The paper presents results of the analysis of violins recorded during the final stage of the international violin-makers competition held in Poznan in 2011. In the quest for attributes that are both efficient for machine learning and interpretable for human experts we referred to the research of violin acousticians: Duennwald, Buen, and Fritz and calculated violin sound power in frequency bands recommended by these researchers. The resulting features, obtained for the averaged spectra of the musical pieces played at the competition, were used for clustering and classification experiments. Results are discussed, and a notable experiment is presented where the classifier assigns each analyzed violin to an instrument from the precedent violinmakers’ competition (2001) and compares their ranking.
(See document for exact affiliation information.)
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member Join the AES. If you need to check your member status, login to the Member Portal.
