Microsoft Teams uses a trained model fed with 30,000 hours of speech samples to retain ideal voice quality while suppressing unwanted audio signals
Microsoft has announced the roll-out of the new AI enhancements for its video conferencing software Microsoft Teams, which will enhance audio and video quality on calls. Announced in a Microsoft blog post, the features include echo cancellation, real-time screen optimisations for video playback, de-reverberation, and brightness and soft-focus video enhancements. Microsoft Teams uses AI technology to differentiate between a speaker’s sound and the user’s voice. This feature eliminates the echo without suppressing speech or multiple users’ ability to speak simultaneously.
De-reverberation can adjust poor room acoustics using machine learning to convert captured audio signals to sound as if users are speaking into a close-range mic. This can fix the specific sound issues where the room acoustics may cause sound to bounce or reverberate, forcing the user’s voice to sound shallow.
Another feature is Interruptibility, the ability to interrupt for clarification or validation, resulting in more natural conversations. The software will use full-duplex (two-way) transmission of audio that will allow users to speak and hear others simultaneously.
Microsoft Teams uses a trained model fed with 30,000 hours of speech samples to retain ideal voice quality while suppressing unwanted audio signals, thus resulting in more fluid dialogues. This will solve the issue of removing echo while maintaining full-duplex audio when using headphones.
The new Microsoft Teams software update has background noise suppression, with the company’s proprietary AI-based noise suppression a default feature for all users. Moreover, AI will help with better video and screen sharing quality.
