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.
106
datasets available to search
ShareScore release 0.9.0
Dataset results
106 results for “Audio Dataset”
University of Rochester Audio-Visual Solo Singing Performance (URSing) Dataset
<p>We introduce a dataset for facilitating audio-visual analysis of singing performances. The dataset comprises a number of songs where singers’ solo voices are recorded in isolation. For each song, we provide the high-quality audio recordings of the solo singing voice and mix with accompaniments, and the video recording of the upper body of the vocal soloist which contains facial expressions and lip movements. We anticipate that the dataset will be useful for developing audiovisual source separation systems. Note that some of the accompaniment tracks come with the backing vocals, which introduces extra challenges of developing an audio-based singing voice separation system, and encourages researchers to integrate the soloists’ visual information to facilitate the separation process. We also anticipate that the dataset will be useful for other multi-modal information retrieval techniques such as audiovisual expressions analysis, audio-visual correspondence, audiovisual lyrics transcription, etc.</p>
Acoustic models of Brazilian Portuguese Speech based on Neural Transformers - Pretraining Datasets raw audios from CORAA
<p>This repository contains all the pretraining datasets used in the paper: Acoustic models of Brazilian Portuguese Speech based on Neural Transformers by Marcelo Gauy and Marcelo Finger. These datasets are part of a collection of datasets from the TaRSila project (see https://sites.google.com/view/tarsila-c4ai). The audios published here were in part also published with annotations and transcriptions as the CORAA dataset (see https://github.com/nilc-nlp/CORAA). Here we publish the original raw audios from the following datasets (without transcriptions) - ALIP, C-Oral, SP2010, NURC-Recife, NURC-São Paulo and Programa Certas Palavras. In total, the datasets contain about 800 hours of Brazilian Portuguese Speech.</p> <p>The audios have been converted to mp3 to facilitate the upload. ALIP, C-Oral and SP2010 are integrally contained in one file each. Programa Certas Palavras and NURC-Recife are split in 3 parts each, while NURC-SP is split in 7 parts of roughly equal size. More information on the datasets can be found in the paper Acoustic models of Brazilian Portuguese Speech based on Neural Transformers as well as on the original references which created these datasets.</p>
Co-located Multi-device Audio Experiences Dataset
<p>This dataset contains the survey responses obtained from the survey on co-located multi-device audio experiences, in the form of a .csv file.</p>
Gunshot/Gunfire Audio Dataset
<p>This dataset presents a multi-firearm, multi-orientation gunshot/gunfire audio dataset collected using several edge devices positioned in and around an outdoor firearm range. This dataset will aid researchers who are interested in the task of gunshot detection, firearm classification, and direction of arrival, thereby helping advance the capabilities of Internet-of-Battlefield-Things (IoBT) systems. </p>
Video-Audio Neural Network Ensemble For Comprehensive Screening Of Autism Spectrum Disorder in Young Children (Openpose ADOS Dataset)
<p>Here, we share a de-identify subsample of the data used in the <a href="https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0308388">research article</a>, that will allow interested scientists to test the <a href="https://github.com/AutismBrainBehavior/Video-Neural-Network-ASD-screening">shared code</a>, as well as, develop alternatives for achieving better prediction accuracy. We have prepared a subsample of pose estimation videos for the first 10 minutes of ADOS examination videos for each of the 160 children including in the current study (80 ASD and 80 TD, 80 Training set and 80 Testing set).</p> <p>With this subset of the full dataset, our trained model achieved an accuracy of 68.75% over 80 videos (40 ASD & 40 TD) by training the Visual Geometry Group 16 Long short term memory recurrent neural network (VGG16 LSTM RNN) over 80 training videos (40 ASD & 40 TD) at 64 batch size and 120 epochs.</p>
