MusicNet is a collection of 330 freely-licensed classical music recordings, together with over 1 million annotated labels indicating the precise time of each note in every recording, the instrument that plays each note, and the note's position in the metrical structure of the composition. The labels are acquired from musical scores aligned to recordings by dynamic time warping. The labels are verified by trained musicians; we estimate a labeling error rate of 4%. We offer the MusicNet labels to the machine learning and music communities as a resource for training models and a common benchmark for comparing results. This dataset was introduced in the paper "Learning Features of Music from Scratch." [1]

This repository consists of 3 top-level files:

musicnet.tar.gz - This file contains the MusicNet dataset itself, consisting of PCM-encoded audio wave files (.wav) and corresponding CSV-encoded note label files (.csv). The data is organized according to the train/test split described and used in "Invariances and Data Augmentation for Supervised Music Transcription". [2]

musicnet_metadata.csv - This file contains track-level information about recordings contained in MusicNet. The data and label files are named with MusicNet ids, which you can use to cross-index the data and labels with this metadata file.

musicnet_midis.tar.gz - This file contains the reference MIDI files used to construct the MusicNet labels.

A PyTorch interface for accessing the MusicNet dataset is available on GitHub. For an audio/visual introduction and summary of this dataset, see the MusicNet inspector, created by Jong Wook Kim. The audio recordings in MusicNet consist of Creative Commons licensed and Public Domain performances, sourced from the Isabella Stewart Gardner Museum, the European Archive Foundation, and Musopen. The provenance of specific recordings and midis are described in the metadata file.

[1] Learning Features of Music from Scratch. John Thickstun, Zaid Harchaoui, and Sham M. Kakade. In International Conference on Learning Representations (ICLR), 2017. ArXiv Report.

@inproceedings{thickstun2017learning, title={Learning Features of Music from Scratch}, author = {John Thickstun and Zaid Harchaoui and Sham M. Kakade}, year={2017}, booktitle = {International Conference on Learning Representations (ICLR)} }

[2] Invariances and Data Augmentation for Supervised Music Transcription. John Thickstun, Zaid Harchaoui, Dean P. Foster, and Sham M. Kakade. In International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 2018. ArXiv Report.

@inproceedings{thickstun2018invariances, title={Invariances and Data Augmentation for Supervised Music Transcription}, author = {John Thickstun and Zaid Harchaoui and Dean P. Foster and Sham M. Kakade}, year={2018}, booktitle = {International Conference on Acoustics, Speech, and Signal Processing (ICASSP)} }

MusicNet is a collection of 330 freely-licensed classical music recordings, together with over 1 million annotated labels indicating the precise time of each note in every recording, the instrument that plays each note, and the note s position in the metrical structure of the composition. The labels are acquired from musical scores aligned to recordings by dynamic time warping. The labels are verified by trained musicians; we estimate a labeling error rate of 4%. We offer the MusicNet labels to the machine learning and music communities as a resource for training models and a common benchmark for comparing results.

MusicNet offers music characterizations. This Data Set contains the results obtained after the analysis of pieces of MIDI music. The analysis was performed by applying the concept of Fundamental Scale: the set of symbols producing the minimal entropy when reading a text consisting of a large sequence of characters.

Each piece is characterized by its character-length. The results are presented in terms of symbolic diversity, entropy, and second order entropy. These characterizations of music allow to represent music pieces, composers, styles and periods in a 3-dimensional space with symbolic diversity, entropy, and second order entropy as its axis. The MIDI pieces analyzed for each style of music can be found at these DOIs:Medieval: 10.6084/m9.figshare.5435983Renaissance: 10.6084/m9.figshare.5435992Baroque: 10.6084/m9.figshare.5435995Classical: 10.6084/m9.figshare.5436010Romantic: 10.6084/m9.figshare.5436013Impressionistic: 10.6084/m9.figshare.5436016Twentieth: 10.6084/m9.figshare.5436019Chinese: 10.6084/m9.figshare.5436022Hindu Raga: 10.6084/m9.figshare.5436025Venezuelan: 10.6084/m9.figshare.5436106Rock: 10.6084/m9.figshare.5436031

Movie Themes: 10.6084/m9.figshare.5436034Rock: 10.6084/m9.figshare.5436031

Movie Themes: 10.6084/m9.figshare.5436034

Context MusicNet is a collection of 330 freely-licensed classical music recordings, together with over 1 million annotated labels indicating the precise time of each note in every recording, the instrument that plays each note, and the note's position in the metrical structure of the composition. The labels are acquired from musical scores aligned to recordings by dynamic time warping. The labels are verified by trained musicians; a labeling error rate of 4% has been estimated. The MusicNet labels are offered to the machine learning and music communities as a resource for training models and a common benchmark for comparing results.

