Skip to content

Motivation

When audio data is streamed through the internet on platforms such as Zoom, the audio is compressed before transmission. This allows for faster loading times, meaning that you internet does not have to be as fast to receive the information in real time. However, this also means that the quality of the audio is reduced. More specifically, the audio is sampled less frequently (like with lower frame rates on videos), and the highest frequencies of audio are removed altogether. In extreme cases, this can make audio sound like it is underwater.

This is where DeciBull can help. DeciBull will take audio data coming in, and transform it using machine learning. Our algorithm will detect where audio information is missing, and work to fill in the gaps. The result is audio with a higher frequency range and sampling rate, which will sound clearer and more crisp to the human ear. DeciBull can upsample audio in real time, so there will be no additional lag when used with live-streamed audio.