Find research datasets worth reusing
Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.
112
datasets available to search
ShareScore release 0.9.0
Dataset results
112 results for “music dataset”
MSMD - Multimodal Sheet Music Dataset
<p>MSMD is a synthetic dataset of 497 pieces of (classical) music that contains both audio and score representations of the pieces aligned at a fine-grained level (344,742 pairs of noteheads aligned to their audio/MIDI counterpart). It can be used for training and evaluating multimodal models that enable crossing from one modality to the other, such as retrieving sheet music using recordings or following a performance in the score image.</p> <p>Please find further information and a corresponding Python package on this Github page: <a href="https://github.com/CPJKU/msmd">https://github.com/CPJKU/msmd</a></p> <p>If you use this dataset, please cite:<br> [1] Matthias Dorfer, Jan Hajič jr., Andreas Arzt, Harald Frostel, Gerhard Widmer.<br> <a href="https://transactions.ismir.net/articles/10.5334/tismir.12/">Learning Audio-Sheet Music Correspondences for Cross-Modal Retrieval and Piece Identification</a> (<a href="https://transactions.ismir.net/articles/10.5334/tismir.12/galley/8/download/">PDF</a>).<br> Transactions of the International Society for Music Information Retrieval, issue 1, 2018.</p>
D-PLACE dataset derived from Bertolo et al. 2023 'Cross-cultural music corpus: The Expanded Natural History of Song Discography'
<p>Cite the source of the dataset as:</p> <blockquote> <p>Mila Bertolo, Martynas Snarskis, Manvir Singh, & Samuel Mehr. (2023, August 8). Cross-cultural music corpus: The Expanded Natural History of Song Discography. Zenodo. https://doi.org/10.5281/zenodo.8378337</p> </blockquote>
Music Streams Labelled with Listening Situation - [User/Track/Device/Situation] Dataset
<p>This is a contextual music dataset labeled with the listening situation associated with each stream. Each stream is composed of the user, track, and device data labelled with a situation. The dataset is collected from Deezer for the period of August 2019 from France and Brazil. The dataset is composed of 3 subsets of situations corresponding to 4, 8, and 12 different situations. The situations are extracted based on keyword matching with the associated playlist title in the Deezer catalog. The full set of situational tags are: "<strong>work, gym, party, sleep | morning, run, night, dance | car, train, relax, club"</strong>.</p> <p>Each instance contains the track/user/deviice triplets, and a situational tag indicating that this user listens to the track in the associated situation wth the corresponding data recieved from the device. The device data contain: "l<strong>inear-time, linear-day, circular-time X, circular-time Y,circular-day X, circular-day Y, device-type, network-type</strong>". The users are represented as <strong>embeddings</strong> based on their listening history computed through the matrix factorization of the user/track matrix. Additionally, the users are also represented with their demographic data of : "<strong>age, country, gender</strong>".</p> <p>The creation of the dataset and our experimental results are described in the paper: Karim M. Ibrahim, Elena V. Epure, Geoffroy Peeters, and Gaël Richard. "Audio Autotagging as Proxy for Contextual MusicRecommendation" [Under Revision]. The source code of the paper is available here: <a href="https://github.com/KarimMibrahim/Situational_Session_Generator.git">https://github.com/KarimMibrahim/Situational_Session_Generator.git</a></p> <p>The dataset is composed of the media_id which is the ID of the track in the Deezer catalog. The 30 seconds track previews used to train the model in the paper can be accessed through the Deezer API: <a href="https://developers.deezer.com/api">https://developers.deezer.com/api</a>. Each user is represented with an <strong>anonymized</strong> <strong>user_id</strong> which is associated with the user embedding available in the user_embeddings.npy file. Note: The index of the embeddings in the user_embeddings arrary corresponds to the user_id, i.e. user_id = 100 have its embeddings at user_embeddings[100]. </p> <p>Finally, the dataset also contains the splits used in our experiments. Our splits were conditioned by one of three conditions: <em>ColdTrack</em> (no overlap of tracks between the splits), <em>ColdUser</em> (no overlap of users between the splits), and <em>WarmCase</em> (overlaps allowed). Each condition is split into 4 subsets for cross-validation marked with a "<strong>fold</strong>" number in each condition. </p>
Background music and cognitive task performance: systematic review dataset
<p>This repository contains the raw data used for a systematic review of the impact of background music on cognitive task performance (Cheah et al., 2022). Our intention is to facilitate future updates to this work.</p><p><strong>Contents description</strong></p><p>This repository contains eight Microsoft Excel files, each containing the synthesised data pertaining to each of the six cognitive domains analysed in the review, as well as task difficulty, and population characteristics:</p><ul><li><i>raw-data-attention</i></li><li><i>raw-data-inhibition</i></li><li><i>raw-data-language</i></li><li><i>raw-data-memory</i></li><li><i>raw-data-thinking</i></li><li><i>raw-data-processing-speed</i></li><li><i>raw-data-task-difficulty</i></li><li><i>raw-data--population</i></li></ul><p><strong>Files description</strong></p><p><i>Tabs organisation</i></p><p>The files pertaining to each cognitive domain include individual tabs for each cognitive task analysed (c.f. Figure 2 in the original paper for the list of cognitive tasks). The file with the population characteristics data also contains separate tabs for each characteristic (extraversion, music training, gender, and working memory capacity).</p><p><i>Tabs contents</i></p><p>In all files and tabs, each row corresponds to the data of a test. The same article can have more than one row if it reports multiple tests. For instance, the study by Cassidy and MacDonald (2007; cf. <i>Memory.xlsx</i>, tab: <i>Memory-all</i>) contains two experiments (immediate and delayed free recall) each with multiple test (immediate free recall: tests 25 – 32; delayed free recall: tests 58 – 61). Each test (one per row), in this experiment, pertains to comparisons between conditions where the background music has different levels of arousal, between groups of participants with different extraversion levels, between different tasks material (words or paragraphs) and different combinations of the previous (e.g., high arousing music vs silence test among extraverts whilst completing an immediate free recall task involving paragraphs; cf. test 30).</p><p>The columns are organised as follows:</p><ul><li>"TESTS": the index of the test in a particular tab (for easy reference);</li><li>"ID": abbreviation of the cognitive tasks involved in a specific experiment (see glossary for meaning);</li><li>"REFERENCE": the article where the data was taken from (see main publications for list of articles);</li><li>"CONDITIONS": an abbreviated description of the music condition of a given test;</li><li>"MEANS (music)": the average performance across all participants in a given experiment with background music;</li><li>"MEANS (silence)": the average performance across all participants in a given experiment without background music.</li></ul><p>Then, in horizontal arrangement, we also include groups of two columns that breakdown specific comparisons related to each test (i.e., all tests comparing the same two types of condition, e.g., L-BgM vs I-BgM, will appear under the same set of columns). For each one, we indicate mean difference between the respective conditions ("MD" column) and the direction of effect ("Standard Metric" column). Each file also contains a "Glossary" tab that explains all the abbreviations used in each document.</p><p><strong>Bibliography</strong></p><p>Cheah, Y., Wong, H. K., Spitzer, M., & Coutinho, E. (2022). Background music and cognitive task performance: A systematic review of task, music and population impact. <i>Music & Science, 5</i>(1), 1-38. <a href="https://doi.org/10.1177/0305735699272005">https://doi.org/10.1177/20592043221134392</a></p>
