The goals of the project are to develop music information retrieval algorithms for folk music. The work is based on the Slovenian archive of folk music and dance culture "Ethnomuse", which gathers a large variety of multimedia materials, including audio, scores, videos, lyrics and textual metadata (including GIS locations). Music includes solo and choir singing, instrumental music, as well as bell chiming performances. We plan to develop tools and methods that will be used by folk music researchers to interact with the archive and include segmentation, transcription and search-by-melody in symbolic and audio domains. We are also developing algorithms for data mining and visualisation of the archive’s contents. Folk music is passed on orally, so songs change over the time, mix with other songs etc. Complex relations exist between song variants in the archive, which encourages us to research methods of discovering and visualizing these relations.
Researches
We are pursuing the following research goals:
Publications
@inproceedings{7368532,
author = {Marolt, Matija},
booktitle = {ISMIR 2009 : proceedings of the 10th {International} {Society} for {Music} {Information} {Retrieval} {Conference}, {October} 26-30, 2009, {Kobe}, {Japan}},
year = {2009},
pages = {75--80},
title = {Probabilistic segmentation and labeling of ethnomusicological field recordings},
}
@inproceedings{7224404,
author = {Marolt, Matija},
booktitle = {SMC 2009 : proceedings of the 6th {Sound} and {Music} {Computing} {Conference}, 23-25 {July} 2009 {Casa} da {M}{\' u}sica, {Porto} - {Portugal}},
year = {2009},
pages = {137--142},
title = {Non-negative matrix factorization with selective sparsity constraints for transcription of bell chiming recordings},
}