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.
33
datasets available to search
ShareScore release 0.7.1
Dataset results
33 results for “source separation”
Dataset for Close Miking Empirical Practice Verification: A Source Separation Approach
<p>A dataset consisting of audio files which serve as support material to the following work: "Close Miking Empirical Practice Verification: A Source Separation Approach" by K. Drossos, S.I. Mimilakis, A. Floros, T. Virtanen and G. Schuller.</p> <p>The audio files contain information about two kind of signals a noise (pink noise) and a guitar (musical source) signal. For each signal multiple recordings, using a sampling frequency of 44.1kHz@16-bit, exist which are dependent upon the following variables:<br> - Mircophone type (polar pattern: Omni-directional and cardioid lobe) @Corresponding Folders: Omni & Card<br> - Microphone angle (30 and 45 degrees only for the cardioid microphone) @Corresponding Folders: C30 & C45<br> - Sound pressure levels for the source (SPLs) and the noise(SPLn) (SPLs : 94dB, 97dB, 100dB || SPLn: 88dB, 91dB, 94dB, 97dB, 100dB)<br> - 12 Distances in meters: 0.03m - 0.30m with a step size of 0.03m & 0.30m - 1.00m with a step size of 0.35m</p> <p>Each recording set contains also the original "clean" sources and their mixture for all the above configurations.<br> <br> The equipment used to record the above signals:<br> -Sound level meter (SLM): B&K 2250 Type A SLM <br> -Mic. A: & Shure SM57, dynamic, cardioid<br> -Mic. B: & Behringer ECM8000, condenser, omni-directional<br> -Laptop: Macbook Pro 15''<br> -Recording software: Digidesign ProTools M-Powered 8<br> -Musical instrument amplifier: Behringer V-Tone GMX212<br> -Digital sound card: M-Audio Fast Track Ultra<br> -Loudspeaker: Electrovoice SX300<br> <br> All recordings took place on the main stage of an empty municipal theater in Lixoyri, Kefallonia, Greece.<br> The authors would like to thank the Department of Technology of Sound and Musical Instruments,<br> Technological Educational Institute of Ionian Islands, for providing the equipment for the measurements.</p>
Stimuli for the paper Perceptual Evaluation of Source Separation for Remixing Music
<p>Stimuli for the paper</p> <p>H. Wierstorf, D. Ward, R. Mason, E. M. Grais, C. Hummersone, M. D. Plumbley, "Perceptual Evaluation of Source Separation for Remixing Music," in 143rd Convention of the Audio Engineering Society, 2017.</p> <p>The files used for the experimental procedure are available at https://doi.org/10.5281/zenodo.835191</p> <p>The stimuli in this publication are based on the DSD100 and submission files of the SiSEC challenge, see https://www.sisec17.audiolabs-erlangen.de</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>
Blind Source Separation of Musical Notes
<p>BLIND SOURCE SEPARATION OF MUSICAL NOTES.</p> <p>This dataset contains 12 groups of three musical notes whose pitches are not in harmonic relation. They were extracted from the RWC Musical Instrument Sound Database [1] and were used for performance evaluation of an audio source separation system, proposed in my PhD thesis [2], in Chapter 5, Section 5.8.2, on page 150.</p> <p>[1] M. Goto, H. Hashiguchi, T. Nishimura, and R. Oka, "RWC music database: music genre database and musical instrument sound database," in Proceedings of the 4th International Conference on Music Information Retrieval, pp. 229-230, 2003.</p> <p>[2] Delgado Castro, A. "Iterative separation of note events from single-channel polyphonic recordings". Ph.D. University of York. 2019.</p>
SynthSOD: Developing an Heterogeneous Dataset for Orchestra Music Source Separation
<p>The SynthSOD dataset contains more than 47 hours of multitrack music obtained by synthesizing orchestra and ensemble pieces from the <a href="https://qsdfo.github.io/LOP/database.html" target="_blank" rel="noopener">Symbolic Orchestral Database (SOD)</a> using Spitfire BBC Symphony Orchestra Professional Library. To synthesize the MIDI files from the SOD, we needed to fix the original files into the General MIDI standard, select a subsect of files that fitted into our requirements (e.g., containing only instruments that we could synthesize), and develop a new system to generate musically-motivated random annotations about tempo, dynamic, and articulation. The code to replicate this process is available in <a href="https://github.com/repertorium/HQ-SOD-generator" target="_blank" rel="noopener">our repository</a> and all the details can be read in <a href="https://doi.org/10.1109/OJSP.2025.3528361" target="_blank" rel="noopener">our paper</a>. We have also published the code to train and evaluate the baseline and the pre-trained models in a <a href="https://github.com/repertorium/SynthSOD-Baseline" target="_blank" rel="noopener">GitHub repository</a>.</p> <p>We have also published the aligned score information for most of the pieces <a href="https://doi.org/10.5281/zenodo.14971533">here</a>.</p>
