Microsoft is continuing to work on improving audio and video functionality in its Teams group-collaboration platform. On June 13, officials announced another handful of features -- including much-requested echo cancellation -- meant to improve the audio functionality in Teams.
Officials said Teams will use machine learning to recognize the difference between the sound from a speaker and a user's voice. This will help alleviate issues that can happen when participants have their microphones too close to their speakers, resulting in a sound loop between the input and output devices. This echo cancellation functionality won't suppress speech or result in multiple parties being unable to speak all at once, they said.
Teams also will use a machine-learning model to convert captured audio signals to sound, resulting in less reverberation, which can cause people to sound like they're speaking in a cave. And officials said that teams will use a model trained with 30,000 hours of speech samples to improve "interruptability" in natural conversations.
"Instead, we either used publicly available data or crowdsourcing to collect specific scenarios. We also ensured that we had a balance of female and male speech, as well as 74 different languages," officials explained in another blog post.
Echo cancellation, improved interruptability, and dereverberation are rolling out for Teams users on Windows and Mac devices and will be coming soon to mobile platforms, according to officials.