Two recent updates (thanks to Frederik Heerde!):
– AMUSE can be started now with Java 1.8 (in the older version, using of Java 1.8 led to problems for classification with trained models).
– Integrated RapidMiner library was updated to version 5.
The course on music data analysis (in German) based on the textbook Music Data Analysis: Foundations and Applications will again take place at TU Dortmund, Department of Statistics in September 2017 (see also the course web page from the last year)
The program of the following 39th SIGMA meeting on 25.07.2017, 14:00-16:00, which takes place at Faculty of Computer Science, TU Dortmund, Otto-Hahn-Str. 14, room E04
(talks in German):
14:00-14:05 Welcome greetings
14:05-14:35 Bachelor thesis
Andreas Jöbges: Optical recognition of bagpipe notation
14:35-15:05 Bachelor thesis
Oliver Magiera: Recognition of melodic patterns in symbolic music data
15:05-15:35 Project group
PG 607: Reactive Jukebox – state after the 1st term
15:35-15:55 Planning of lectures on music data analysis in summer terms 2017 and 2018
15:55-16:00 Miscellaneous, next meeting
The thesis „Andreas Jöbges: Optische Erkennung von Dudelsacknotation“ (Optical recognition of bagpipe notation, PDF in German) presents a system which produces digital score in „BWW“ format from scanned music pieces with the help of neural networks trained for the recognition of notes and specific symbols in bagpipe notation.
E-Services Research Group, TU Dortmund currently conducts a study about creating music playlists and would like to invite you to participate in an online experiment.
The experiment takes about 10-15 minutes. You will be asked to create a short playlist for a particular occasion and to give us some feedback on the process.
There you can, of course, discover new and exciting music, too.
Start the experiment here: http://bit.ly/playlisting-en.
The German-speaking workshop „Musik trifft Informatik“ („music meets computer science“) will be supported with a tutorial on automatic music processing. Please see details below (in German):
GI Workshop: Musik trifft Informatik
47. Jahrestagung der Gesellschaft für Informatik e.V. (GI)
Chemnitz, Germany, 26 September 2017
https://www.audiolabs-erlangen.de/resources/MIR/2017-GI-Workshop-Musik/
GI Tutorial: Automatisierte Methoden der Musikverarbeitung
47. Jahrestagung der Gesellschaft für Informatik e.V. (GI)
Chemnitz, Germany, 25 September 2017
https://www.audiolabs-erlangen.de/resources/MIR/2017-GI-Tutorial-Musik/
The following papers were accepted:
Kamehkhosh, I. and Jannach, D.: User Perception of Next-Track Music Recommendations. UMAP 2017, Bratislava, 2017
[describes a user case study on the evaluation of different algorithms for playlist generation with regard to perceived quality]
Scholz, F., Vatolkin, I., and Rudolph, G.: Singing Voice Detection across Different Music Genres. AES 2017), Erlangen, 2017
[compares three methods for recognition of vocals using distinct training and validation sets from six music genres]
Im Rahmen der 47. Jahrestagung der Gesellschaft für Informatik e.V. (GI) in Chemnitz findet ein Workshop zum Thema „Musik trifft Informatik“ statt. Details zu Beiträgen und Inhalten findet man auf den folgenden Webseiten:
https://www.audiolabs-erlangen.de/resources/MIR/2017-GI-Workshop-Musik/
https://informatik2017.de/ws01-musik-trifft-informatik/
Wichtige Termine
Weiterhin wird es ein Tutorial zum Thema „Musikverarbeitung“ geben. Details findet man auf den folgenden Webseiten:
https://www.audiolabs-erlangen.de/resources/MIR/2017-GI-Tutorial-Musik/
https://informatik2017.de/tut01-musikverarbeitung/
Termin: 25.09.2017
The student project group „The Reactive Jukebox“ (supervised by D. Jannach and I. Vatolkin) will start in summer term 2017.
The goal of the project is to develop an application for music recommendation with an interactive interface to listener requests. Such interaction should be done in an intuitive and user-friendly way, for instance providing feeback in a spoken language or describing music with the help of interpretable semantic features („please recommend me faster songs with more distorted guitars and more vocal segments“). For the indexing of music, it is planned to integrate features of different sources, like community tags, meta data, audio descriptors, or symbolic characteristics.
F. Ostermann, I. Vatolkin, G. Rudolph: Evaluation Rules for Evolutionary Generation of Drum Patterns in Jazz Solos. Proceedings of the 6th International Conference on Computational Intelligence in Music, Sound, Art and Design (EvoMUSART 2017), Amsterdam, 2017
I. Vatolkin: Generalisation Performance of Western Instrument Recognition Models in Polyphonic Mixtures with Ethnic Samples. Proceedings of the 6th International Conference on Computational Intelligence in Music, Sound, Art and Design (EvoMUSART 2017), Amsterdam, 2017
