Microsoft Teams now uses artificial intelligence to enhance echo, interruptions, and acoustics

Microsoft has spent the last two years introducing flashy new productivity capabilities to Teams, and now, due to artificial intelligence, the company is revamping the fundamentals. They have all been on a call where someone’s room acoustics make it difficult to hear them, or seen two people trying to communicate at the same time, resulting in an awkward “no, you go ahead” situation. Microsoft’s new AI-powered voice quality improvements could help to reduce or even eliminate these annoyances on a daily basis.

Microsoft is now utilising machine learning models to improve room acoustics, so you won’t sound like you’re in a cave any longer. In an interview, Microsoft’s Robert Aichner, a principal program director for intelligent conversation and communications cloud, says, “While we have been trying our best with digital signal processing to do a really good job in Teams, we have now started using machine learning for the first time to build echo cancellation where you can truly reduce echo from all the different devices.”

Microsoft has been testing this for months, putting its models to the test in the real world to make sure Teams users see the echo reduction and call quality improvements. 30,000 hours of speech were used to train the software’s algorithms, and thousands of devices were captured using crowd sourcing, in which Teams users are paid to record their voice and playback audio from their device.

“We also simulate about 100,000 different rooms… the room acoustics play a big role in echo cancellation,” Aichner explains. As a result, call audio quality has improved significantly, and echo has been eliminated, allowing numerous people to speak at the same time. In the video above, you can see all of the improvements in action.

If Teams detects sound bouncing or reverberating in a room, the model will convert and process captured audio to make it sound like Teams players are speaking into a close-range microphone rather than an echoey mess.

The most impressive feature is the ability for people on Teams calls to interrupt each other without the awkward overlap where you can’t hear the other person because to the echo. All of this work, as well as earlier enhancements to AI-based noise suppression, is now available in Microsoft Teams. Instead of using the cloud, all processing is done locally on client devices.

“We said we want to do to do it on the client, because the cloud is still expensive if you want to do every call processed in the cloud… and obviously we’d have to pass that cost onto the customer” explains Aichner. That would possibly limit these crucial Teams enhancements to paying customers, whereas the on-device route ensures that features like noise suppression are available on 90% of Teams-enabled devices.

All of these new Microsoft Teams features, as well as certain real-time screen optimizations for text in videos and AI-based bandwidth improvements during video or screen-sharing sessions, are now available.