Automatic determination notes of melodic whistling sounds

The computer is one of the most important discoveries of the last century, and is used almost all scientific fields. One of these fields is the music. Primarily for the production and sharing music with other people, a good musical knowledge and musical notes knowledge are required. Although there is not a lot of people that have this information, however, everyone can play almost all the music with a whistle. So, the whistle can be enough tool to develop new melodies.In this study, we describe an automatic system for transforming man-made melodic whistles into notes. Sound signals transferred to a computer via the microphone become a time series consisting of certain range values. Many mathematical operations can be performed on this series. Due to have a stochastic character, in this series the data brings to the fore the use of statistical methods. Firstly, we apply DC shifting to our audio signal. So the mean of signal is zero. Then we suppress the noise using various filters. We determine whistles and silence regions. Then we find the duration of this whistles regions. Otsu algorithm is used for finding whistling and silence regions. We determine the fundamental frequencies of whistling parts with spectral analysis. Eventually, these fundamental frequencies are grouped and so the notes of audio signal are found.