HHDS - Spanish HipHop Dataset for Music Source Separation
<p><strong>What is HHDS?</strong></p> <p>HHDS is a reduced compilation of Hip Hop songs, used to train a Convolutional Neural Network (CNN) for audio source separation in [1], built on top of the DeepConvSep framework [2] developed at the Music Technology Group (MTG), Universitat Pompeu Fabra.</p> <p>The structure of HHDS follows the convention of DSD100 [3] (Demixing Secrets Dataset). HHDS contains the separated tracks for the categories of bass, drums, vocals and others in monophonic WAV les with a sampling rate of 44100Hz. The mixture is calculated by normalizing the sum of the tracks. The main difference with respect to DSD100 is that in HHDS there are HipHop songs only, instead of many different genres. The total number of songs is 18, from which 13 are used for training and 5 are used for evaluation.</p> <p>A detailed list of the songs included in the dataset can be found inside the .zip file provided. The reader can also find the code for this dataset in the DeepConvSep repository in the path examples/hiphopss.</p> <p> </p> <p><strong>References</strong></p> <p>[1] "A Deep Learning Approach to Source Separation and Remixing of HipHop music", Héctor Martel, Undergraduate Thesis, Universitat Pompeu Fabra 2016-2017. </p> <p>[2] DeepConvSep repository on GitHub: https://github.com/MTG/DeepConvSep</p> <p>[3] Demixing Secrets Dataset (DSD100), SiSEC2016: http://liutkus.net/DSD100.zip</p>
Experimental procedure for the paper Perceptual Evaluation of Source Separation for Remixing Music
<p>All the files you need to rerun the experiment described in the paper:</p> <p>H. Wierstorf, D. Ward, R. Mason, E. M. Grais, C. Hummersone, M. D. Plumbley, "Perceptual Evaluation of Source Separation for Remixing Music," in 143rd Convention of the Audio Engineering Society, 2017.</p> <p>The actual stimuli are not part of this publication, but can be regenerated or be downloaded from https://doi.org/10.5281/zenodo.835182<br> </p>
Figures and data for the paper Perceptual Evaluation of Source Separation for Remixing Music
<p>Listening test results and figures for the paper:</p> <p>H. Wierstorf, D. Ward, R. Mason, E. M. Grais, C. Hummersone, M. D. Plumbley,<br> "Perceptual Evaluation of Source Separation for Remixing Music," in 143rd<br> Convention of the Audio Engineering Society, 2017.</p> <p>`fig02/data/` contains the results from single listeners and median results across<br> listeners.<br> `fig02/fig02.plt` is the code to regenerate `fig02/fig02.pdf`.</p> <p>`fig03/data/` contains average medians across all songs.<br> `fig03/fig03.plt` is the code to regenerate `fig03/fig03.pdf`.</p> <p>The listening test results were extracted from the raw data stored together with<br> the experimental procedure in https://doi.org/10.5281/zenodo.835191.<br> There the file `experiment/saves/ratings/analyze_results.py` can be run in order<br> to regenerate the result files provided with this publication.<br> </p>
Scenario data, model source code and plotting routine for manuscript: Separating CO2 emission from removal targets comes with limited cost impacts
<p>This data archive contains REMIND model setup, results data and data analysis files for manuscript:<br><strong>Separating CO2 emission reduction from removal targets comes with limited cost impact.<br><br>plotting</strong>(directory) contains results data, manuscript specific data analysis and plotting routine scripts used to generate the figures of the manuscript.<br><strong>remind</strong>(directory) contains REMIND model source code and scenario set-up. Detailed scenario configurations are set in remind/config/scenario_config_SepMark.csv.<br><strong>remind2</strong>(directory) contains the slightly modified R-library package used for post-processing of REMIND output.<br><br>AMENDMENT<br><strong>Plots_SeparateMarkets_afterReviewProcess.Rmd</strong> After the review process, the new plotting script was added including the additional figures in the Supplementary Material. This file should replace the previous R-markdown file SepMark_essential/plotting/Plots_SeparateMarkets.Rmd.</p>
Relative Transfer Matrix for Low SNR Speech Separation from Noisy Sources in Reverberant Rooms