BirdVox-DCASE-20k: a dataset for bird audio detection in 10-second clips
<p>BirdVox-DCASE-20k: a dataset for bird audio detection in 10-second clips<br> =====================================================<br> Version 2.0, March 2018.</p> <p><br> Created By<br> -------------</p> <p>Vincent Lostanlen (1, 2, 3), Justin Salamon (2, 3), Andrew Farnsworth (1), Steve Kelling (1), and Juan Pablo Bello (2, 3).</p> <p>(1): Cornell Lab of Ornithology (CLO)<br> (2): Center for Urban Science and Progress, New York University<br> (3): Music and Audio Research Lab, New York University</p> <p>https://wp.nyu.edu/birdvox</p> <p> </p> <p>Description<br> --------------</p> <p>The BirdVox-DCASE-20k dataset contains 20,000 ten-second audio recordings. These recordings come from ROBIN autonomous recording units, placed near Ithaca, NY, USA during the fall 2015. They were captured on the night of September 23rd, 2015, by six different sensors, originally numbered 1, 2, 3, 5, 7, and 10.</p> <p>Out of these 20,000 recording, 10,017 (50.09%) contain at least one bird vocalization (either song, call, or chatter).</p> <p>The dataset is a derivative work of the BirdVox-full-night dataset [1], containing almost as much data but formatted into ten-second excerpts rather than ten-hour full night recordings.</p> <p>In addition, the BirdVox-DCASE-20k dataset is provided as a development set in the context of the "Bird Audio Detection" challenge, organized by DCASE (Detection and Classification of Acoustic Scenes and Events) and the IEEE Signal Processing Society.</p> <p>The dataset can be used, among other things, for the development and evaluation of bioacoustic classification models.</p> <p><br> We refer the reader to [1] for details on the distribution of the data and [2] for details on the hardware of ROBIN recording units.</p> <p>[1] V. Lostanlen, J. Salamon, A. Farnsworth, S. Kelling, J.P. Bello. "BirdVox-full-night: a dataset and benchmark for avian flight call detection", Proc. IEEE ICASSP, 2018.</p> <p>[2] J. Salamon, J. P. Bello, A. Farnsworth, M. Robbins, S. Keen, H. Klinck, and S. Kelling. Towards the Automatic Classification of Avian Flight Calls for Bioacoustic Monitoring. PLoS One, 2016.</p> <p> </p> <p>Data Files<br> ------------</p> <p>The wav folder contains the recordings as WAV files, sampled at 44,1 kHz, with a single channel (mono). The original sample rate was 24 kHz.</p> <p>The name of each wav file is a random 128-bit UUID (Universal Unique IDentifier) string, which is randomized with respect to the origin of the recording in BirdVox-full-night, both in terms of time (UTC hour at the start of the excerpt) and space (location of the sensor).</p> <p>The origin of each 10-second excerpt is known by the challenge organizers, but not disclosed to the participants.</p> <p> </p> <p>Metadata Files<br> --------------</p> <p>A table containing a binary label "hasbird" associated to every recording in BirdVox-DCASE-20k is available on the website of the DCASE "Bird Audio Detection" challenge: http://machine-listening.eecs.qmul.ac.uk/bird-audio-detection-challenge/</p> <p>These labels were automatically derived from the annotations of avian flight call events in the BirdVox-full-night dataset.</p> <p>If your evaluation procedure requires the precise timestamps of each avian flight call (at a fine time scale of 50 ms), and is agnostic to non-flight call avian vocalizations (e.g. geese, crows, owls, etc.), we kindly suggest you to use the BirdVox-full-night dataset rather than BirdVox-DCASE-20k: wp.nyu.edu/birdvox/birdvox-full-night</p> <p>On the other hand, if your evaluation procedure encompasses all avian vocalizations, and is performed at a coarse time scale of 10 seconds, then BirdVox-DCASE-20k is the appropriate dataset.</p> <p>The annotation campaign of avian flight calls in BirdVox-full-night was performed by Andrew Farnsworth and lasted 102 hours.</p> <p>The additional annotation campaign of non-flight call avian vocalizations was performed by Vincent Lostanlen and lasted 10 hours.</p> <p>The accuracy of the labeling is estimated to be somewhere between 99.5% (100 mislabelings) and 99.95% (10 mislabelings).