Audio Analytics writes:
Enabled devices can then intelligently respond through alerting consumers or triggering a wide range of actions from connected smart home products. The home’s personal assistant can either disturb the uninvited guest or let the owner know that family are home.
This new sound profile complements our suite of security and safety sound profiles (glass break, dog bark, smoke alarm, CO alarm and anomaly detection) to provide consumers with even greater peace of mind when it comes to unattended homes. For the consumer tech brands that embed our software into their products, they can add unique, intelligent benefits that enhance value propositions and retailer ranging decisions.
“A major selling point for smart home cameras is their ability to detect people but these products have limitations, such as the field of view. Our new speech detection capability enables leading tech brands to offer 360-degree coverage on their products. Intelligent sound recognition enables device manufacturers to tackle these limitations, enhance value and offer consumers the ability to always be in, even when they’re out,” said Audio Analytic founder and CEO Dr Chris Mitchell.
Different to common voice activity detection (VAD), our speech detection sound profile delivers a higher level of accuracy for home security solutions, performing well beyond off-the-shelf VAD options. In addition, the compact size of our sound profiles means that our customers can deploy new features and benefits over-the-air to devices already in the market. For more information, visit our software page.
Security and safety are two applications driving the fast-growing smart home market. We forecast that around 14m cameras will be sold globally in 2018 (including video doorbells, inside and outside cameras). By 2023 we expect 65m cameras will be sold, with the majority of devices offering intelligent sound recognition.
While cameras and doorbells are a rapidly growing market, the opportunity for enhancing consumer security is not limited to these products. We expect smart speakers, hubs, TV set-top boxes, alarms, and even light bulbs to start taking more of an active role in smart home security.
By training our software to delineate human speech from other sounds we ensure that our customers’ products don’t indiscriminately alert users to just any noise, but offer reliable, actionable information based on real-world sounds. This is because consumers have a low tolerance for false alerts and successful products can’t afford to set expectations that they can’t live up to.