Indian Art Music Raga Recognition Dataset (features)
<p>The <strong>Rāga Recognition Datasets (features)</strong> comprise two sizable datasets, one for each music tradition: the <strong>Carnatic Music Dataset (CMD)</strong> and the <strong>Hindustani Music Dataset (HMD)</strong>. Each dataset entry includes features such as <strong>pitch</strong>, <strong>tonic</strong>, and <strong>nyas</strong> and <strong>tani</strong> segments. These datasets can be used to develop and evaluate approaches for automatic rāga recognition in Indian art music. To the best of our knowledge, they are the largest and most comprehensive datasets (in terms of available metadata) ever used for studying this task.</p> <p>This repository only contains the metadata and computed features for the dataset, and shared in open access. To get the audio, please refer <a href="https://zenodo.org/records/7278511" target="_blank" rel="noopener">to this zenodo entry</a> and submit your request.</p> <p> </p> <p>Please cite the following publications if you use the material shared here in your research work.</p> <blockquote> <p>Gulati, S., Serrà, J., Ganguli, K. K., ¸Sentürk, S., & Serra, X. (2016). Time-delayed melody surfaces for raga recognition. In Proceedings of the 17th International Society for Music Information Retrieval Conference (ISMIR), pp. 751–757. New York, USA. [<a href="http://hdl.handle.net/10230/33117">Postprint PDF</a>]</p> </blockquote> <blockquote> <p>Gulati, S., Serrà, J., Ishwar, V., ¸Sentürk, S., & Serra, X. (2016). Phrase-based raga recognition using vector space modeling. In Proceedings of the 41st IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 66–70. Shanghai, China. [<a href="http://hdl.handle.net/10230/32879">Postprint PDF</a>]</p> </blockquote> <p> </p> <h2>Annotation Format</h2> <p>We provide both tsv files and json files that contain information about each audio recording in terms of its mbid, the path of the audio/feature files and the associated rāga identifier. Each rāga is assigned a unique identifier by Dunya, which is similar to the mbid in terms of purpose. We also provide a mapping of the rāga id to its transliterated name.</p> <h2>Mirdata</h2> <p>This dataset is included in <a href="https://github.com/mir-dataset-loaders/mirdata">mirdata</a>. Use the following code snippet to access the dataset in mirdata.</p> <pre><code># Import midata import mirdata # Initialize dataset dataset_name = 'compmusic_raga' data_home = 'mirdata/dataset' dataset = mirdata.initialize(dataset_name, data_home=data_home) # Download dataset dataset.download() # Validate dataset dataset.validate() # Load dataset as a dictionary with track ids as keys and track objects as values data = dataset.load_tracks()</code></pre> <p>In order to load the audio files in mirdata, they must be requested beforehand and placed in the data home directory.</p> <h2>Contact </h2> <p>If you have any questions or comments about the dataset, please feel free to email:</p> <p><a href="mailto:mtg-info@upf.edu">mtg-info@upf.edu</a></p> <p> </p>
MuMu: Multimodal Music Dataset
<p>MuMu is a Multimodal Music dataset with multi-label genre annotations that combines information from the Amazon Reviews dataset and the Million Song Dataset (MSD). The former contains millions of album customer reviews and album metadata gathered from Amazon.com. The latter is a collection of metadata and precomputed audio features for a million songs. </p> <p>To map the information from both datasets we use MusicBrainz. This process yields the final set of 147,295 songs, which belong to 31,471 albums. For the mapped set of albums, there are 447,583 customer reviews from the Amazon Dataset. The dataset have been used for multi-label music genre classification experiments in the related publication. In addition to genre annotations, this dataset provides further information about each album, such as genre annotations, average rating, selling rank, similar products, and cover image url. For every text review it also provides helpfulness score of the reviews, average rating, and summary of the review. </p> <p>The mapping between the three datasets (Amazon, MusicBrainz and MSD), genre annotations, metadata, data splits, text reviews and links to images are available here. Images and audio files can not be released due to copyright issues.</p> <ul> <li>MuMu dataset (mapping, metadata, annotations and text reviews)</li> <li>Data splits and multimodal feature embeddings for ISMIR multi-label classification experiments </li> </ul> <p>These data can be used together with the Tartarus deep learning library https://github.com/sergiooramas/tartarus.</p> <p>NOTE: This version provides simplified files with metadata and splits.</p> <p><strong>Scientific References</strong></p> <p>Please cite the following papers if using MuMu dataset or Tartarus library.</p> <p>Oramas, S., Barbieri, F., Nieto, O., and Serra, X (2018). Multimodal Deep Learning for Music Genre Classification, Transactions of the International Society for Music Information Retrieval, V(1).</p> <p>Oramas S., Nieto O., Barbieri F., & Serra X. (2017). Multi-label Music Genre Classification from audio, text and images using Deep Features. In Proceedings of the 18th International Society for Music Information Retrieval Conference (ISMIR 2017). https://arxiv.org/abs/1707.04916</p> <p> </p>
Dataset: Music Industry Professionals' Perspectives on Music Streaming Services and Recommendation
