more reliable connected home solutions without cloud connectivity. Traditional setup processes for cloud-based solutions require multiple steps, including purchasing the product, installing the app, connecting to Wi-Fi, and enabling the skill or device, which can be a cumbersome process and prone to errors. 

Building a Home Hub with Local Speech Technologies

When it comes to building a home hub, leveraging local speech technologies can significantly enhance the user experience and address common pain points associated with smart home devices.

Here are some key features and benefits of a Sensory-powered home hub:

• Custom Wake Word: Users can define their own wake words (hotwords, trigger words, wake up words) or use enrolled fixed wake words, allowing for personalized and brand-specific interactions
• Natural Language Control: The home hub can control various satellite products such as lights, plugs, fans, air conditioners, blinds, and appliances using natural language commands through TrulyNatural STT (Speech-to-Text) technology
• Communication Protocols: The home hub supports multiple communication protocols including Bluetooth, Wi-Fi, Matter, and ZigBee, ensuring seamless integration with a wide range of devices
• No Reliance on Cloud Partners: The system operates independently of cloud services, eliminating the risk of changes in pricing, support, or availability from cloud partners
• Sound ID and Custom Sounds: The home hub can be equipped with Sound ID, which can detect standard sounds like alarms, glass breaks, doorbells, dogs barking, babies crying, and more. It also supports custom, user-defined sounds
• Language Models and NLU: Sensory has developed custom home control language models that interpret user intentions and convert them into specific functions. These models are designed to be ultra-low power and efficient, even in environments with limited resources

Sensory also offers options to customize your Home Hub with custom hotwords, speaker verification and Speaker ID, and Sound ID. 

On-Device Speech Recognition is Highly Beneficial for Businesses and Consumers Alike

By leveraging Sensory’s technology, companies can ensure data privacy, reduced costs, and a more reliable user experience. As the landscape of speech technologies continues to evolve, Sensory is at the forefront, offering innovative solutions that meet the growing demand for local, on-device processing.

Watch the full webinar recording, or contact us to learn how Sensory can help your business implement on-device speech recognition.