<p>This folder contains the supplementary audio files for the paper "Relative Transfer Matrix for Low SNR Speech Separation from Noisy Sources in Reverberant Rooms" submitted to <em>The Journal of the Acoustical Society of America</em>.</p>
DCASE 2024 Task 9: Language-Queried Audio Source Separation | Validation Set
<p>This is the <strong>validation set for Task 9, Language-Queried Audio Source Separation (LASS), in DCASE 2024 Challenge</strong>. </p> <p>This validation split is meant to be used for Task 9 at the scientific challenge DCASE 2024. This split is not meant to be used for training LASS methods. This split is meant to be used for evaluating LASS methods during the model development stage.</p> <p>This validation set consists of 1000 audio files sourced from Freesound [1], uploaded between April and October 2023. Each audio file has been manually annotated with three captions. In the annotation guidance, we instructed annotators to describe the content of audio clips using 5-20 words (similar to the caption style in Clotho [3] and AudioCaps [4] datasets). The tags of each audio file were verified and revised according to the FSD50K [2] sound event categories. Each audio file has been chunked into a 10-second clip and downsampled to 16kHz.</p> <p><strong>== Details ==</strong></p> <p>The audio files in the archives:</p> <ul> <li>lass_validation.zip</li> </ul> <p>and the associated metadata (including tags and captions) in the JSON file:</p> <ul> <li>lass_validation.json</li> </ul> <p>Participants will evaluate their LASS models using synthetic mixture data in the development stage. Specifically, given an audio clip A1 and its corresponding caption C, we select an additional audio clip, A2, to serve as background noise, thereby creating a mixed audio, A3. We anticipate that the LASS system, given A3 and C as inputs, will be able to separate the A1 source. We use the revised tags information to ensure that the two audio clips used in each mix do not share overlapping sound source classes. Three thousand synthetic audio mixtures with signal-to-noise ratios (SNR) ranging from -15dB to 15dB will be generated for the validation of LASS model development. These synthetic mixtures can be generated based on the provided CSV file:</p> <ul> <li>lass_synthetic_validation.csv</li> </ul> <p>The evaluation tool can be found at: https://github.com/Audio-AGI/dcase2024_task9_baseline/blob/main/dcase_evaluator.py</p> <p><strong>== References ==</strong></p> <p>[1] Fonseca E, Pons Puig J, Favory X, et al. Freesound datasets: a platform for the creation of open audio datasets. International Society for Music Information Retrieval (ISMIR), 2017.</p> <p>[2] Fonseca E, Favory X, Pons J, et al. FSD50k: an open dataset of human-labeled sound events. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2021, 30: 829-852.</p> <p>[3] Drossos K, Lipping S, Virtanen T. Clotho: An audio captioning dataset. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). 2020: 736-740.</p> <p>[4] Kim C D, Kim B, Lee H, et al. AudioCaps: Generating captions for audios in the wild. Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics (NAACL). 2019: 119-132.</p>
DCASE 2024 Task 9: Language-Queried Audio Source Separation | Development Set
<p><strong>== Description == </strong></p> <p>The development set is composed of audio samples from FSD50K [1] and Clotho v2 [2] datasets. FSD50K contains over 51k audio clips (~100 hours) manually labeled using 200 classes drawn from the AudioSet Ontology. For each audio clip in the FSD50K dataset, we generated one automatic caption for each audio clip by prompting ChatGPT (GPT-4) with its sound event tags. All audio files should be converted to mono 16 kHz audio for training LASS models. </p> <p>Clotho v2: <a href="../records/4783391">https://zenodo.org/records/4783391</a></p> <p>FSD50K: <a href="../records/4060432">https://zenodo.org/records/4060432</a></p> <p>Automatic captions generated for FSD50K:</p> <ul> <li>fsd50k_dev_auto_caption.json</li> <li>fsd50k_eval_auto_caption.json</li> </ul> <p>Prompt for generating captions:</p> <blockquote> <p>I will give you a number of lists containing sound events. Please write an one-sentence audio caption to describe these sounds.</p> <p>Make sure you are using grammatical subject-verb-object sentences. Directly describe the sounds and avoid using the word “heard”. Please don't describe the temporal order of these sound events. The caption should be less than 20 words.