Reflect on a moment at a holiday gathering or networking event. With multiple people talking and some music humming in the background, do you ever wonder how it is possible to tune that all out and focus on a private conversation? According to the American Academy of Audiology, this is due to “the cocktail party effect”; a phenomenon described as the “ability to focus one’s attention on a particular stimulus while filtering out a range of other stimuli.”
Now think about a listening device that you’re trying to command with your voice during the gathering. Likewise, they need to recognize and capture your voice in order to play the song you’ve requested.
The startup company, Yobe, has created a patented approach for such applications. Using a combination of signal processing, advanced artificial intelligence, and broadcast studio techniques, they can identify unique voice biometrics and track individuals even within noisy environments. The team discusses why they chose MATLAB to support their development in this month’s feature video.
To leave a comment, please click here to sign in to your MathWorks Account or create a new one.