Algorithm decodes peculiarities of different yodeling styles

An algorithm confirms what yodelers have been telling each other for a long time: Northwestern Swiss and Central Swiss yodels can be distinguished.

Men and women yodeling in traditional clothes

The Innerrhoder "Rugguusseli", the Ausserrhoder "Zäuerli" or the Innerschweizer "Juiz": The yodel has many names in Switzerland. But to what extent do the yodel styles actually differ? This is what musicologist Yannick Wey from the Lucerne University of Applied Sciences and Arts and computer scientist Cornelia Metzig from Queen Mary University in London investigated. For this purpose, they focused on the yodeling strongholds of central Switzerland with the cantons of Obwalden and Nidwalden, and northeastern Switzerland with the cantons of Appenzell Innerrhoden and Ausserrhoden, as well as Toggenburg.

A machine learning algorithm served as a tool, which the interdisciplinary duo fed with sheet music from a total of 217 yodel melodies. Based on the melodic properties, the program attempts to recognize characteristic features of the different styles. The "Melodyfeatures" toolkit, which extracts the features from melodies and thus makes them usable for algorithms, was developed by Metzig and has already been used to classify various folk songs worldwide.

Result of the yodel study: The melodies differ in all cantons. But within the regions - Central Switzerland and Northeastern Switzerland - the differences were too weak, so that the algorithm made relatively many false assignments. The situation was different between the regions: Here, the algorithm correctly assigned more than three out of four yodels.

Written by: sda

Photos: keystone

Read more