</p> </blockquote> <p>In addition to the development set, participants are free to use any external data (including private data) but are not allowed to use audio in Freesound uploaded between April and October 2023. Participants must specify all external resources utilized in their submission in the technical report.</p> <p><strong>== References ==</strong></p> <p>[1] Fonseca E, Favory X, Pons J, et al. FSD50k: an open dataset of human-labeled sound events. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2021, 30: 829-852.</p> <p>[2] Drossos K, Lipping S, Virtanen T. Clotho: An audio captioning dataset. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). 2020: 736-740.</p> <p><strong>== Contact ==</strong></p> <p>Xubo Liu, xubo.liu@surrey.ac.uk</p>
Multiple Speaker Separation from Noisy Sources in Reverberant Rooms using Relative Transfer Matrix
<p>This folder is the supplementary audio files for the paper "Multiple Speaker Separation from Noisy Sources in<br>Reverberant Rooms using Relative Transfer Matrix" submitted to the European Signal Processing Conference (EUSIPCO).</p>
soundscape_IR: A source separation toolbox for exploring acoustic diversity in soundscapes
<p>1. Soundscapes contain rich acoustic information associated with animal behaviors, environmental characteristics, and human activities, providing opportunities for predicting biodiversity changes and associated drivers. However, assessing the diversity of animal vocalizations remains challenging due to the interference of environmental and anthropogenic noise. A tool for separating sound sources and delineating changes in acoustic signals is crucial for an effective assessment of acoustic diversity.</p> <p>2. We present soundscape_IR, an open-source Python toolbox dedicated to soundscape information retrieval in which non-negative matrix factorization is applied. This toolbox provides algorithms for supervised and unsupervised source separation (SS). It also enables the use of a snapshot recording for model training and subsequently applying adaptive and semi-supervised SS when target species produce sounds with varying features and when unseen sound sources are encountered.</p> <p>3. Our results demonstrated that SS could enhance the vocalizations of target species, characterize the complexity of vocal repertoires, and investigate the spatio-temporal divergence of soundscapes. In tropical forest soundscapes, the application of SS effectively detected the rutting vocalizations of sika deer and revealed a graded structure in their acoustic characteristics. In subtropical estuarine soundscapes, SS automated the process of identifying distinct biotic and abiotic sounds, and the result uncovered divergent sound compositions between inshore and offshore waters.</p> <p>4. Implementation of SS in soundscape analysis offers a promising method for streamlining the assessment of acoustic diversity in diverse environments. Future application of SS will open new directions to acoustically quantify ecological interactions across individual, species, and ecosystem levels.</p>
CrossNet-Open-Unmix for Music Source Separation (X-UMXL)
<p>Weights of CrossNet-Open-Unmix (X-UMX) trained on the internal 100h dataset which is larger than <a href="https://sigsep.github.io/datasets/musdb.html">MUSDB18</a>, named X-UMX Large (X-UMXL). The weights can be used with <a href="https://github.com/asteroid-team/asteroid/tree/master/egs/musdb18/X-UMX">X-UMX on Asteroid (PyTorch)</a>. The details of X-UMX are described in <a href="https://ieeexplore.ieee.org/document/9414044">here</a>.</p>
DCASE 2024 Task 9: Language-Queried Audio Source Separation | Evaluation Set
<p>This is the <strong>evaluation set for Task 9, Language-Queried Audio Source Separation (LASS), in DCASE 2024 Challenge</strong>. </p> <p>This evaluation set is meant to be used for Task 9 at the scientific challenge DCASE 2024. This split is not meant to be used for training LASS methods. This split is meant to be used for evaluating LASS methods in the final testing & ranking stage. All audio clips are sourced from Freesound, uploaded between April and October 2023. Each audio file has been segmented into 10-second clips and converted to mono 16 kHz.