</p> <p><br> Please Acknowledge BirdVox-DCASE-20k in Academic Research<br> --------------------------------------------------------------------------------</p> <p>When BirdVox-70k is used for academic research, we would highly appreciate it if scientific publications of works partly based on this dataset cite the following publication:</p> <p>V. Lostanlen, J. Salamon, A. Farnsworth, S. Kelling, J. Bello. "BirdVox-full-night: a dataset and benchmark for avian flight call detection", Proc. IEEE ICASSP, 2018.</p> <p>@inproceedings{lostanlen2018icassp,<br> title = {BirdVox-full-night: a dataset and benchmark for avian flight call detection},<br> author = {Lostanlen, Vincent and Salamon, Justin and Farnsworth, Andrew and Kelling, Steve and Bello, Juan Pablo},<br> booktitle = {Proc. IEEE ICASSP},<br> year = {2018},<br> published = {IEEE},<br> venue = {Calgary, Canada},<br> month = {April},<br> }</p> <p>The creation of this dataset was supported by NSF grants 1125098 (BIRDCAST) and 1633259 (BIRDVOX), a Google Faculty Award, the Leon Levy Foundation, and two anonymous donors.</p> <p> </p> <p>Conditions of Use<br> ---------------------</p> <p>Dataset created by Vincent Lostanlen, Justin Salamon, Andrew Farnsworth, Steve Kelling, and Juan Pablo Bello.</p> <p>The BirdVox-DCASE-20k dataset is offered free of charge under the terms of the Creative Commons Attribution 4.0 International (CC BY 4.0) license:<br> https://creativecommons.org/licenses/by/4.0/</p> <p>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, Cornell Lab of Ornithology is not liable for, and expressly excludes all liability for, loss or damage however and whenever caused to anyone by any use of the BirdVox-DCASE-20k dataset or any part of it.</p> <p> </p> <p>Feedback<br> -----------</p> <p>Please help us improve BirdVox-DCASE-20k by sending your feedback to:<br> * Vincent Lostanlen: vincent.lostanlen@gmail.com for feedback regarding data pre-processing,<br> * Andrew Farnsworth: af27@cornell.edu for feedback regarding data collection and ornithology, or<br> * Dan Stowell: dan.stowell@qmul.ac.uk for feedback regarding the DCASE "Bird Audio Detection" challenge.</p> <p>In case of a problem, please include as many details as possible.</p> <p> </p> <p><br> Acknowledgements<br> ------------------------</p> <p>We thank Jessie Barry, Ian Davies, Tom Fredericks, Jeff Gerbracht, Sara Keen, Holger Klinck, Anne Klingensmith, Ray Mack, Peter Marchetto, Ed Moore, Matt Robbins, Ken Rosenberg, and Chris Tessaglia-Hymes for designing autonomous recording units and collecting data.<br> We acknowledge that the land on which the data was collected is the unceded territory of the Cayuga nation, which is part of the Haudenosaunee (Iroquois) confederacy.</p> <p> </p>
Pyramic Dataset : 48-Channel Anechoic Audio Recordings of 3D Sources
<p>The Pyramic Dataset contains recordings done using the<br> <a href="https://github.com/LCAV/Pyramic">Pyramic</a> 48 channel microphone array in an<br> anechoic chamber. The recordings consist of 8 different samples (2x sweeps, 1x<br> noise, 5x speech) repeated at 180 angles (every 2 degrees) and from 3 different<br> heights. The audio samples recorded are</p> <ul> <li>Linear and exponential sweeps</li> <li>Noise sequence</li> <li>2x male and 3x female speech</li> </ul> <p>This dataset allows to evaluate the performance of array processing algorithms<br> on real-life recordings done using MEMS microphones similar to those used in<br> mobile phones with all the non-idealities involved. The dataset is suitable for both 2D<br> and 3D scenarios. By subsampling the 48<br> microphones, a large number of array configurations can be tested. Example of<br> algorithms are:</p> <ul> <li>Direction of arrival (DOA) estimation</li> <li>Beamforming</li> <li>Source separation</li> <li>Array calibration</li> </ul> <p>Another application is the generation of realistic room impulse by combining<br> the impulse responses of microphones from sources at multiple angles with a<br> variant of the image source model.</p> <p>In addition to the raw (compressed or not) and segmented<br> recordings, the impulse responses of all the microphones for every source<br> locations were recovered from the exponential sweep measurements and are<br> distributed together with the dataset. The initial manual measurement of loudspeakers<br> and microphones locations was improved upon using a blind calibration method.