Specifically, MusicNet labels is proposed as a tool to address the following tasks:

Identify the notes performed at specific times in a recording. Classify the instruments that perform in a recording. Classify the composer of a recording. Identify precise onset times of the notes in a recording. Predict the next note in a recording, conditioned on history. Content (Raw - recommended) The raw data is available in standard wav audio format, with corresponding label files in csv format. These data and label filenames are MusicNet ids, which you can use to cross-index the data, labels, and metadata files.

(Python) The Python version of the dataset is distributed as a NumPy npz file. This is a binary format specific to Python (WARNING: if you attempt to read this data in Python 3, you need to set encoding='latin1' when you call np.load or your process will hang without any informative error messages). This format has three dependencies:

Python - This version of MusicNet is distributed as a Python object. NumPy - The MusicNet features are stored in NumPy arrays. intervaltree - The MusicNet labels are stored in an IntervalTree. Acknowledgements The MusicNet labels apply exclusively to Creative Commons and Public Domain recordings, and as such we can distribute and re-distribute the MusicNet labels together with their corresponding recordings. The music that underlies MusicNet is sourced from the Isabella Stewart Gardner Museum, the European Archive, and Musopen.

This work was supported by the Washington Research Foundation Fund for Innovation in Data-Intensive Discovery, and the program "Learning in Machines and Brains" (CIFAR).

About this version:

This particular variant of the MusicNet dataset has been resampled to a 16 kHz-mono-16-bit-wav format, which makes it more suitable for certain audio processing tasks, particularly those that require lower sampling rates. We redistribute this data as a part of YourMT3 project. The license for redistribution is attached.

Moreover, this version of the dataset includes various split options derived from previous works on automatic music transcription as python dictionary (see README.md). Below is a brief description of available split options:

MUSICNET_SPLIT_INFO = {
  'train_mt3': [], # the first 300 songs are synth dataset, while the remaining 300 songs are acoustic dataset. 
  'train_mt3_synth' : [], # Note: this is not the synthetic dataset of EM (MIDI Pop 80K) nor pitch-augmented. Just recording of MusicNet MIDI, split by MT3 author's split. But not sure if they used this (maybe not).
  'train_mt3_acoustic': [],
  'validation_mt3': [1733, 1765, 1790, 1818, 2160, 2198, 2289, 2300, 2308, 2315, 2336, 2466, 2477, 2504, 2611],
  'validation_mt3_synth': [1733, 1765, 1790, 1818, 2160, 2198, 2289, 2300, 2308, 2315, 2336, 2466, 2477, 2504, 2611],
  'validation_mt3_acoustic': [1733, 1765, 1790, 1818, 2160, 2198, 2289, 2300, 2308, 2315, 2336, 2466, 2477, 2504, 2611],
  'test_mt3_acoustic': [1729, 1776, 1813, 1893, 2118, 2186, 2296, 2431, 2432, 2487, 2497, 2501, 2507, 2537, 2621],
  'train_thickstun': [], # the first 327 songs are synth dataset, while the remaining 327 songs are acoustic dataset. 
  'test_thickstun': [1819, 2303, 2382],
  'train_mt3_em': [], # 293 tracks. MT3 train set - 7 missing tracks[2194, 2211, 2227, 2230, 2292, 2305, 2310], ours
  'validation_mt3_em': [1733, 1765, 1790, 1818, 2160, 2198, 2289, 2300, 2308, 2315, 2336, 2466, 2477, 2504, 2611], # ours
  'test_mt3_em': [1729, 1776, 1813, 1893, 2118, 2186, 2296, 2431, 2432, 2487, 2497, 2501, 2507, 2537, 2621], # ours
  'train_em_table2' : [], # 317 tracks. Whole set - 7 missing tracks[2194, 2211, 2227, 2230, 2292, 2305, 2310] - 6 test_em
  'test_em_table2' : [2191, 2628, 2106, 2298, 1819, 2416], # strings and winds from Cheuk's split, using EM annotations
  'test_cheuk_table2' : [2191, 2628, 2106, 2298, 1819, 2416], # strings and winds from Cheuk's split, using Thickstun's annotations
}

About MusicNet:

The MusicNet dataset, originally released in 2016 by Thickstun et al., "Learning Features of Music from Scratch". It is a collection of music recordings annotated with labels for various tasks, such as automatic music transcription, instrument recognition, and genre classification. The original dataset contains over 330 hours of audio, sourced from various public domain recordings of classical music, and is labeled with instrument activations and note-wise annotations.

About MusicNet EM:

MusicNetEM are refined labels for the MusicNet dataset, in the form of MIDI files. They are aligned with the recordings, with onset timing within 32ms. They were created using an EM process, similar to the one described in the Ben Maman and Amit H. Bermano, "Unaligned Supervision for Automatic Music Transcription in The Wild". Their split (Table 2 of this paper) derived from another paper, Kin Wai Cheuk et al., "ReconVAT: A Semi-Supervised Automatic Music Transcription Framework for Low-Resource Real-World Data".