<p><strong>Questionnaire response data set</strong><br> Here, we include the data retrieved from participants at Eurosonic Noorderslag 2023, as described in the paper cited above.<br> When using, analyzing, or publishing this data in any way, please make sure to attribute it to the authors and cite it accordingly.<br> <br> We include the data in .xlsx, .csv format (semicolon-separated, and .tsv format (tab-separated). We suggest using the Excel file, as its layout makes it more easily readable.<br> <br> The complete question list as used in the questionnaire is published separately on <a href="https://doi.org/10.5281/zenodo.8121151">https://doi.org/10.5281/zenodo.8121151</a>.<br> <br> <strong>Paper title</strong><br> Looking at the FAccTs: Exploring Music Industry Professionals’ Perspectives on Music Streaming Services and Recommendations<br> <br> <strong>Paper abstract</strong><br> Music recommender systems, commonly integrated into streaming services, help listeners find music. Previous research on such systems has focused on providing the best possible recommendations for these services' consumers, as well as on fairness for artists who release their music on streaming services. While those insights are imperative, another group of stakeholders has been omitted so far: the many other professionals working in the music industry. They, too, are (in)directly affected by music streaming services. Therefore, this work explores the perspective of music industry professionals. We present a study that addresses the role of streaming services and recommender systems in their jobs. Results indicate this role is significant. Furthermore, participants feel that music recommender systems lack transparency and are insufficiently controllable, for both customers and artists. Finally, participants desire that music streaming services take charge of increasing recommendation diversity, and variety in consumers' listening behavior and taste.</p> <p><strong>Citation</strong><br> Karlijn Dinnissen, Isabella Saccardi, Marloes Vredenborg, and Christine Bauer. 2023. Looking at the FAccTs: Exploring Music Industry Professionals’ Perspectives on Music Streaming Services and Recommendations. In 2nd International Conference of the ACM Greek SIGCHI Chapter (CHIGREECE 2023), September 27–28, 2023, Athens, Greece. ACM, New York, NY, USA, 5 pages. https://doi.org/10.1145/3609987.3610011</p>
Datasets from the RecSys 2020 article "Carousel Personalization in Music Streaming Apps with Contextual Bandits"
<p>We publicly release the anonymized <em>user_features.csv</em> and <em>playlist_features.csv</em> datasets, from the music streaming platform Deezer, as described in the article "<em>Carousel Personalization in Music Streaming Apps with Contextual Bandits"</em> published in the proceedings of the 14th ACM Conference on Recommender Systems (<em>RecSys 2020</em>). The paper is available <a href="https://arxiv.org/abs/2009.06546">here</a>.</p> <p>These datasets are used in the GitHub repository <a href="https://github.com/deezer/carousel_bandits">deezer/carousel_bandits</a> to reproduce experiments from the article.</p> <p>Please cite our paper if you use our code or data in your work.</p>
Bipolar EEG dataset - music
<p>Dataset setting out to investigate neural responses to continuous musical pieces with bipolar EEG. Analysis code (and usage instructions) to derive neural responses to the temporal fine structure of the stimuli is <a href="https://github.com/octaveEtard/EEGmusic2020">on Github</a>. The EEG data processed to this end is provided here, as well as the raw data to enable different analyses (e.g. slower cortical responses).</p> <p><strong># Introduction</strong></p> <p>This dataset contains bipolar scalp EEG responses of 17 subjects listening to continuous musical pieces (Bach's Two-Part Inventions), and performing a vibrato detection task.</p> <p>Naming conventions:</p> <p>- The subject IDs are <code>EBIP01, EBIP02 ... EBIP17</code>.<br> - The different conditions are labelled to indicate the instrument that was being attended: <code>fG</code> and <code>fP</code> for the Guitar and Piano in quiet (Single Instrument (SI) conditions), respectively; and <code>fGc</code> and <code>fPc</code> for Competing conditions where both the instruments are playing together, but where the subjects should be selectively attending to the Guitar or Piano, respectively (Competing Instrument (CI) conditions).<br> - An appended index from 2 to 7 designates the invention that was played (index 1 corresponds to the training block for which no EEG data was recorded). Note that this index does not necessarily corresponds to the order in which the stimuli were played (order was pseudo-randomised).<br> <br> For example, the EEG file named <code>EBIP08_fGc_4</code> contains EEG data from subject <code>EBIP08</code> performing the competing instrument task (CI condition), attending to the guitar (ignoring the piano), and the stimulus that was played was the invention #4.</p> <p><strong># Content</strong></p> <p>The general organisation of the dataset is as follow:<br> <br> <code>data</code><br>  <code>├─── behav</code>              <em>folder containing the behavioural data</em><br>  <code>├─── EEG</code>            <em>folder containing the EEG data</em><br>  <code>│   ├─── processed</code><br>  <code>│   └─── raw</code><br>  <code>├─── linearModelResults</code>       <em>folder containing the results from the analysis code</em><br>  <code>└─── stimuli</code> <em>folder containing the stimuli</em><br>      <code>├─── features</code><br>      <code>├─── processedInventions</code><br>      <code>└─── rawInventions</code></p> <p>This general organisation is the one expected by the code. The location of the <code>data</code> folder and/or these main folders can be personalised in the <code>functions/+EEGmusic2020/getPath.m</code> function in the Github repository. The architecture of the sub-folders in each of these folders is specified by the functions <code>makePathEEGFolder</code>, <code>makePathFeatureFiles</code> and <code>makePathSaveResults</code>. The naming of the files within them is implemented by <code>makeNameEEGDataFile</code> and <code>makeNameEEGDataFile</code> (all these functions being in <code>functions/+EEGmusic2020</code>).</p> <p> </p> <p>- The <code>behav</code> folder is structured as follow:<br> <br> <code>behav</code><br> <code>├─── EBIP02</code><br> <code>│ ├─── EBIP02_keyboardInputs_fGc_2.mat</code> <em>file containing variables:</em><br> <code>│ │ ├─── timePressed</code> <em>key press time (in seconds, relative to stimulus onset)</em><br> <code>│ │ └─── keyCode</code> <em>ID of the keys that were pressed</em><br> <code>│ └─── ...</code><br> <code>├─── ...</code><br> <code>├─── vibTime </code><br> <code>│ ├─── vibTime_2.mat</code> <em>file containing variables:</em><br> <code>│ │ ├─── idxNoteVib</code> <em>index (in the MIDI files) of the notes in which vibratos were inserted</em><br> <code>│ │ ├─── instrumentOrder</code> <em>order of the instruments in <code>idxNoteVib</code> and <code>vibTiming</code> variables</em><br> <code>│ │ └─── vibTiming</code> <em>timing of vibrato onsets in the track (in s)</em><br> <code>│ ├─── ...</code><br> <code>└─── clickPerformance_RT_2.0.mat</code> <em>file containing behavioural results for all subjects (FPR, TPR, etc.):</em></p> <p><code>instrumentOrder</code> indicates to what instrument each column of <code>idxNoteVib</code> and <code>vibTiming</code> refers to. The data for <code>EBIP01</code> missing due to a technical error.</p> <p> </p> <p>- The <code>EEG/raw</code> folder contains unprocessed EEG data for all subjects, and files indicating the order in which the inventions were played. It is structured as follow:<br> <br> <code>EEG</code><br> <code>├─── raw</code><br> <code>│ ├─── EBIP01</code><br> <code>│ │ ├─── EBIP01_EEGExpParam.mat</code> <em>file containing variables:</em><br> <code>│ │ │ ├─── conditionOrder </code> <em>whether this subject started by listening to the guitar or piano</em><br> <code>│ │ │ └─── partsOrder </code> <em>order in which the inventions were presented to this subject</em><br> <code>│ │ ├─── EBIP01_fGc_2.[eeg/vhdr/vmrbk]</code> <em>raw EEG data files</em><br> <code>│ │ ├─── ...</code><br> <code>│ ├─── ...