</p> <p>This record contains</p> <ul> <li>The compressed recordings (TTA format)</li> <li>Segmented recorded samples</li> <li>Impulse responses</li> <li>Documentation and code (also available on <a href="https://github.com/fakufaku/pyramic-dataset">github</a>)</li> </ul> <p>The raw measurements in wav format are available as a separate <a href="https://zenodo.org/deposit/1209005">record</a> (10.5281/zenodo.1209005).</p> <p>The best way to get started is to only get the documentation and code from <a href="https://github.com/fakufaku/pyramic-dataset">github</a> (a copy is available in pyramic-dataset-doc-d2a456b4.zip) and follow the instructions in the README. The version on github is most up-to-date. If possible, please use that one.</p>
Pyramic Dataset : 48-Channel Anechoic Audio Recordings of 3D Sources (Raw)
<p>The Pyramic Dataset contains recordings done using the<br> <a href="https://github.com/LCAV/Pyramic">Pyramic</a> 48 channel microphone array in an<br> anechoic chamber. The recordings consist of 8 different samples (2x sweeps, 1x<br> noise, 5x speech) repeated at 180 angles (every 2 degrees) and from 3 different<br> heights. The audio samples recorded are</p> <ul> <li>Linear and exponential sweeps</li> <li>Noise sequence</li> <li>2x male and 3x female speech</li> </ul> <p>This dataset allows to evaluate the performance of array processing algorithms<br> on real-life recordings done using MEMS microphones similar to those used in<br> mobile phones with all the non-idealities involved. The dataset is suitable for both 2D<br> and 3D scenarios. By subsampling the 48<br> microphones, a large number of array configurations can be tested. Example of<br> algorithms are:</p> <ul> <li>Direction of arrival (DOA) estimation</li> <li>Beamforming</li> <li>Source separation</li> <li>Array calibration</li> </ul> <p>Another application is the generation of realistic room impulse by combining<br> the impulse responses of microphones from sources at multiple angles with a<br> variant of the image source model.</p> <p>In addition to the raw (compressed or not) and segmented<br> recordings, the impulse responses of all the microphones for every source<br> locations were recovered from the exponential sweep measurements and are<br> distributed together with the dataset. The initial manual measurement of loudspeakers<br> and microphones locations was improved upon using a blind calibration method.</p> <ul> </ul> <p>This record contains only the raw measurements in wav format and archive of the documentation and code.</p> <p>The post-processed data is available in a separate <a href="https://zenodo.org/record/1209563">record</a> that contains:</p> <ul> <li>The compressed recordings (TTA format)</li> <li>Segmented recorded samples</li> <li>Impulse responses</li> <li>Documentation and code (also available on <a href="https://github.com/fakufaku/pyramic-dataset">github</a>)</li> </ul> <p>The best way to get started is to only get the documentation and code from <a href="https://github.com/fakufaku/pyramic-dataset">github</a> and download the data as needed into the unzipped archive. Then follow the instructions in README.md.</p>
Turkish Makam Music Audio-Score Alignment Dataset
<p>This release contains the annotations and the scores to test the audio-score alignment methodology explained in:</p> <blockquote> <p><em>Şentürk, S., Gulati, S., and Serra, X. (2014). <strong>Towards alignment of score and audio recordings of Ottoman-Turkish makam music.</strong> In Proceedings of 4th International Workshop on Folk Music Analysis, pages 57–60, Istanbul, Turkey.</em></p> </blockquote> <p>The dataset in this release is derived from the transcription test dataset used in the paper:</p> <blockquote> <p><em>Benetos, E. & Holzapfel, A. (2013). <strong>Automatic transcription of Turkish makam music.</strong> In Proceedings of 14th International Society for Music Information Retrieval Conference, 4 - 8 Nov 2013, Curitiba, PR, Brazil.