Register for free and continue reading
Join our growing army of changemakers and get unlimited access to our premium content
A new app makes meetings more efficient by recognising and cancelling out all unwanted background noise and echoes
Spotted: By now, we have probably all been there – you are in a virtual meeting and the first 10 minutes are spent reminding everyone to mute their mics before the sound of barking dogs, talking housemates, and feedback becomes too distracting. That is the situation that startup Krisp aims to solve. The company has developed an AI-powered app that removes all types of background noise and echoes from your microphone.
Krisp works by adding a virtual filter between the users’ microphone and their calling app – whether that is Teams, Zoom or your mobile phone. Krisp’s noice cancelling technology uses a deep neural network which has been trained to recognise background sounds using more than 20,000 noises and 50,000 different speakers.
Krisp can be downloaded as an app and turned on and off with a click. Once on, it filters the audio from any audio calling apps automatically. Because Krisp operates locally, conducting all audio processing entirely on the users’ device, it is completely secure. The app will remove background noise on the user’s end, and also any noise or echoes coming from those on the other end of the call.
By removing disruptive background noise, Krisp points out that its app can help improve efficiency and customer satisfaction. On its website, the company says that, “Krisp removes both the echo resonating from walls of empty room and the echo that occurs from your own voice during the call,” adding, “Your mic will no longer act like a parrot.” The company should know – most of their employees work remotely.
Over the last two years, we have seen a huge number of innovations aimed at improving the now ubiquitous online meeting. These have included the use of Minecraft for holding meetings in virtual offices, and a virtual whiteboarding tool designed for brainstorming.