</p> <p>This evaluation set consists of<strong> evaluation set (synth)</strong> and an <strong>evaluation set (real)</strong>. </p> <p><strong>== Evaluation set (synth) ==</strong></p> <p>This evaluation set is created using 1,000 audio clips. Each clip is annotated with three captions describing the content of the clip. We created 3,000 synthetic mixtures with signal-to-noise ratios (SNR) ranging from -15 to 15 dB. Each synthetic mixture includes one natural language query and its corresponding target source. We used annotated tag information to ensure that the two audio clips used in each mix do not share overlapping sound source classes. The original audio files used to create these mixtures are not released. The mixtures and language queries are available for evaluation.</p> <p>The audio files in the archives:</p> <ul> <li>lass_evaluation_synth.zip</li> </ul> <p>and the associated metadata (including audio filename and text queries) in the CSV file:</p> <ul> <li>lass_synthetic_evaluation.csv</li> </ul> <p><strong>== Evaluation set (real) ==</strong></p> <p>This evaluation set consists of 100 audio clips. Each audio clip contains at least two overlapping sound sources. For each audio clip, we manually annotated their component sources using text descriptions, so that each clip can be used as a 'mixture' from which to extract one or more of the component sources based on a text query. Each audio clip in evaluation (real) was labeled with two such text queries.</p> <p>The audio files in the archives:</p> <ul> <li>lass_evaluation_real.zip</li> </ul> <p>and the associated metadata (including audio filename and text queries) in the CSV file:</p> <ul> <li>lass_real_evaluation.csv</li> </ul>
Data related to the utilisation of source-separated biowaste for the production of bio-based chemicals
<p><span>The main objective of this dataset was to demonstrate the technological solution of biosolvents production (mainly bioethanol) via the utilisation of urban biowaste within the city context of Athens, Greece. More specifically, the aim of the dataset is the demonstration of a pre-existing system, namely PILOT 5 for the production of bioethanol. </span></p> <p><span>The results achieved were very promising for the viability of the process, either with dried or wet feedstock. It is important to consider the energy consumption, as it constitutes more than 50% in the overall ethanol production cost. In comparison, the total energy consumption for the production of ethanol with dried feedstock is 26% higher than with wet feedstock.</span><span> </span><span>The most energy-intensive stage is drying, followed by distillation. </span></p> <p> </p>
Support Material for: S.I. Mimilakis et al., ``Examining the Perceptual Effect of Alternative Objective Functions for Deep Learning Based Music Source Separation''
<p>The audio corpus used for conducting the listening tests reported in S.I. Mimilakis et al., ``Examining the Perceptual Effect of Alternative Objective Functions for Deep Learning Based Music Source Separation''.</p>
soundscape_IR: A source separation toolbox for exploring acoustic diversity in soundscapes
Open the record for dataset details and reuse information.
Data from: Separating sources of density-dependent and density-independent establishment limitation in invading species
Successful colonization by invasive species depends on both the ability to disperse seeds to a site and an ability to establish once seeds have arrived. While seed and establishment limitation are known to jointly influence colonization, decomposing establishment limitation into density-dependent and density-independent components has remained challenging. Here, we couple theoretical models of recruitment with a multispecies invasion experiment conducted within a natural gradient of soil moisture and productivity to assess how variation in establishment limitation shapes outcomes for invasion. Recruitment was affected by both density-dependent and density-independent sources of establishment limitation in three of four species. Soil moisture stress and productivity both increased density-independent mortality in one species, whereas density-dependent mortality increased in locations with favourable soil moisture. Synthesis. Successful establishment of invading species can be limited by both density-dependent and density-independent mechanisms. In particular, the strength of density-independent limitation may depend on natural gradients in abiotic factors. The varying strengths of establishment limitation suggest that patterns of invasion are likely to be uneven both in space and in time. Understanding how intraspecific competitive constraints and density-independent limitation vary with abiotic gradients can assist with predicting when invasions are likely to occur, information that can be harnessed in the development of better methods for control.
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.