</em></p> </blockquote> <p>The scores for each composition are obtained from the SymbTr collection explained in:</p> <blockquote> <p><em>Karaosmanoğlu, K. (2012). <strong>A Turkish makam music symbolic database for music information retrieval: SymbTr.</strong> In Proceedings of 13th International Society for Music Information Retrieval Conference (ISMIR), pages 223–228.</em></p> </blockquote> <p>From the annotated score onsets for some of the above recordings only the main singing voice segments have been selected. Further separately only a subset of vocal onsets crresponding to phoneme transitions rules have been explicitly annotated as annotationOnsets.txt</p> <blockquote> <p><a href="http://mtg.upf.edu/biblio/author/810">Dzhambazov, G.</a>, <a href="http://mtg.upf.edu/biblio/author/644">Srinivasamurthy A.</a>, <a href="http://mtg.upf.edu/biblio/author/494">Şentürk S.</a>, & <a href="http://mtg.upf.edu/biblio/author/1012">Serra X.</a> (2016). <a href="http://mtg.upf.edu/node/3492">On the Use of Note Onsets for Improved Lyrics-to-audio Alignment in Turkish Makam Music</a>. 17th International Society for Music Information Retrieval Conference (ISMIR 2016</p> </blockquote> <p><strong>Using this dataset</strong></p> <p>Please cite the above publications if you use this dataset in a publication.</p> <p>We are interested in knowing if you find our datasets useful! If you use our dataset please email us at <a href="mailto:mtg-info@upf.edu">mtg-info@upf.edu</a> and tell us about your research.</p> <p> </p> <p><a href="http://compmusic.upf.edu/node/233">http://compmusic.upf.edu/node/233 </a></p>
JAAH: Audio-aligned jazz harmony dataset
<p>The dataset consists of annotations of 113 tracks selected from “The Smithsonian Collection of Classic Jazz” and “Jazz: The Smithsonian Anthology,” covering a range of performers, subgenres, and historical periods. Annotations were made by a jazz musician and contain information about the meter, structure, and chords for entire audio tracks.</p> <p>For referencing the JAAH, and obtaining a more detailed description of it, please refer to:</p> <blockquote>Eremenko, V., Demirel, E., Bozkurt, B., & Serra, X. (2018). <a href="https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=3&ved=2ahUKEwiU2cGnvoLgAhUkiKYKHT96AW0QFjACegQICBAC&url=https%3A%2F%2Fzenodo.org%2Frecord%2F1291834%2Ffiles%2FJazzHarmonyDataset_Ismir2018.pdf&usg=AOvVaw38mi5jVoPpKLWiIJYt5CGl">Audio-aligned jazz harmony dataset for automatic chord transcription and corpus-based research.</a> International Society for Music Information Retrieval Conference.</blockquote>
Evaluation datasets for DCASE 2018 Bird Audio Detection
<p>Evaluation data audio for <a href="http://dcase.community/challenge2018/task-bird-audio-detection">the DCASE 2018 Bird Audio Detection task (Task 3)</a>.</p> <ol> <li> <p><strong>Crowdsourced dataset, UK ("warblrb10k")</strong> - a held-out set of 2,000 recordings from the same conditions as the Warblr development dataset.</p> </li> <li> <p><strong>Remote monitoring dataset, Chernobyl ("Chernobyl")</strong> - 6,620 audio clips collected from unattended remote monitoring equipment in the Chernobyl Exclusion Zone (CEZ). This data was collected as part of the <a href="https://wiki.ceh.ac.uk/display/NRT/NERC+RATE+TREE+Home">TREE</a> (Transfer-Exposure-Effects) research project into the long-term effects of the Chernobyl accident on local ecology. The audio covers a range of birds and includes weather, large mammal and insect noise sampled across various CEZ environments, including abandoned village, grassland and forest areas.</p> </li> <li> <p><strong>Remote monitoring night-flight calls, Poland ("PolandNFC")</strong> - 4,000 recordings from Hanna Pamuɫa's PhD project of monitoring autumn nocturnal bird migration. The recordings were collected every night, from September to November 2016 on the Baltic Sea coast, Poland, using Song Meter SM2 units with microphones mounted on 3–5 m poles. For this challenge, we use a subset derived from 15 nights with different weather conditions and background noise including wind, rain, sea noise, insect calls, human voice and deer calls.</p> </li> </ol>
Audio file of Malay diphthong (dataset)
