ai3™ is capable of recognising a number of sounds
We are adding new sound profiles all the time, providing additional opportunities for customers to enhance products already in the market. The small file size of each means that customers can deploy additional sound profiles over the air.
Window glass break
Using a standard MEMs microphone from at least six meters away, ai3™ is capable of recognising the breakage of single and double glazed windows.
Target products include a wide range of smart home devices deployed for safety or security applications, such as (but not limited to) smart speakers, security cameras, smart lightbulbs, etc.
Dog bark
Using a typical MEMS microphone from at least six meters away, ai3™ is capable of recognising the bark of a dog due to frustration, distress or excitement, or to gain attention or threaten intruders.
Target products include a wide range of smart home devices deployed for safety, security or pet wellness applications, such as (but not limited to) smart speakers, security cameras, smart lightbulbs, driver assist, etc.
Smoke and CO alarm
Using a typical MEMS microphone from at least ten meters away, ai3™ is capable of recognising the loud tonal sounds emitted by smoke and CO alarm devices, irrespective of whether they are modulated or continuous tones.
Target products include a wide range of smart home devices deployed for safety or security applications, such as (but not limited to) smart speakers, security cameras, smart lightbulbs, etc.
Baby cry
Using a typical MEMS microphone from at least five meters away, ai3™ is capable of recognising the sound of a crying baby between the age of 0-12 months.
Target products include a wide range of smart home devices deployed for safety or wellness applications, such as (but not limited to) baby monitors, smart speakers, security cameras, smart lightbulbs, etc.
Anomaly detection
Using a typical MEMS microphone, ai3™ is capable of recognising changes in the ambient acoustic profile of a home, recognising when something happens that is out of the ordinary.
Target products include a wide range of smart home devices deployed for safety, security or wellness applications, such as (but not limited to) smart speakers, security cameras, smart lightbulbs, etc.
Want to see our future roadmap?
Cloud-based AI? Think again…
ai3™ is designed to run on the device, not in the cloud. This means that all sound identification, analysis and decision making is done locally and instantly while using very little memory or CPU capacity. What’s more, this approach offers consumers the level of privacy that they expect.
Responsive
On-device recognition in real-time
Private
No need to stream audio to the cloud
Scalable
Multiple sounds recognised by a single device
Accurate & precise
Unique audio scene analyser engine
Efficient
Low memory and power footprint
Adaptive
Learns and adapts to the local sound environment
We are with you all the way
As your long-term sound recognition partner we’ll use our experience to help you through the process of integrating our software into your products.
Our dedicated customer engagement team will give you the peace of mind you need during the integration calibration and certification rigorous process that we employ.
Integration
Calibration
Certification
ARM and Audio Analytic publish whitepaper on sound recognition
Get in touch
With offices in Palo Alto and Cambridge, UK we can discuss your requirements and our technology in more detail.
We can send you our mutual NDA, or use yours, and can understand your plans and help you understand how artificial audio intelligence could enhance your products.
Teaching tech to hear
Our embedded, flexible sound recognition software platform is based on the pioneering work that we do on data collection, data analysis and algorithm development.