</code></p> <p>The <code>conditionOrder</code> variable can assume two values: either <code>{'fG','fP'}</code> indicating the subject started by listening to the guitar or <code>{'fP','fG'}</code> indicating the subject started by listening to the piano. The <code>partsOrder</code> variable is a 2 x 6 matrix containing the indices (2 to 7) of the inventions that were played, ordered in the presentation order. During the first block, the instrument <code>conditionOrder{1}</code> was attended, and the invention # <code>partsOrder(1,1)</code> was played. During the second block, the instrument <code>conditionOrder{2}</code> was attended, and the invention <code>#partsOrder(2,1)</code> was played, etc.</p> <p>Each EEG files contains 3 channels: 2 are the bipolar electrophysiological channels, and one (labelled <code>Sound</code>) contains a recording of the stimuli that were played and that was simultaneously recorded at the same sampling rate as the EEG data (5 kHz) by the amplifier through an acoustic adapter. The files also contain triggers that indicate the beginning and end of the stimuli (labelled <code>S 1</code> and <code>S 2</code> respectively). The sound channel and triggers can be used to temporally align the EEG data and stimuli.</p> <p> </p> <p>The <code>EEG/processed</code> folder contains processed EEG data for all subjects, as required for the analyses carried out in the <a href="https://github.com/octaveEtard/EEGmusic2020">code.</a> It is organised as follow:<br> <br> <code>EEG</code><br> <code>├─── processed</code><br> <code>│ └─── Fs-5000</code> <em>sampling rate</em><br> <code>│ └─── HP-130</code> <em>processing that was applied</em><br> <code>│ │ ├─── EBIP01</code><br> <code>│ │ │ ├─── ...</code> <em>processed EEG data files</em><br> <code>│ │ ├─── ...</code><br> <code>│ └─── noProc</code><br> <code>│ ├─── ...</code></p> <p>This structure is specified by the <code>makePathEEGFolder</code> function, and the file names by <code>makeNameEEGDataFile</code>. In the files in the <code>noProc</code> folder, the EEG data was simply aligned with the stimuli, but is otherwise unprocessed. Events were added to mark stimulus onset and offset (labelled <code>stimBegin</code> and <code>stimEnd</code>). In the other folders, the EEG data was furthermore high-pass filtered at 130 Hz (HP-130).</p> <p> </p> <p>- The <em>linearModelResults</em> folder contains the results from the linear model analyses:<br> <br> <code>linearModelResults</code><br> <code>└─── Fs-5000</code> <em>sampling rate</em><br> <code>│ └─── HP-130</code> <em>processing of the EEG data</em><br> <code>│ │ └─── LP-2000</code> <em>processing of the stimulus feature</em><br> <code>│ │ ├─── ...</code> <em>result files</em><br> <code>│ ├─── ...</code></p> <p>This structure and file names are specified by the <code>makePathSaveResults</code> function.</p> <p> </p> <p>- The <code>rawInventions</code> folder contains the orignal data that was used to construct the stimuli:<br> <br> <code>rawInventions</code><br> <code>├─── invent1</code> <em>invention index</em><br> <code>│ ├─── invent1_60bpm.mid</code> <em>MIDI file</em><br> <code>│ ├─── invent1_60bpm_guitar.wav </code> <em>guitar track</em><br> <code>│ └─── invent1_60bpm_piano.wav </code> <em>piano track</em><br> <code>│</code><br> <code>├─── ...</code></p> <p>In this folder (and <strong>only </strong>in this folder), the numbering of the inventions differs from the one otherwise used throughout. The correspondence is as shown below:<br> Raw invention # | Feature, etc. #<br> 1, 2, 3, 4 -> 1, 2, 3, 4<br> 7, 8, 9 -> 5, 6,7</p> <p> </p> <p>- The <code>processedInventions</code> contains invention waveforms that have been transformed. The instrument and invention index are indicated by a suffix in the file names ('G': guitar, 'P': piano). 'zv' indicates that the vibratos were replaced by zeros. 'noOnset30ms' indicates that the onset of the notes was suppressed in a 30 ms window.</p> <p> </p> <p>- The <code>features</code> folder contains specific features of the stimuli for use in the models:<br> <br> <code>features</code><br> <code>└─── Fs-5000</code> <em>sampling rate of the feature</em><br> <code> └─── LP-2000</code> <em>processing of the feature </em><br> <code> ├─── waveform </code> <em>feature name (here: stimulus waveform)</em><br> <code> │ ├─── ... </code> <em>feature files</em><br> <code> └─── WNO</code> <em>Waveform No Onset (stimulus waveform with note onsets removed)</em><br> <code> ├─── ... </code></p> <p>The naming convention is as highlighted above for the <code>processedInventions</code> folder. SI conditions correspond to 'G' & 'P' files, and CI conditions to 'PG' files. In the latter case, 'fG' indicates the attended instrument is the guitar and 'fP' the piano.<br> These files notably contain the variables <code>attended</code> and <code>ignored</code> that contains the feature for the attended and ignored instruments (the <code>ignored</code> field is only present in the CI conditions).<br> Note that a pair of two files corresponding to the same invention in a CI condition ('PGfG' & 'PGfP') effectively contain the same information with the <code>attended</code> and <code>ignored</code> variables flipped.</p>
CROCUS: Dataset of Musical Performance Critique
<p>CROCUS (CRitique dOCUmentS): Dataset of Musical Performance Critique Documents (in Japanese) CC BY-NC-ND 4.0</p> <p>This open dataset contains 90 musical performances and 239 critiques of those performances.</p> <p>For more information, please visit the project page below.<br> <a href="https://masaki-cb.github.io/crocus/">https://masaki-cb.github.io/crocus/</a></p> <p>--</p> <p>Naming Rule:<br> For Recording Data<br> PieceID-PlayerID.wav<br> e.g. Perfomance of beethoven's piece by player 1<br> n01-bee-sym3-p01.wav</p> <p>For Critique Data<br> PieceID-PlayerID-CriticID.wav<br> e.g. Critique by teacher 3 on perfomance of beethoven's piece by player 1<br> n01-bee-sym3-p01-c03.txt</p> <p>List of Piece<br> PieceID Composer Title<br> n01-bee-sym3 L. v. Beethoven Symphony No.3 in E flat Major 'Eroica', Op.55<br> n02-ros-silk G. A. Rossini La scala di seta, Overture<br> n03-sch-sym8 F. Schubert Symphony No.8 in B Minor, D.759 'Unfinished'<br> n04-bra-vcon J. Brahms Violin Concerto in D Major, Op.77<br> n05-tch-sym4 P. I. Tchaikovsky Symphony No.4 in f Minor, Op.36<br> n06-tch-swan P. I. Tchaikovsky "The Swan Lake", Ballet Suite, Op.20a<br> n07-rim-sche N. Rimsky-Korsakov "Scheherazade", Symphonic Suite, Op.35<br> n08-str-donj R. Strauss "Don Juan", Symphonic Poem, Op.20<br> n09-rav-tomb M. Ravel Le Tombeau de Couperin I.Prelude<br> n10-pro-pete S. Prokofiev "Peter and the Wolf", Symphonic Tale, Op.67</p>
Shared Acoustic Codes Underlie Emotional Communication in Music and Speech - Evidence from Deep Transfer Learning (Datasets)