<p>This dataset provides the spoken audio for the diphthong /au/, /ai/, /oi/ and /ui/ from district regions of Selangor (standard), Perlis, Kedah, and Pulau Pinang located in Malaysia. The overall recording of the audio spontaneous speech is approximately 4 hours. A total of 193 diphthongs were extracted from the spoken words and the file was grouped into district and diphthong as shown below. This is the clean and processed data with clear pronunciation of the diphthongs. The dataset have been used to analyze the pronunciation of the diphthong between the districts regions in Malaysia and can be used further for camparing between other languages.</p> <p> </p> <p>DiphData.zip contains all of these file:</p> <p> </p> <p>Selangor’s diphthong file</p> <p>S(au).file</p> <p>S(ai).file</p> <p>S(oi).file</p> <p> </p> <p>Perlis’s diphthong file</p> <p>P(au).file</p> <p>P(ai). file</p> <p>P(oi).file</p> <p>P(ui).file</p> <p> </p> <p>Kedah’s diphthong file</p> <p>K(au).file</p> <p>K(ai).file</p> <p>K(oi).file</p> <p>K(ui).file</p> <p> </p> <p>Pulau Pinang’s diphthong file</p> <p>PP(au).file</p> <p>PP(ai).file</p> <p>PP(oi).file</p> <p>PP(ui).file</p> <p> </p>
Hit Song Prediction (Million Song Dataset and Audio Features)
<p><strong>Hit Song Prediction Dataset</strong></p> <p>This dataset is based on the Million Song Dataset (MSD), which contains one million songs that are representative for western commercial music released between 1922 and 2011. The dataset contains release year information for 515,576 of the MSD songs. Please refer to http://millionsongdataset.com/ for further information on the million song dataset.</p> <p>For our hit song prediction experiments, we extract high- and low-level audio features using the Essentia toolkit (cf. https://essentia.upf.edu/). For the high-level features, we make use of the pre-trained classifiers as provided by Essentia. For a detailed description of the features, please visit the Essentia documentation.</p> <p><br> The dataset hence contains:</p> <ul> <li><strong>Audio features</strong>: the compressed msd_audio_features.tar.gz file contains the low- and high-level features for each track, stored as json files. Please note that we organize all MSD audio feature files based on the track's identifier with one folder holding all tracks with the same first letter of the track identifier to keep the files manageable. For each track, we provide two files: one containing the high-level and one containing the low-level features extracted by Essentia.</li> <li><strong>Billboard data:</strong> the folder billboard_data contains two files: msd_bb_matches.csv contains information about the MSD tracks that were also featured in the Billboard Hot 100 charts. Here, we provide the MSD id, Echo Nest id, artist name, track title, release year, peak position in Billboard charts and the number of weeks in the charts. The second file, msd_bb_non_matches.csv contains meta-information about the tracks of the MSD that were not featured in the Billboard Hot 100 and hence were used as negative samples. Here, we provide the MSD id, Echo Nest id, artist name, track title and the release year.</li> </ul> <p><br> If you make use of the dataset, please kindly cite the following paper:</p> <p>Eva Zangerle, Michael Vötter, Ramona Huber, and Yi-Hsuan Yang. Hit Song Prediction: Leveraging Low- and High-Level Audio Features. In Proceedings of the 20th International Society for Music Information Retrieval Conference 2019 (ISMIR 2019), 2019.</p> <p><br> @inproceedings{zangerle_ismir19,<br> title = {{Hit Song Prediction: Leveraging Low- and High-Level Audio Features}},<br> author = {Eva Zangerle and Ramona Huber and Michael V\"{o}tter and Yi-Hsuan Yang},<br> year = {2019},<br> booktitle = {{Proceedings of the 20th International Society for Music Information Retrieval Conference 2019 (ISMIR 2019)}},<br> }</p>
BirdVox-scaper-10k: a synthetic dataset for multilabel species classification of flight calls from 10-second audio recordings
