Of all the different types of music, a cappella is special because there's no instrumental accompaniment; it is the voice that matters. It exemplifies the beauty of simplicity, but learning an a cappella song can be challenging, especially when the new song from your favorite singer has no sheet music available to learn. To distinguish each voice part is mostly a manual process that requires expertise in music, while the task is prone to mistakes and time-consuming. At Building 21, I proposed to use machine learning to automatically separate and transcribe voice parts in acapella songs. Ideally, such a tool can make learning a capella much easier and encourage more people to have a taste of the beauty of a capella.