QUARTET DATASET

About the dataset The dataset contains 96 multimodal recordings of string quartet performance. Acquired data includes the audio of each performance in separate tracks: up to two ambient tracks and individual tracks acquired through contact microphones. In addition, each recording includes instrumental motion capture and a set of derived instrumental bowing descriptors for each performer, upper body movement for each performer, as well as high-quality video. The recordings were carried out as part of the experiments on ensemble performance inter-dependence reported in the following doctoral dissertation: Papiotis, P. (2015). A computational approach to studying interdependence in string quartet performance. Ph.D. Thesis, Universitat Pompeu Fabra. For further information about the multimodal representation refer to: Maestre, Esteban and Papiotis, Panagiotis and Marchini, Marco and Llimona, Quim and Mayor, Oscar and Pérez, Alfonso and Wanderley, Marcelo M, "Enriched Multimodal Representations of Music Performances: Online Access and Visualization," in IEEE MultiMedia, vol. 24, no. 1, pp. 24-34, Jan.-Mar. 2017, doi: 10.1109/MMUL.2017.3. License Dataset compiled and authored by Esteban Maestre, Panos Papiotis, Marco Marchini and Alfonso Perez. Copyright © 2015 Music Technology Group, Universitat Pompeu Fabra. All Rights Reserved. The QUARTET dataset is offered free of charge for internal non-commercial use only. You may not redistribute, publically communicate or modify it, unless expressly permitted by the Universitat Pompeu Fabra (UPF) or by applicable law. The individual contents of the dataset may be protected by copyright, and a non-transferrable limited license to reproduce and use the same is granted for the indicated purpose only. The dataset and its contents are made available on an “as is” basis and without warranties of any kind, including without limitation satisfactory quality and conformity, merchantability, fitness for a particular purpose, accuracy or completeness, or absence of errors. Subject to any liability that may not be excluded or limited by law, the UPF is not liable for, and expressly excludes, all liability for loss or damage however and whenever caused to anyone by any use of the QUARTET dataset or any part of it. Structure of the dataset Initially, we created a RepoVizz datapack for each of the recordings which can be browsed at: http://repovizz.upf.edu. We strongly recommend using RepoVizz for sensor visualization. Both in the RepoVizz platform and in this zenodo repository each datapack/subdirectory has an XML file where all the included files are described. To use all of the advanced functionalities of RepoVizz (including WebGL 3D visualization), it is recommended to use Chrome. For further information about the RepoVizz platform see the tutorial. Each recording contains a number of data files corresponding to the various data modalities captured in the experiment. The files are organized inside the datapack structure (sub-directories) in the following nodes: An AudioGroup node, containing the recorded audio from each of the utilized microphones (a cardioid microphone, a binaural microphone, and one bridge pickup microphone per performer) Each audio recording is accompanied