<p>This repository contains the datasets used in the article "Shared Acoustic Codes Underlie Emotional Communication in Music and Speech - Evidence from Deep Transfer Learning" (Coutinho & Schuller, 2017). </p> <p>In that article four different data sets were used: SEMAINE, RECOLA, ME14 and MP (acronyms and datasets described below). The SEMAINE (speech) and ME14 (music) corpora were used for the unsupervised training of the Denoising Auto-encoders (domain adaptation stage) - only the audio features extracted from the audio files in these corpora were used and are provided in this repository. The RECOLA (speech) and MP (music) corpora were used for the supervised training phase - both the audio features extracted from the audio files and the Arousal and Valence annotations were used. In this repository, we provide the audio features extracted from the audio files for both corpora, and Arousal and Valence annotations for some of the music datasets (those that the author of this repository is the data curator).</p> <p>Below, you can find description of the various corpora, the details about the data stored in this repository and information on how to obtain the rest of the data used by Coutinho and Schuller (2017).</p> <p><strong>SEMAINE (speech)</strong></p> <p>The SEMAINE corpus (McKeown, Valstar, Cowie, Pantic & Schroder, 2012) was developed specifically to address the task of achieving emotion-rich interactions, and it is adequate for this task as it comprises a wide range of emotional speech. It includes video and speech recordings of spontaneous interactions between human and emotionally stereotyped `characters'. Coutinho & Schuller (2017) used a subset of this database (called <em>Solid-SAL</em>). The <em>Solid-SAL</em> dataset is freely available for scientific research purposes (see http://semaine-db.eu). This repository includes the audio features used in Coutinho & Schuller (2017) (under features/SEMAINE).</p> <p><strong>RECOLA (speech)</strong></p> <p>The RECOLA database (Ringeval, Sonderegger, Sauer & Lalanne, 2013) consists of multimodal recordings (audio, video, and peripheral physiological activity) of spontaneous dyadic interactions between French adults. Coutinho & Schuller (2017) used the RECOLA-Audio module which consists of the audio recordings of each participant in the dyadic phase of the task. In particular, they used the non-segmented high-quality audio signals (WAV format, 44.1kHz, 16bits), obtained through unidirectional headset microphones, of the first five minutes of each interaction. Annotations consist of time-continuous ratings of the level of Arousal and Valence dimensions of emotion perceived by each rater while seeing and listening the audio-visual recordings of each participant task. The publicly available annotated dataset includes only part of the data which amounts to a total number of 23 instances. The time frame length used by Coutinho & Schuller (2017) is 1s (the original annotations were downsampled). This repository includes the audio features used in Coutinho & Schuller (2017) (under features/RECOLA). To obtain the annotations you should contact the author of the original study (see https://diuf.unifr.ch/diva/recola/download.html for further details).</p> <p><strong>ME14 (music)</strong></p> <p>The MediaEval ``Emotion in Music'' task is dedicated to the estimation of Arousal and Valence scores continuously in time and value for song excerpts from the Free Music Archive. Coutinho and Schuller (2017) used the whole corpus (development and test sets for the 2014 challenge) which includes 1,744 songs belonging to 11 musical styles -- Soul, Blues, Electronic, Rock, Classical, Hip-Hop, International, Folk, Jazz, Country, and Pop (maximum of five songs per artist). This repository includes the audio features used in Coutinho & Schuller (2017) (under features/ME14). The full dataset (including annotations) can be obtained from http://www.multimediaeval.org/mediaeval2014/emotion2014/.</p> <p><strong>MP (music)</strong></p> <p>This is a corpus compiled specifically for this work described in Coutinho & Schuller (2017) using data collected in four previous studies. It consists of emotionally diverse full music pieces from a variety of musical styles (Classical and contemporary Western Art, Baroque, Bossa Nova, Rock, Pop, Heavy Metal, and Film Music). Annotations were obtained in controlled laboratory experiments whereby the emotional character of each piece was evaluated time-continuously in terms of levels of Arousal and Valence perceived by listeners (ranging between 35 to 52 in the four studies). In what follows, some details about the various studies are described.</p> <ul> <li>MP<sub>DB1</sub>: This subset of the MP corpus consists of the data reported by Korhonen (2004), and gently made available by the author. This dataset includes six full (or long excerpts) music pieces ranging from 151s to 315s in length (only classical music). Each piece was annotated by 35 participants (14 females). The time series correspondents to each music piece were collected at 1Hz. The golden standard for each piece was computed by averaging the individual time series across all raters. This repository includes the audio features used in Coutinho & Schuller (2017) (under features/MP/DB1). To obtain the labels please contact the author of the original study.</li> <li>MP<sub>DB2</sub>: The dataset by Coutinho & Cangelosi (2011) includes 9 full pieces (43s to 240s long) of classical music (romantic repertoire) annotated by 39 subjects (19 females). Values were recorded every time the mouse was moved with a precision of 1 ms. The resultant timeseries were then resampled (moving average) to a synchronous rate of 1 Hz. The golden standard for each piece was computed by averaging the individual time series across all raters. This repository includes the audio features (under features/MP/DB2) and labels (under annotations/MP/DB2) used in Coutinho & Schuller (2017).</li> <li>MP<sub>DB3</sub>: This dataset was collected by Coutinho & Dibben (2012) and it consists of 8 pieces of film music (84s to 130s long) taken from the late 20th century Hollywood film repertoire. Emotion ratings were given by 52 participants (26 females). The annotation procedure, data processing, and golden standard calculations were identical to MP<sub>DB2</sub>. This repository includes the audio features (under features/MP/DB3) and labels (under annotations/MP/DB3) used in Coutinho & Schuller (2017).</li> <li>MP<sub>DB4</sub>: This dataset was collected by Grewe, Nagel, Kopiez and Altenmüller (2007), and gently made available by the authors. It includes seven music pieces (127s to 502s in length) of heterogeneous styles (e.g., Rock, Pop, Heavy Metal, Classical). Each music piece was annotated by 38 participants (29 females) using an identical methodology to MP<sub>DB2</sub> and MP<sub>DB3</sub>. Data processing and golden standard calculations were also identical. This repository includes the audio features (under features/MP/DB4) used in Coutinho & Schuller (2017). To obtain the labels contact the authors of the original study</li> </ul> <p> </p> <p><strong>Bibliography</strong></p> <p>Coutinho, E., & Cangelosi, A. (2011). Musical emotions: predicting second-by-second subjective feelings of emotion from low-level psychoacoustic features and physiological measurements. <em>Emotion</em>, <em>11</em>(4), 921.</p> <p>Coutinho, E., & Dibben, N. (2013). Psychoacoustic cues to emotion in speech prosody and music. <em>Cognition & Emotion</em>, <em>27</em>(4), 658-684.</p> <p>Coutinho E, Schuller B (2017) Shared acoustic codes underlie emotional communication in music and speech—Evidence from deep transfer learning. PLoS ONE 12(6): e0179289. https://doi. org/10.1371/journal.pone.0179289.</p> <p>Grewe, O., Nagel, F., Kopiez, R., Altenmüller, E. (2007). Emotions over time: synchronicity and development of subjective, physiological, and facial affective reactions to music. <em>Emotion, 7</em>(4), pp. 774-788. DOI: 10.1037/1528-3542.7.4.774.</p> <p>Korhonen, M. (2004). Modeling Continuous Emotional Appraisals of Music Using System Identification. Available from: http://hdl.handle.net/10012/879.</p> <p>McKeown, G., Valstar, M., Cowie, R., Pantic, M., Schroder, M. (2012). The SEMAINE Database: Annotated Multimodal Records of Emotionally Colored Conversations between a Person and a Limited Agent. <em>IEEE Transactions on Affective Computing</em>, 3, pp. 5-17. DOI: http://doi.ieeecomputersociety.org/10.1109/T-AFFC.2011.20.</p> <p>Ringeval, F., Sonderegger, A., Sauer, J. & Lalanne, D. (2013). Introducing the RECOLA Multimodal Corpus of Remote Collaborative and Affective Interactions. In <em>Proceedings of the 2nd International Workshop on Emotion Representation, Analysis and Synthesis in Continuous Time and Space (EmoSPACE 2013)</em>, Shanghai, China. IEEE</p>
