Masterarbeit MSTR-2020-70

Bibliograph.
Daten
Dzamashvili, Anastasia: Analyzing Music Instrument Playing History via an Interactive Visual Diary.
Universität Stuttgart, Fakultät Informatik, Elektrotechnik und Informationstechnik, Masterarbeit Nr. 70 (2020).
53 Seiten, englisch.
Kurzfassung

As live-logging becomes an increasingly habitual practice, Personal Visual Analytics face the inevitable challenge of adapting to all forms of tracked data. Our concept offers the input and visualization of recorded MIDI data to those who use musical instruments, ranging from amateur musicians to professional performers. Users can oversee and reflect on the development of their repertoire throughout the span of months, condensed into two different formats. The Overview maps each recording in regard to the start and end times and places it onto the respective day with a color code corresponding to its title. Interaction with each separate recording provides linkage through arcs to others analogous with respect to either title or note similarity. For the latter, we support the Levenshtein distance and the Gotoh algorithm as similarity measures. The Weekview produces a more concentrated canvas, mapping the tracks with a starker focus on the day of the week at which the user prefers to play. What allows this view to display data in a more concise form, is that the duration and exact save time information are not taken into consideration, making it easier to read. We evaluated our approach and prototypical implementation in a case study. For this, we first used a small set of user data and, due to time constraints, further carried out the examination using artificially generated data, simulating different cases of users with varying recording habits. Our implementation performs well under up to several hundreds of recordings. However, the recordings are harder to spot when the recording time exceeds a year or when there are multiple recordings saved within a small time frame. Our concept provides a novel way of tracking and exploring recorded music for those already intrigued by the idea of visualizing their recording library, while potentially attracting musicians who track less, but wish to discover patterns in their practice regimes.

Abteilung(en)Universität Stuttgart, Institut für Visualisierung und Interaktive Systeme, Visualisierung und Interaktive Systeme
BetreuerSedlmair, Jun.-Prof. Michael; Heyen, Frank
Eingabedatum22. April 2021
   Publ. Informatik