skip to primary navigationskip to content

AudioTelligence Ltd

Improving speech recognition in noisy environments. HEAR MORE ... RECOGNISE MORE

We are an experienced team of signal processing experts based in Cambridge, UK with a track record of bringing innovative products to market. We have developed a technology in collaboration with the Cambridge University Signal Processing Group that uses Bayesian statistics and elements of Machine Learning to extract clean audio from noisy environments e.g. speech in the presence of vehicle noise/radio/music/children/air-conditioning/household appliances/TV etc.

The technology greatly improves the command recognition rate of services such as Alexa in the presence of any noise. We benchmarked it against the Amazon Dot using a recorded voice adding 100 items to a shopping list. In background noise our technology in conjunction with the Alexa Voice Service achieved 91% recognition rate in conditions where the Dot achieved 22%.

The technology achieved similar results in an independent in-car case study devised and carried out by Edinburgh University Centre for Speech Technology Research.

Our technology requires minimal resources and uses low cost uncalibrated mics so it is easily integrated into consumer products to improve usability in real world conditions. Applications include home assistants, smartphone, in-car, hearing assistance, amongst others, and we are already in early discussions with several potential partners.

Our business model is IP licensing.

Company Information

Year Founded: 2017


Linkedin: AudioTelligence