PDMX: A Large-Scale Public Domain MusicXML Dataset for Symbolic Music Processing
<p>We introduce <strong>PDMX</strong>: a <strong>P</strong>ublic <strong>D</strong>omain <strong>M</strong>usic<strong>X</strong>ML dataset for symbolic music processing. Refer to our <a title="PDMX Paper" href="https://arxiv.org/abs/2409.10831" target="_blank" rel="noopener">paper</a> for more information, and our <a title="PDMX GitHub Repository" href="https://github.com/pnlong/PDMX/" target="_blank" rel="noopener">GitHub repository</a> for any code-related details. Please cite both our paper and <a href="https://arxiv.org/abs/2410.02084" target="_blank" rel="noopener">our collaborators' paper</a> if you use this dataset (see our GitHub for more information).</p> <p>Upon further use of the PDMX dataset, we discovered a discrepancy between the public-facing copyright metadata on the <a href="https://musescore.com/">MuseScore website</a> and the internal copyright data of the MuseScore files themselves, which affected 31,221 (12.29% of) songs. We have decided to proceed with the former given its public visibility on Musescore (i.e. this is what the MuseScore website presents its users with). We have noted files with conflicting internal licenses in the <em><strong>license_conflict</strong></em> column of PDMX. We recommend using the <em><strong>no_license_conflict</strong></em> subset of PDMX (which still includes 222,856 songs) moving forward.</p> <p>Additionally, for each song in PDMX, we not only provide the <em>MusicRender</em> and metadata JSON files, but we also try to include the associated compressed MusicXML (MXL), sheet music (PDF), and MIDI (MID) files when available. Due to the corruption of 42 of the original MuseScore files, these songs lack those associated files (since they could not be converted to those formats) and only include the <em>MusicRender</em> and metadata JSON files. The <em><strong>all_valid</strong></em> subset of PDMX describes the songs where all associated files are valid.</p>
Help Me study! Music Listening Habits While Studying (Dataset)
<p>This repository contains the raw data used for a research study that examined university students' music listening habits while studying. There are two experiments in this research study. Experiment 1 is a retrospective survey, and Experiment 2 is a mobile experience sampling research study.</p><p>This repository contains five Microsoft Excel files with data obtained from both experiments. The files are as follows:</p><ul><li><i>onlineSurvey_raw_data.xlsx</i></li><li><i>esm_raw_data.xlsx</i></li><li><i>esm_music_features_analysis.xlsx</i></li><li><i>esm_demographics.xlsx</i></li><li><i>index.xlsx</i></li></ul><p><strong>Files Description</strong></p><p><i><strong>File: onlineSurvey_raw_data.xlsx</strong></i></p><p>This file contains the raw data from Experiment 1, including the (anonymised) demographic information of the sample. The sample characteristics recorded are:</p><ul><li>studentship</li><li>area of study</li><li>country of study</li><li>type of accommodation a participant was living in</li><li>age</li><li>self-identified gender</li><li>language ability (mono- or bi-/multilingual)</li><li>(various) personality traits</li><li>(various) musicianship</li><li>(various) everyday music uses</li><li>(various) music capacity</li></ul><p>The file also contains raw data of responses to the questions about participants' music listening habits while studying in real life. These pieces of data are:</p><ul><li>likelihood of listening to specific (rated across 23) music genres while studying and during everyday listening.</li><li>likelihood of listening to music with specific acoustic features (e.g., with/without lyrics, loud/soft, fast/slow) music genres while studying and during everyday listening.</li><li>general likelihood of listening to music while studying in real life.</li><li>(verbatim) responses to participants' written responses to the open-ended questions about their real-life music listening habits while studying.</li></ul><p><i><strong>File: esm_raw_data.xlsx</strong></i></p><p>This file contains the raw data from Experiment 2, including the following variables:</p><ul><li>information of the music tracks (track name, artist name, and if available, Spotify ID of those tracks) each participant was listening to during each music episode (both while studying and during everyday-listening)</li><li>level of arousal at the onset of music playing and the end of the 30-minute study period</li><li>level of valence at the onset of music playing and the end of the 30-minute study period</li><li>specific mood at the onset of music playing and the end of the 30-minute study period</li><li>whether participants were studying</li><li>their location at that moment</li><li>(if studying) whether they were studying alone</li><li>(if studying) the types of study tasks</li><li>(if studying) the perceived level of difficulty of the study task</li><li>whether participants were planning to listen to music while studying</li><li>(various) reasons for music listening</li><li>(various) perceived positive and negative impacts of studying with music</li></ul><p>Each row represents the data for a single participant. Rows with a record of a participant ID but no associated data indicate that the participant did not respond to the questionnaire (i.e., missing data).</p><p><i><strong>File: esm_music_features_analysis.xlsx</strong></i></p><p>This file presents the music features of each recorded music track during both the study-episodes and the everyday-episodes (retrieved from Spotify's "Get Track's Audio Features" API). These features are:</p><ul><li>energy level</li><li>loudness</li><li>valence</li><li>tempo</li><li>mode</li></ul><p>The contextual details of the moments each track was being played are also presented here, which include:</p><ul><li>whether the participant was studying</li><li>their location (e.g., at home, cafe, university)</li><li>whether they were studying alone</li><li>the type of study tasks they were engaging with (e.g., reading, writing)</li><li>the perceived difficulty level of the task</li></ul><p><i><strong>File: esm_demographics.xlsx</strong></i></p><p>This file contains the demographics of the sample in Experiment 2 (N = 10), which are the same as in Experiment 1 (see above).</p><p>Each row represents the data for a single participant. Rows with a record of a participant ID but no associated demographic data indicate that the participant did not respond to the questionnaire (i.e., missing data). </p><p><i><strong>File: index.xlsx</strong></i></p><p>Finally, this file contains all the abbreviations used in each document as well as their explanations.</p>