by an AudioDesc node, containing the audio features extracted using the Essentia library. A VideoGroup node, containing the recorded HQ video A Score node, containing the note onset and offset times for the audio of each instrument as they were obtained by the score-performance alignment A DescriptorGroup node, containing the instrumental gesture features for each instrument computed from data acquired using the wired motion capture system A MoCapGroup node, containing the raw 6DOF (position & orientation) data acquired using the Polhemus wired motion capture system A MoCapGroup node, containing the upper body marker data, instrument marker data, and bow marker data acquired using the Qualisys optical motion capture system As a result of errors during the recording (especially in regard to the optical motion capture software and the video camera), some datapacks may not contain all of the above nodes. Audio recordings, score-performance alignment, instrumental gesture features and raw motion capture data from the wired system are present in all datapacks. In the list of recording below, we denote the contents of each recording as follows: A for audio, S for score-performance alignment, P for wired motion capture data, Q for wireless motion capture data, and V for video. By accessing the database, you accept the conditions of use. The datapacks can be found in the list below. List of recordings I1S1 - Intonation Exercise 1 Solo violin 1 [A|S|P|Q|V] I1S2 - Intonation Exercise 1 Solo violin 2 [A|S|P|Q|V] I1S3 - Intonation Exercise 1 Solo viola [A|S|P|Q|V] I1S4 - Intonation Exercise 1 Solo cello [A|S|P|Q|V] I1E1 - Intonation Exercise 1 Ensemble (no annotations) [A|S|P|Q|V] I1E2 - Intonation Exercise 1 Ensemble (with annotations) [A|S|P|Q|V] IP1S1 - Intonation Piece excerpt 1 Solo violin 1 [A|S|P|Q|V] IP1S2 - Intonation Piece excerpt 1 Solo violin 2 [A|S|P|Q|V] IP1S3 - Intonation Piece excerpt 1 Solo viola [A|S|P|Q|V] IP1S4 - Intonation Piece excerpt 1 Solo cello [A|S|P|Q|V] IP1E - Intonation Piece excerpt 1 Ensemble [A|S|P|Q|V] IP2S1/DP2S1 - Intonation / Dynamics Piece excerpt 2 Solo violin 1 [A|S|P|Q|V] IP2S2/IP2S2 - Intonation / Dynamics Piece excerpt 2 Solo violin 2 [A|S|P|Q|V] IP2S3/DP2S3 - Intonation / Dynamics Piece excerpt 2 Solo viola [A|S|P|Q|V] IP2S4/DP2S4 - Intonation / Dynamics Piece excerpt 2 Solo cello [A|S|P|Q|V] IP2E/DP2E - Intonation / Dynamics Piece excerpt 2 Ensemble [A|S|P|Q|V] D1S1 - Dynamics Exercise 1 Solo violin 1 [A|S|P|Q|V] D1S2 - Dynamics Exercise 1 Solo violin 2 [A|S|P|Q|V] D1S3 - Dynamics Exercise 1 Solo viola [A|S|P|Q|V] D1S4 - Dynamics Exercise 1 Solo cello [A|S|P|Q|V] D1E - Dynamics Exercise 1 Ensemble [A|S|P|Q|V] D2S1 - Dynamics Exercise 2 Solo violin 1 [A|S|P|Q|V] D2S2 - Dynamics Exercise 2 Solo violin 2 [A|S|P|Q|V] D2S3 - Dynamics Exercise 2 Solo viola [A|S|P|Q|V] D2S4 - Dynamics Exercise 2 Solo cello [A|S|P|Q|V] D2E - Dynamics Exercise 2 Ensemble [A|S|P|Q|V] DP1S1 - Dynamics Piece excerpt 1 Solo violin 1 [A|S|P|Q|V] DP1S2 - Dynamics Piece excerpt 1 Solo violin 2 [A|S|P|Q|V] DP1S3 - Dynamics Piece excerpt 1 Solo viola [A|S|P|Q|V] DP1S4 - Dynamics Piece excerpt 1 Solo cello [A|S|P|Q|V] DP1E - Dynamics Piece excerpt 1 Ensemble [A|S|P|Q|V] T1S1 - Timbre exercise Solo violin 1 [A|S|P|Q|V] T1S2 - Timbre exercise Solo violin 2 [A|S|P|Q|V] T1S3 - Timbre exercise Solo viola [A|S|P|Q|V] T1S4 - Timbre exercise Solo cello [A|S|P|Q|V] T1E - Timbre exercise Ensemble [A|S|P|Q|V] R1S1 - Rhythm exercise 1 Solo violin 1 [A|S|P|Q|V] R1S2 - Rhythm exercise 1 Solo violin 2 [A|S|P|Q|V] R1S3 - Rhythm exercise 1 Solo viola [A|S|P|Q|V] R1S4 - Rhythm exercise 1 Solo cello [A|S|P|Q|V] R1E - Rhythm exercise 1 Ensemble [A|S|P|Q|V] R2S1 - Rhythm exercise 2 Solo violin 1 [A|S|P|Q|V] R2S2 - Rhythm exercise 2 Solo violin 2 [A|S|P|Q|V] R2S3 - Rhythm exercise 2 Solo viola [A|S|P|Q|V] R2S4 - Rhythm exercise 2 Solo cello [A|S|P|Q|V] R2E - Rhythm exercise 2 Ensemble [A|S|P|Q|V] RP1S1 - Rhythm excerpt 1 Solo violin 1 [A|S|P|Q|V] RP1S2 - Rhythm excerpt 1 Solo violin 2 [A|S|P|Q|V] RP1S3 - Rhythm excerpt 1 Solo viola [A|S|P|Q|V] RP1S4 - Rhythm excerpt 1 Solo cello [A|S|P|Q|V] RP1E - Rhythm excerpt 1 Ensemble [A|S|P|Q|V] RP2S1 - Rhythm excerpt 2 Solo violin 1 [A|S|P|Q|V] RP2S2 - Rhythm excerpt 2 Solo violin 2 [A|S|P|Q|V] RP2S3 - Rhythm excerpt 2 Solo viola [A|S|P|Q|V] RP2S4 - Rhythm excerpt 2 Solo cello [A|S|P|Q|V] RP2E - Rhythm excerpt 2 Ensemble [A|S|P|Q|V] U1S1 - Unity of Execution exercise 1 Solo violin 1 [A|S|P|Q|V] U1S2 - Unity of Execution exercise 1 Solo violin 2 [A|S|P|Q|V] U1S3 - Unity of Execution exercise 1 Solo viola [A|S|P|Q|V] U1S4 - Unity of Execution exercise 1 Solo cello [A|S|P|Q|V] U1E - Unity of Execution exercise 1 Ensemble [A|S|P|Q|V] U2S1 - Unity of Execution exercise 2 Solo violin 1 [A|S|P|Q|V] U2S2 - Unity of Execution exercise 2 Solo violin 2 [A|S|P|Q|V] U2S3 - Unity of Execution exercise 2 Solo viola [A|S|P|Q|V] U2S4 - Unity of Execution exercise 2 Solo cello [A|S|P|Q|V] U2E - Unity of Execution exercise 2 Ensemble [A|S|P|Q|V] UP1S1 - Unity of Execution excerpt 1 Solo violin 1 [A|S|P|Q|V] UP1S2 - Unity of Execution excerpt 1 Solo violin 2 [A|S|P|Q|V] UP1S3 - Unity of Execution excerpt 1 Solo viola [A|S|P|Q|V] UP1S4 - Unity of Execution excerpt 1 Solo cello [A|S|P|Q|V] UP1E - Unity of Execution excerpt 1 Ensemble [A|S|P|Q|V] UP2S1 - Unity of Execution excerpt 2 Solo violin 1 [A|S|P|Q|V] UP2S2 - Unity of Execution excerpt 2 Solo violin 2 [A|S|P|Q|V] UP2S3 - Unity of Execution excerpt 2 Solo viola [A|S|P|Q|V] UP2S4 - Unity of Execution excerpt 2 Solo cello [A|S|P|Q|V] UP2E - Unity of Execution excerpt 2 Ensemble [A|S|P|Q|V] UP3S1/DP3S1 - Unity of Execution/Dynamics excerpt 3 Solo violin 1 [A|S|P|Q|V] UP3S2/DP3S2 - Unity of Execution/Dynamics excerpt 3 Solo violin 2 [A|S|P|Q|V] UP3S3/DP3S3 - Unity of Execution/Dynamics excerpt 3 Solo viola [A|S|P|Q|V] UP3S4/DP3S4 - Unity of Execution/Dynamics excerpt 3 Solo cello [A|S|P|Q|V] UP3E/DP3E - Unity of Execution/Dynamics excerpt 3 Ensemble [A|S|P|Q|V] P1S1 - Phrasing exercise 1 Solo violin 1 [A|S|P|Q|V] P1S2 - Phrasing exercise 1 Solo violin 2 [A|S|P|Q|V] P1S3 - Phrasing exercise 1 Solo viola [A|S|P|Q|V] P1S4 - Phrasing exercise 1 Solo cello [A|S|P|Q|V] P1E - Phrasing exercise 1 Ensemble [A|S|P|Q|V] PP1S1 - Phrasing excerpt 1 Solo violin 1 [A|S|P|Q|V] PP1S2 - Phrasing excerpt 1 Solo violin 2 [A|S|P|Q|V] PP1S3 - Phrasing excerpt 1 Solo viola [A|S|P|Q|V] PP1S4 - Phrasing excerpt 1 Solo cello [A|S|P|Q|V] PP1E - Phrasing excerpt 1 Ensemble [A|S|P|Q|V] PP2S1/RP3S1 - Phrasing excerpt 2 / Rhythm excerpt 3 Solo violin 1 [A|S|P|Q|V] PP2S2/RP3S2 - Phrasing excerpt 2 / Rhythm excerpt 3 Solo violin 2 [A|S|P|Q|V] PP2S3/RP3S3 - Phrasing excerpt 2 / Rhythm excerpt 3 Solo viola [A|S|P|Q|V] PP2S4/RP3S4 - Phrasing excerpt 2 / Rhythm excerpt 3 Solo cello [A|S|P|Q|V] PP2E/RP3E - Phrasing excerpt 2 / Rhythm excerpt 3 Ensemble [A|S|P|Q|V] For each recorded exercise and piece excerpt, five corresponding datapacks have been uploaded: Solo violin 1, Solo violin 2, Solo viola, Solo cello, and Ensemble. A single exception to the above exists for the Intonation exercise (I1), where two ensemble datapacks exist - one without the visual annotations seen in Figure 3.1, and one with. Each datapack has a short-form title that serves as an identifier, with each letter or number in the title denoting a characteristic of the datapack. For example, D2S1 corresponds to Dynamics exercise 2 Solo violin 1, D2S3 corresponds to Dynamics exercise 2 Solo viola, D2E corresponds to Dynamics exercise 2 Ensemble, while DP2E corresponds to Dynamics piece excerpt 2 Ensemble. A legend for the symbols used in the identifier of each datapack can be seen in Table 3.2. Besides the short-form title, each shared datapack has a description (providing details on the datapack itself), an abstract (providing details on the recorded exercise/piece), and a list of keywords by which it is searchable. Each recording contains a number of data files corresponding to the various data modalities captured in the experiment. The files are organized inside the datapack structure in the following nodes: An AudioGroup node, containing the recorded audio from each of the utilized microphones (a cardioid microphone, a binaural microphone, and one bridge pickup microphone per performer) Each audio recording is accompanied by an AudioDesc node, containing the audio features extracted using the Essentia library. A VideoGroup node, containing the recorded HQ video A Score node, containing the note onset and offset times for the audio of each instrument as they were obtained by the score-performance alignment A DescriptorGroup node, containing the instrumental gesture features for each instrument computed from data acquired using the wired motion capture system A MoCapGroup node, containing the raw 6DOF (position & orientation) data acquired using the Polhemus wired motion capture system A MoCapGroup node, containing the upper body marker data, instrument marker data, and bow marker data acquired using the Qualisys optical motion capture system As a result of errors during the recording (especially in regard to the optical motion capture software and the video camera), some datapacks may not contain all of the above nodes. Audio recordings, score-performance alignment, instrumental gesture features and raw motion capture data from the wired system are present in all datapacks. Data Curation This is a backup copy of the dataset hosted on the RepoVizz platform. The copy has been made in order to provide QUARTET DATASET with more visibility, accessibility and security. This copy excludes the audio descriptors computed in the original dataset with essentia library. Copy made in May 2020. Data curators: Pedro Ramoneda, Marius Miron and Esteban Maestre.