The paper presents a method for segmentation and labeling of ethnomusicological field recordings. Our goal was to design a robust algorithm that would approximate manual segmentation of field recordings. Audio fragments are classified into speech, solo singing, choir singing, instrumental or bell chiming performance. Then, a set of candidate segment boundaries is obtained, and finally the recording is segmented with a probabilistic model that maximizes the posterior probability of segments. Evaluation of the algorithm on a set of field recordings from the Ehtnomuse archive is presented.
COBISS.SI-ID: 7368532
Paper explores whether we can derive a melodic similarity measure, correlating to variant types and measuring songs belonging to the same variant type as more similar, in contrast to songs from different variant types. The measure would be useful for folk song retrieval based on variant types, classification of unknown tunes, as well as a measure of similarity between variant types. We evaluated the measure on the task of classifying an unknown melody into a set of existing variant types. The proposed measure gives the correct variant type in the top 10 list for 68% of queries in our data set.
COBISS.SI-ID: 7368788