PodcastMix - a dataset for separating music and speech in podcasts
<p><strong>Note: due to zenodo limitations here we host solely the metadata. the whole dataset can be found at: https://drive.google.com/drive/u/0/folders/1tpg9WXkl4L0zU84AwLQjrFqnP-jw1t7z </strong></p> <p>We introduce PodcastMix, a dataset formalizing the task of separating background music and foreground speech in podcasts. It contains audio files at 44.1kHz and the corresponding metadata. For further details check the following paper and the associated GitHub repository: </p> <ul> <li>N. Schmidt, J. Pons, M. Miron, "PodcastMix - a dataset for separating music and speech in podcasts", Interspeech (2022)</li> <li>N. Schmidt, "PodcastMix - a dataset for separating music and speech in podcasts", Masters thesis, MTG, UPF (2021) https://zenodo.org/record/5554790#.YXLHvNlByWA </li> <li>https://github.com/MTG/Podcastmix</li> </ul> <p>This dataset contains four parts. Due to zenodo file size limitation we host the training dataset on google drive. We highlight the content of the zenodo archives within brackets:</p> <ul> <li>[metadata] PodcastMix-synth train: large and diverse training set that is programatically generated (with a validation partition). The mixtures are created programatically with music from Jamendo and speech from the VCTK dataset. </li> <li>[metadata] PodcastMix-synth test a programatically generated test set with reference stems to compute evaluation metrics. The mixtures are created programatically with music from Jamendo and speech from the VCTK dataset. </li> <li>[audio and metadata] PodcastMix-real with-reference : a test set with real podcasts with reference stems to compute evaluation metrics. The podcasts are recorded by one of the authors and the source of the music is the FMA dataset. </li> <li>[audio and metadata] PodcastMix-real no-reference: a test set with real podcasts with only the podcasts mixes for subjective evaluation. The podcasts are compiled from the internet. </li> </ul> <p>The training dataset, PodcastMix-synth may be found at our google drive repository: https://drive.google.com/drive/folders/1tpg9WXkl4L0zU84AwLQjrFqnP-jw1t7z?usp=sharing . The archive comprises 450GB of audio and metadata with the following structure:</p> <ul> <li>[metadata and audio] PodcastMix-synth train: large and diverse training set that is programatically generated (with a validation partition). The mixtures are created programatically with music from Jamendo and speech from the VCTK dataset. </li> <li>[metadata and audio] PodcastMix-synth test a programatically generated test set with reference stems to compute evaluation metrics. The mixtures are created programatically with music from Jamendo and speech from the VCTK dataset. </li> </ul> <p>Make sure you maintain the folder structure of the original dataset when you uncompress these files. </p> <p><br> This dataset is created by Nicolas Schmidt, Marius Miron, Music Technology Group - Universitat Pompeu Fabra (Barcelona) and Jordi Pons. This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 Unported License (CC BY-SA 4.0).</p> <p><br> Please acknowledge PodcastMix in Academic Research. When the present dataset is used for academic research, we would highly appreciate if authors quote the following publications:</p> <ul> <li>N. Schmidt, J. Pons, M. Miron, "PodcastMix - a dataset for separating music and speech in podcasts", Interspeech (2022)</li> <li>N. Schmidt, "PodcastMix - a dataset for separating music and speech in podcasts", Masters thesis, MTG, UPF (2021) https://zenodo.org/record/5554790#.YXLHvNlByWA </li> </ul> <p><br> 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 dataset or any part of it.</p> <p><br> PURPOSES. The data is processed for the general purpose of carrying out research development and innovation studies, works or projects. In particular, but without limitation, the data is processed for the purpose of communicating with Licensee regarding any administrative and legal / judicial purposes.<br> </p>
Multi-modal dataset for music genre recognition based on six different modalities for LMD-aligned and SLAC datasets
<p>Multi-modal dataset for music genre recognition based on six different modalities for the LMD-aligned [1] and SLAC [2] datasets. Further details are provided in [3].</p> <p><strong>Descriptions of files</strong></p> <table> <thead> <tr> <th scope="col">Link</th> <th scope="col">Description</th> </tr> </thead> <tbody> <tr> <td><a href="https://zenodo.org/record/5651429/files/LMD-aligned_Filelist.arff">LMD-aligned_Filelist.arff</a></td> <td>File list with 1575 music tracks selected from the LMD-aligned dataset [1] with tagtraum genre annotations [4] (only a subset of LMD-aligned is used, which includes only pieces for which all six modalities were accessible, and which includes only well-represented genres)</td> </tr> <tr> <td><a href="https://zenodo.org/record/5651429/files/LMD-aligned_ExtractedFeatures.tar.gz">LMD-aligned_ExtractedFeatures.tar.gz</a></td> <td>Raw audio signal and model-based features extracted with AMUSE [5]</td> </tr> <tr> <td><a href="https://zenodo.org/record/5651429/files/LMD-aligned_ProcessedFeatures.tar.gz">LMD-aligned_ProcessedFeatures.tar.gz</a></td> <td>Processed features: audio signal and model-based features aggregated for 4 s time frames with 2 s step size / all other features (see the table below) with the same values for all time frames</td> </tr> <tr> <td><a href="https://zenodo.org/record/5651429/files/LMD-aligned_Datasets.tar.gz">LMD-aligned_Datasets.tar.gz</a></td> <td>Training, optimization, and test datasets for 3 splits for the recognition of 5 genres in [3]</td> </tr> <tr> <td><a href="https://zenodo.org/record/5651429/files/SLAC_Filelist.arff">SLAC_Filelist.arff</a></td> <td>File list with 250 music tracks from the SLAC dataset [2] (genres and sub-genres are provided in the folder structure)</td> </tr> <tr> <td><a href="https://zenodo.org/record/5651429/files/SLAC_ExtractedFeatures.tar.gz">SLAC_ExtractedFeatures.tar.gz</a></td> <td>Raw audio signal and model-based features extracted with AMUSE [5]</td> </tr> <tr> <td><a href="https://zenodo.org/record/5651429/files/SLAC_ProcessedFeatures.tar.gz">SLAC_ProcessedFeatures.tar.gz</a></td> <td>Processed features: audio signal and model-based features aggregated for 4 s time frames with 2 s step size / all other features (see the table below) with the same values for all time frames</td> </tr> <tr> <td><a href="https://zenodo.org/record/5651429/files/SLAC_Datasets.tar.gz">SLAC_Datasets.tar.gz</a></td> <td>Training, optimization, and test datasets for 3 splits for the recognition of 5 genres and 10 sub-genres in [3]</td> </tr> </tbody> </table> <p><strong>Modalities and feature sub-groups</strong></p> <table> <thead> <tr> <th scope="col">Modality</th> <th scope="col">Sub-group</th> <th scope="col"> <p>Dimensions in processed</p> <p>features of LMD-aligned</p> </th> <th scope="col"> <p>Dimensions in processed</p> <p>features of SLAC</p> </th> </tr> </thead> <tbody> <tr> <td>Audio signal</td> <td>Low-level</td> <td>1-524</td> <td>1-524</td> </tr> <tr> <td>Audio