- The company’s usage of AI in a call audio quality has resulted in significant improvements and an elimination of echo that allows multiple users to converse simultaneously.
- A stunning aspect of the application is that it allows people to interrupt another user on Teams calls now without the overlap of voices due to the echo.
Microsoft has been adding flashy new productivity features to Teams for the past two years. Now, the company is overhauling how the fundamentals function, all thanks to Artificial Intelligence (AI).
Microsoft’s new AI-powered voice quality improvements will enhance or eliminate day-to-day annoyances like someone having poor room voice acoustics or an awkward situation when two people speak simultaneously.
The multinational technology corporation is now using a Machine Learning model to improve room acoustics that doesn’t make a user’s voice echo.
“While we have been trying our best with digital signal processing to do an excellent 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,” explained Robert Aichner, a principal program manager for intelligent conversation and communications cloud at Microsoft.
The software company has been conducting tests for months, measuring its models in the real world to ensure Teams’ users notice the echo reduction and improvements in call quality.
Microsoft used 30,000 hours of speech to assist and train its models and has captured thousands of devices with the help of crowdsourcing, where it pays users to record voices and playback audio from their devices.
“We also simulate about 100,000 different rooms… the room acoustics play a big role in echo cancellation,” said Aichner.
The company’s usage of AI in call audio quality has resulted in significant improvements in call audio quality and the elimination of echo that allows multiple users to converse simultaneously.
When the Teams application detects a bounce or reverberation of sound in a room resulting in poor audio, the model converts the recorded audio and processes it to make it sound like Teams participants are talking in close range instead of an echoey mess.
A stunning aspect of the application is that it allows people to interrupt another user on Teams calls now without the overlap of voices where you can’t hear the other person due to the echo.
Microsoft is now moving all this work in Teams, alongside its improvements with AI-based noise suppression. The processing is done locally on the client’s device instead of in the cloud.
“We said we want 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,” explained Aichner.
This change would drastically restrict important Teams’ improvements to paying customers. The on-device route means that features like noise suppression are available on 90% of users’ devices using Teams.
The latest enhancements made by Microsoft Teams are now live, along with some real-time screen optimizations for subject matter in videos and AI-based improvements to bandwidth limitations during video or share-sharing calls.