Online  /  November 08, 2021  -  November 12, 2021

International Society for Music Information Retrieval ISMIR 2021

The ISMIR conference is the world’s leading research forum on processing, searching, organising and accessing music-related data. The 22nd International Society for Music Information Retrieval Conference will take place from Nov 8-12, 2021 in an online format. 

Fraunhofer IDMT's Semantic Music Technologies group is happy to meet you at the ISMIR 2021. Arrange a meeting via the ISMIR Lab Showcase, or get in touch with us via the contact form on this page. 

Live Slots at ISMIR Lab Showcase 

  • Nov 9, 2021 from 16:30-17:30
  • Nov 12, 2021 from 12:00-13:00

A Novel Dataset for Time-Dependent Harmonic Similarity between Chord Sequences

Bittner, Franca; Abeßer, Jakob; Nadar, Christon-Ragavan; Lukashevich, Hanna; Kramer, Patrick

State-of-the-art algorithms in many music information retrieval (MIR) tasks such as chord recognition, multipitch estimation, or instrument recognition rely on deep learning algorithms, which require large amounts of data to be trained and evaluated. In this paper, we present the IDMT-SMT-CHORD-SEQUENCES dataset, which is a novel synthetic dataset of 45,000 chord progressions played on 45 different musical instruments. The dataset is organized in a triplet fashion and each triplet includes one “anchor” chord sequence as well as one corresponding similar and dissimilar chord progression. The audio files are synthesized from MIDI data using different sound fonts from FluidSynth. Furthermore, we conducted a benchmark experiment on time-dependent harmonic similarity based on learnt embedding representations. The results show that a convolutional neural network (CNN), which considers the temporal context of a chord progression outperforms a simpler approach based on temporal averaging of input features.

Late Breaking Demos (B), "Main Hall", Nov 11, 2021, 8:00-10:00