Microsoft brings AI-enabled ‘Noise Cancellation’ feature to Skype

Microsoft has been updating its popular team voice and video calling platform Skype regularly with new features that make it easier to connect with others. The platform has also played a vital role for organisations and on an individual level for connecting during the lockdown last year. But the firm is not done updating the software. This time, it has brought noise cancellation to its desktop version.

In the blog post, the Skype team states that this tech was originally made for Microsoft Teams. “We are pleased to announce the release of our latest background noise suppression feature in the Skype desktop app. Originally developed for Microsoft Teams, this new feature is designed to silence just about everything except for your voice when you’re meeting on Skype,” it adds.

This feature can be found under the ‘Settings’ option and can be toggled with different options - Auto, Low and High.

It is worth mentioning that the noise cancellation feature is not yet available on the web version or mobile versions of Skype. You can get it in Windows desktop app and on Mac as well.

Talking more about the tech behind this feature, the blog post adds that the new noise cancellation tech works by analysing your audio feed and using trained deep neural networks to remove the noise without affecting the speaker’s voice. This is different as traditional noise suppression algorithms can address simple, consistent noises like that of a fan, it cannot differentiate when it comes to more complex noises like typing on the keyboard, crunch of a food wrapper or when your pet dog is howling.

Also read: Microsoft finally adds support for blurred backgrounds on Skype for Android

“This technology relies on machine learning (ML) to learn the difference between clean speech and noise and is frequently referred to as artificial intelligence (AI). A representative dataset is used to train the ML model to work in most of the situations our Skype users experience. There needs to be enough diversity in the dataset in terms of the clean speech, noise types, and the environments from which our users are joining online calls,” states the post.

To achieve this level of tech, the team used approximately 760 hours of clean speech data and 180 hours of noise data in their datasets for training. For noise data, they included 150 noise types to cover the diverse scenarios including keyboard typing, running water, snoring and more.