Audio Analytic

About Audio Analytic

Audio Analytic is the global leader in intelligent sound recognition. Our cloudless AI technology gives consumer products a sense of hearing beyond just speech and music and, as a result, a greater understanding of context.

Our embedded software platform (ai3™ and ai3-nano™) are suitable for a wide range of products from smart speakers and video doorbells, to smartphones and true wireless earbuds. They have been licensed to some of the world’s most prominent consumer technology companies and can be found in many products available today.

Highly-accurate, compact, edge-based AI is only possible because of our world-leading core technology integrated into a specialist ML pipeline.

Machine learning is all about the data and Alexandria™ is the world’s largest, commercially-exploitable audio dataset for machine learning, with over 30 million labelled recordings across 1,000 sound classes. All the audio data has been primarily sourced through dedicated data collection campaigns, involving a global network of volunteers or through our purpose-built anechoic sound labs.

We have also designed AuditoryNET™ - a range of compact, sophisticated, topologically-optimised DNNs and a framework for sound event and acoustic scene recognition; which together model both the acoustic and temporal characteristics of sound. AuditoryNET™ makes use of a dedicated machine learning pipeline which is optimised for sound recognition and includes our patented Loss Function for model training and three levels of data labelling (fine, episodic and weak), and various compression and acceleration techniques.