<p>BirdVox-scaper-10k: a synthetic dataset for multilabel species classification of flight calls from 10-second audio recordings<br> =============================================================================================<br> Version 1.0, September 2019.</p> <p> </p> <p>Created By<br> -------------</p> <p>Elizabeth Mendoza (1), Vincent Lostanlen (2, 3, 4), Justin Salamon (3, 4), Andrew Farnsworth (2), Steve Kelling (2), and Juan Pablo Bello (3, 4).</p> <p> </p> <p>(1): Forest Hills High School, New York, NY, USA<br> (2): Cornell Lab of Ornithology, Cornell University, Ithaca, NY, USA<br> (3): Center for Urban Science and Progress, New York University, New York, NY, USA<br> (4): Music and Audio Research Lab, New York University, New York, NY, USA</p> <p>https://wp.nyu.edu/birdvox</p> <p> </p> <p>Description<br> --------------</p> <p>The BirdVox-scaper-10k dataset contains 9983 artificial soundscapes. Each soundscape lasts exactly ten seconds and contains one or several avian flight calls from up to 30 different species of New World warblers (Parulidae). Alongside each audio file, we include an annotation file describing the start time and end time of each flight call in the corresponding soundscape, as well as the species of warbler it belongs to.</p> <p>In order to synthesize soundscapes in BirdVox-scaper-10k, we mixed natural sounds from various pre-recorded sources. First, we extracted isolated recordings of flight calls containing little or no background noise from the CLO-43SD dataset [1]. Secondly, we extracted 10-second "empty" acoustic scenes from the BirdVox-DCASE-20k dataset [2]. These acoustic scenes contain various sources of real-world background noise, including biophony (insects) and anthropophony (vehicles), yet are guaranteed to be devoid of any flight calls. Lastly, we "fill" each acoustic scene by mixing it with flight calls sampled at random.</p> <p>Although the BirdVox-scaper-10k does not consist of natural recordings, we have taken several measures to ensure the plausibility of each synthesized soundscape, both from qualitative and quantitative standpoints.<br> <br> The BirdVox-scaper-10k dataset can be used, among other things, for the research, development, and testing of bioacoustic classification models.</p> <p>For details on the hardware of ROBIN recording units, we refer the reader to [2].</p> <p>[1] J. Salamon, J. Bello. Fusing shallow and deep learning for bioacoustic bird species classification. Proc. IEEE ICASSP, 2017.</p> <p>[2] V. Lostanlen, J. Salamon, A. Farnsworth, S. Kelling, and J. Bello. BirdVox-full-night: a dataset and benchmark for avian flight call detection. Proc. IEEE ICASSP, 2018.</p> <p>[3] J. Salamon, J. P. Bello, A. Farnsworth, M. Robbins, S. Keen, H. Klinck, and S. Kelling. Towards the Automatic Classification of Avian Flight Calls for Bioacoustic Monitoring. PLoS One, 2016.</p> <p> </p> <p> </p> <p> </p> <p>@inproceedings{lostanlen2018icassp,<br> title = {BirdVox-full-night: a dataset and benchmark for avian flight call detection},<br> author = {Lostanlen, Vincent and Salamon, Justin and Farnsworth, Andrew and Kelling, Steve and Bello, Juan Pablo},<br> booktitle = {Proc. IEEE ICASSP},<br> year = {2018},<br> published = {IEEE},<br> venue = {Calgary, Canada},<br> month = {April},<br> }</p>
Audio Source Separation Dataset
<p>AUDIO SOURCE SEPARATION DATASET.</p> <p>This dataset has been constructed from audio excerpts taken from the Bach10 dataset by Duan et al. [1]. This database can be used in performance evaluation and results can be compared with the ones presented in my PhD thesis [2], Section 5.8.5, on pages 153-158, in Chapter 5. A percussive sequence from the Open Air Library [3], has also been used in these experiments.</p> <p>[1] Z. Duan, B. Pardo, and C. Zhang, "Multiple fundamental frequency estimation by modeling spectral peaks and non-peak regions," IEEE Transactions on Audio, Speech and Language Processing, vol. 18, no. 8, pp. 2121-2133, 2010.</p> <p>[2] Delgado Castro, A. "Iterative separation of note events from single-channel polyphonic recordings," Ph.D. University of York. 2019.</p> <p>[3] https://www.york.ac.uk/electronic-engineering/research/communication-technologies/projects/open-acoustic-impulse-response-library/</p>
Human animal hunting audio dataset
<p>YouTube collected audio samples from human hunting animals, such as Coyotes, Bobcats, Wild Boars.</p> <p>The samples are in MP3 format, having a maximum duration of 3 seconds.</p> <p>The audio was classified by a semi-supervised process using a sound event classification system. </p>