signal</td> <td>Semantic</td> <td>525-810</td> <td>525-810</td> </tr> <tr> <td>Audio signal</td> <td>Structural complexity</td> <td>811-908</td> <td>811-908</td> </tr> <tr> <td>Model-based</td> <td>Instruments</td> <td>909-1018</td> <td>909-1018</td> </tr> <tr> <td>Model-based</td> <td>Moods</td> <td>1019-1146</td> <td>1019-1146</td> </tr> <tr> <td>Model-based</td> <td>Various</td> <td>1147-1402</td> <td>1147-1402</td> </tr> <tr> <td>Playlists</td> <td>Genres</td> <td>1403-1973</td> <td>1403-1973</td> </tr> <tr> <td>Playlists</td> <td>Styles</td> <td>1974-1695</td> <td>1974-1695</td> </tr> <tr> <td>Symbolic</td> <td>Pitch</td> <td>1696-1757</td> <td>1696-1757</td> </tr> <tr> <td>Symbolic</td> <td>Melodic</td> <td>1758-1781</td> <td>1758-1781</td> </tr> <tr> <td>Symbolic</td> <td>Chords</td> <td>1782-1836</td> <td>1782-1836</td> </tr> <tr> <td>Symbolic</td> <td>Rhythm</td> <td>1837-1935</td> <td>1837-1935</td> </tr> <tr> <td>Symbolic</td> <td>Tempo</td> <td>1936-1963</td> <td>1936-1963</td> </tr> <tr> <td>Symbolic</td> <td>Instrument presence</td> <td>1964-2441</td> <td>1964-2441</td> </tr> <tr> <td>Symbolic</td> <td>Instruments</td> <td>2442-2456</td> <td>2442-2456</td> </tr> <tr> <td>Symbolic</td> <td>Texture</td> <td>2457-2480</td> <td>2457-2480</td> </tr> <tr> <td>Symbolic</td> <td>Dynamics</td> <td>2481-2484</td> <td>2481-2484</td> </tr> <tr> <td>Album covers</td> <td>SIFT</td> <td>2485-2584</td> <td>2485-2584</td> </tr> <tr> <td>Lyrics</td> <td>jLyrics descriptors</td> <td>2585-2603</td> <td>2585-2671</td> </tr> <tr> <td>Lyrics</td> <td>Bag-of-Words</td> <td>2604-2703</td> <td> </td> </tr> <tr> <td>Lyrics</td> <td>Doc2Vec</td> <td>2704-2803</td> <td> </td> </tr> </tbody> </table>
Indian Regional Music Dataset
<p>This dataset is a collection of mel-spectrogram features extracted from Indian regional music containing the following languages:<br> Hindi, Gujarati, Marathi, Konkani, Bengali, Oriya, Kashmiri, Assamese, Nepali, Konyak, Manipuri, Khasi & Jaintia, Tamil, Malayalam, Punjabi, Telugu, Kannada.</p> <p>Five recordings are collected for each language for four artists (2Male + 2Female) each. 2 artists out of 4 for each language are old veteran performers, and the remaining 2 are contemporary performers. Overall, the dataset includes 17 languages and 68 artists (34 Males and 34 Females). There are 340 recordings in the dataset, with a total duration of 29.3 hrs.</p> <p>Mel-spectrogram is extracted from a 3-second segment with a 1/2 second sliding window for each song. Extracted mel-spectrogram for each segment is annotated with language, location, local_song_index, global_song_index, language_id, location_id, artist_id, gender_id and no_of_artists.</p> <p>_________________________________________________________________________________________________________</p> <p>This project was funded under the grant number: ECR/2018/000204 by the Science & Engineering Research Board (SERB).</p> <p> </p>
Indian Folk Music Dataset
<p>This dataset is a collection of mel-spectrogram features extracted from Indian folk music containing the following 15 folk styles:<br> Bauls, Bhavageethe, Garba, Kajri, Maand, Sohar, Tamang Selo, Veeragase, Bhatiali, Bihu, Gidha, Lavani, Naatupura Paatu, Sufi, Uttarakhandi.</p> <p>The number of recordings varies from 16 to 50 in the mentioned folk styles representing the scarcity of availability of given folk styles on the Internet. There are at least 4 artists and a maximum of 22. Overall there are 125 artists (34 female + 91 male) in these 15 folk styles. </p> <p>There is a total of 606 recordings in the dataset, with a total duration of 54.45 hrs.<br> Mel-spectrogram is extracted from a 3-second segment with each song's 1/2 second sliding window. Extracted mel-spectrogram for each segment is annotated with folk_style, state, artist, gender, song, source, no_of_artists, folk_style_id, state_id, artist_id, gender_id.<br> _________________________________________________________________________________________________________<br> This project was funded under the grant number: ECR/2018/000204 by the Science & Engineering Research Board (SERB).</p>
Indian Semi-Classical Music Dataset
<p>This dataset is a collection of mel-spectrogram features extracted from Indian semi-classical music containing the following 9 semi-classical styles:<br> Bhajan, Chaiti, Dadra, Ghazal, Kajri, Natya Sangeet, Qawwali, Tappa, Thumri.</p> <p>The number of recordings varies from 25 (for Chaiti) to 50 in the mentioned styles representing the scarcity of availability of given folk styles on the Internet. There are at least 5 artists and a maximum of 13. Overall there are 48 artists (36 female + 12 male) in these 9 semi-classical styles. <br> There is a total of 425 recordings in the dataset, with a total duration of 54.69 hrs.<br> Mel-spectrogram is extracted from a 3-second segment with each song's 1/2 second sliding window. Extracted mel-spectrogram for each segment is annotated with the genre, artist, gender, song, source, no_of_artists, genre_id, artist_id, gender_id.<br> _________________________________________________________________________________________________________<br> This project was funded under the grant number: ECR/2018/000204 by the Science & Engineering Research Board (SERB).</p>
Music Style Transfer datasets
<p>A collection of MIDI datasets used in my MSc project: Evaluating Music Style Tranfer methods (link TBD).</p>
JAZZVAR: A Dataset of Variations found within Solo Piano Performances of Jazz Standards for Music Overpainting
<p>Release of the MIDI data pairs that constitute the JAZZVAR dataset. See below for the abstract of the publication.</p> <p>The data is also available transposed to C/Am and subsequently, to all keys, with accompanying metadata.</p> <p>Abstract:</p> <p>Jazz pianists often uniquely interpret jazz standards. Passages from these interpretations can be viewed as sections of variation. We manually extracted such variations from solo jazz piano performances. The JAZZVAR dataset is a collection of 502 pairs of Variation and Original MIDI segments. Each Variation in the dataset is accompanied by a corresponding Original segment containing the melody and chords from the original jazz standard. Our approach differs from many existing jazz datasets in the music information retrieval (MIR) community, which often focus on improvisation sections within jazz performances. In this paper, we outline the curation process for obtaining and sorting the repertoire, the pipeline for creating the Original and Variation pairs, and our analysis of the dataset. We also introduce a new generative music task, Music Overpainting, and present a baseline Transformer model trained on the JAZZVAR dataset for this task. Other potential applications of our dataset include expressive performance analysis and performer identification.</p>
ScienceDex guides
Understand access before you commit
These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.
Allen Brain Atlas
Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.
Annotated Behaviour and Observability Dataset (ABODe)
ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.
DANDI Archive for NWB datasets
DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.
International Brain Laboratory public data
The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.
OpenNeuro
OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.