Pitch Detection Algorithms and their application on tom drums

Abstract: Pitch detection is of great value within the audio industry and is defined as the perceived height of sound. In this paper, the challenging task of determining pitch of tom drums is presented, known to have an indefinite pitch. The aim of the study is to explore the possibility of determining their pitch. The thesis is divided into three phases, to further improve our knowledge on the pitch of tom drums. The first phase concludes previous work on pitch detection algorithms with regards to tom drums. The second phase concludes a user study which tasks the test subjects to estimate the pitch of different tom drums. The results of the two first phases are concluded and further analyzed in the third phase where we improve upon the Normalized Square Difference method to match the results of the user study. The final algorithm uses an average peak amplitude algorithm with a base in NSDF, it shows a 93% match towards the data set, however to guarantee this result outside the data set further qualitative studies are required.