Microsoft Teams now uses AI to improve echo, interruptions, and acoustics

Image: Microsoft

Microsoft has spent the past two years adding flashy new productivity features to Teams and now the company is rethinking how the basics work thanks to artificial intelligence. We have all been on a call where someone has poor room acoustics making it hard to hear them, or seen two people try to talk at the same time. These day-to- day annoyances should be eliminated by Microsoft's new voice quality improvements.

Machine learning models are being used by Microsoft to improve room acoustics so you won't sound like you're in a cave. Robert Aichner, a principal program manager for intelligent, said that they have started using machine learning to build echo cancellation where you can reduce echo from all the different devices.

Microsoft has been measuring its models in the real world to make sure Teams users notice the improvements. 30,000 hours of speech was used by the software maker to train its models, and thousands of devices were captured through crowd-sourced audio and video.

The room acoustics are a big part of echo cancellation. The result is an improvement in call audio quality and an elimination of echo that allows multiple people to speak at the same time. The improvements can be seen in the video above.

The model will process captured audio and convert it to sound like teams are speaking into a microphone instead of an echoey mess.

The ability for people to interrupt each other on Teams calls is the most impressive part. Microsoft has made improvements to its noise suppression with the use of artificial intelligence. The processing is done on client devices.

The cloud is still expensive if you want to do every call in the cloud and so we want to do it on the client. The on- device route means that features like noise suppression are available on 90 percent of devices using Teams.

All of the new Microsoft Teams improvements are now live, along with some real-time screen adjustments for text in videos and improvements to bandwidth constraints during video or screen-sharing calls.