Speaker: Diego de Benito Gorrón.
Abstract: Based on https://guitarset.weebly.com/uploads/1/2/1/6/121620128/xi_ismir_2018.pdf. The guitar is a popular instrument for a variety of reasons, including its ability to produce polyphonic sound and its musical versatility. The resulting variability of sounds, however, poses significant challenges to automated methods for analyzing guitar recordings. As data driven methods become increasingly popular for difficult problems like guitar transcription, sets of labeled audio data are highly valuable resources. In this paper we present GuitarSet, a dataset that provides high quality guitar recordings alongside rich annotations and metadata. In particular, by recording guitars using a hexaphonic pickup, we are able to not only provide recordings of the individual strings but also to largely automate the expensive annotation process. The dataset contains recordings of a variety of musical excerpts played on an acoustic guitar, along with time-aligned annotations of string and fret positions, chords, beats, downbeats, and playing style. We conclude with an analysis of new challenges presented by this data, and see that it is interesting for a wide variety of tasks in addition to guitar transcription, including performance analysis, beat/downbeat tracking, and chord estimation.