Beehive piping audio dataset
<p>This dataset of bee piping audio signals which was built by collecting about 48 different recordings that were published by various beekeepers on the YouTube platform.<br> Each recording has a duration varying from 2 to 13 seconds and is annotated according to the beekeeper comment respectively as <em>tooting</em> or <em>quacking</em> signals.<br> We extracted the audio from the video using the ``YouTube soundtrack extraction'' tool which enables to store each recording<br> into the WAVE file with a sampling frequency of <span class="math-tex">\(F_s = 44,1kHz\)</span> or <span class="math-tex">\(F_s = 48kHz\)</span>.<br> <br> After removing the silent frames, the resulting dataset contains 36 tooting signals and 8 quacking signals which correspond to a duration of 145 seconds for tooting and 60 seconds for quacking (total 205 seconds).</p> <p>The <em>DATASET_DESCRIPTION.pdf</em> file contains a precise description of each recording with the links to original YouTube sources.</p>
Audio Piano Triad Dataset
<p>Created by: Agustín Macaya Valladares<br> Date: May 5th, 2021</p> <p>- Dataset contains 43.200 examples of piano triads in .wav format.<br> - Second Version: The audios are the same as the first version, but the octave number in the names were corrected from (2, 3, 4) to (3, 4, 5), respectively.</p> <p>Details:<br> - Sample rate: 16000 Hz.<br> - Data type: 16-bit PCM (int16).<br> - File size: Each example has a file size of 128 kB (5.53 GB for complete dataset).<br> - Duration: 4 seconds.<br> - Sound: Piano (digital).<br> - Chords were played by a human on a velocity-sensitive piano keyboard.<br> - 3 seconds pressed, 1 second released.</p> <p>- 3 octaves (3, 4, 5).<br> - 12 base notes per octave: Cn, Df, Dn, Ef, En, Fn, Gf, Gn, Af, An, Bf, Bn. (n is natural, f is flat).<br> - 4 triad types per note: major (j), minor (n), diminished (d), augmented (a). No inversions.<br> - 3 volumes per triad: forte (f), metsoforte (m), piano (p).</p> <p><br> - 10 original examples per combination of octave, base note, triad type, and volume. (10*3*12*4*3 = 4.320 examples).<br> - x10 data augmentation for each example (4.320 * 10 = 43.200 total examples).<br> - Data augmentation through random temporal and amplitude shifts.<br> - Metadata is in the name of the chord. For example: "piano_4_Af_d_m_45.wav" is a piano chord, (4) 4th octave, (Af) A flat base note, (d) diminished, (m) metsoforte, 45th example.</p> <p>Note:<br> - The audios are in 16-bit PCM (int16) data type to reduce the file size. This means that the dynamic range of values in the array is -32768 to 32768, integers. To normalize the audios in the range -1 to 1 (float) just divide by 32768.</p> <p>Second Version: The audios are the same as the first version, but the octave number in the names were corrected from (2, 3, 4) to (3, 4, 5), respectively.</p>
LOLA. Flamenco tagged audios Dataset
<p>This dataset holds ~1500 audio samples of 3 different flamenco styles. Those styles are:</p> <ul> <li>Bulerías</li> <li>Alegrías</li> <li>Sevillanas</li> </ul> <p>Each audio is in a folder with the same name as the style to which it belongs. Each audio has a duration between 10 and 15 seconds and is saved in mp3 format.</p>
jazznet: A Dataset of Fundamental Piano Patterns for Music Audio Machine Learning Research
<p>Jazznet is a dataset of piano patterns for music audio machine learning research. The dataset comprises chords, arpeggios, scales, and chord progressions in all keys of an 88-key piano and in all the inversions, for a total of 162520 labeled piano patterns, resulting in 95GB of data and more than 26k hours of audio. The data is also accompanied by Python scripts to enable the easy generation of new piano patterns beyond those present in the dataset. The data is broken down into small, medium, and large subsets, comprising 21516, 30328, and 52360 patterns, respectively (with all the chords, arpeggios, and scales being present in all subsets). </p> <p>The GitHub page of the dataset, containing details of the dataset and scripts for generating new data is https://github.com/tosiron/jazznet.</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.