Voice Dataset

Using a standard sound recorder Logitech, that does not filter out the background noise, the voice of the subject is captured twice, resulting in two files with extension .wav. The subject speaks at about six inches from the microphone and is asked to read a text composed of two short sentences followed by a set of numbers: My favorite sport is basketball , I love to support this research and 1, 6, 7, 2, 3, 0, 8, 7, 4, 5, 9, 1. The files are typically 10 to 25 seconds in length.