The technology behind our noise cancellation

Our state of the art AI model removes all background noise from audio streams and is proven via the industry standard PESQ scale to be the highest performing technology in the market.

Our model works by utilising the latest Deep Learning techniques, and training our neural networks on thousands of hours of audio data. Our unique data set has been built up over years and contains terabytes of information on a vast array of both noise categories and clean speech samples. These samples are then combined and augmented with proprietary algorithms to precisely mimic real world communication scenarios.

As the training progresses, over time our model becomes an expert in identifying what is noise and what is voice, and is thus able to completely remove the noisy part of the signal, leaving pristine speech behind.