License:

CC-BY-4.0

epr-labs/music-net dataset hosted on Hugging Face and contributed by the HF Datasets community

UNDER CONSTRUCTION >

About this version:

This particular variant of the MusicNet dataset has been resampled to a 16 kHz-mono-16-bit-wav format, which makes it more suitable for certain audio processing tasks, particularly those that require lower sampling rates. We redistribute this data as a part of YourMT3 project. The license for redistribution is attached.

Moreover, this version of the dataset includes various split options derived from previous works on automatic music transcription as python dictionary (see README.md). Below is a brief description of available split options:

MUSICNET_SPLIT_INFO = { 'train_mt3': [], # the first 300 songs are synth dataset, while the remaining 300 songs are acoustic dataset. 'train_mt3_synth' : [], # Note: this is not the synthetic dataset of EM (MIDI Pop 80K) nor pitch-augmented. Just recording of MusicNet MIDI, split by MT3 author's split. But not sure if they used this (maybe not). 'train_mt3_acoustic': [], 'validation_mt3': [1733, 1765, 1790, 1818, 2160, 2198, 2289, 2300, 2308, 2315, 2336, 2466, 2477, 2504, 2611], 'validation_mt3_synth': [1733, 1765, 1790, 1818, 2160, 2198, 2289, 2300, 2308, 2315, 2336, 2466, 2477, 2504, 2611], 'validation_mt3_acoustic': [1733, 1765, 1790, 1818, 2160, 2198, 2289, 2300, 2308, 2315, 2336, 2466, 2477, 2504, 2611], 'test_mt3_acoustic': [1729, 1776, 1813, 1893, 2118, 2186, 2296, 2431, 2432, 2487, 2497, 2501, 2507, 2537, 2621], 'train_thickstun': [], # the first 327 songs are synth dataset, while the remaining 327 songs are acoustic dataset.
'test_thickstun': [1819, 2303, 2382], 'train_mt3_em': [], # 293 tracks. MT3 train set - 7 missing tracks[2194, 2211, 2227, 2230, 2292, 2305, 2310], ours 'validation_mt3_em': [1733, 1765, 1790, 1818, 2160, 2198, 2289, 2300, 2308, 2315, 2336, 2466, 2477, 2504, 2611], # ours 'test_mt3_em': [1729, 1776, 1813, 1893, 2118, 2186, 2296, 2431, 2432, 2487, 2497, 2501, 2507, 2537, 2621], # ours 'train_em_table2' : [], # 317 tracks. Whole set - 7 missing tracks[2194, 2211, 2227, 2230, 2292, 2305, 2310] - 6 test_em 'test_em_table2' : [2191, 2628, 2106, 2298, 1819, 2416], # strings and winds from Cheuk's split, using EM annotations 'test_cheuk_table2' : [2191, 2628, 2106, 2298, 1819, 2416], # strings and winds from Cheuk's split, using Thickstun's annotations }

About MusicNet:

The MusicNet dataset, originally released in 2016 by Thickstun et al., "Learning Features of Music from Scratch". It is a collection of music recordings annotated with labels for various tasks, such as automatic music transcription, instrument recognition, and genre classification. The original dataset contains over 330 hours of audio, sourced from various public domain recordings of classical music, and is labeled with instrument activations and note-wise annotations.

About MusicNet EM:

MusicNetEM are refined labels for the MusicNet dataset, in the form of MIDI files. They are aligned with the recordings, with onset timing within 32ms. They were created using an EM process, similar to the one described in the Ben Maman and Amit H. Bermano, "Unaligned Supervision for Automatic Music Transcription in The Wild". Their split (Table 2 of this paper) derived from another paper, Kin Wai Cheuk et al., "ReconVAT: A Semi-Supervised Automatic Music Transcription Framework for Low-Resource Real-World Data".

License:

CC-BY-4.0

Dataset

This dataset was created by Rohit Singh

Contents

Explore the historical Whois records related to royalty-free-classical-music.net (Domain). Get insights into ownership history and changes over time.

Tencent Music Entertainment, the music division of the Chinese tech giant Tencent, has demonstrated sustained profitability in recent years. In 2024, the company reported **** billion yuan in net profit with a robust growth of online music-paying users.

Music classification tree MusicNet and the data associated with top levels of the tree.

In 2024, South Korean music label Source Music posted a net profit of around **** billion South Korean won. While this represents a decrease from the previous year, it was also the second year of profit. Their only signed artist for the past few years, K-pop girl group GFRIEND disbanded in May of 2021, leaving the label without a direct income stream for the rest of the year. Having been acquired by HYBE Corporation in 2019, the label debuted a new girl group, LE SSERAFIM, in May of 2022.

seungheondoh/cmd-musicnet-metadata dataset hosted on Hugging Face and contributed by the HF Datasets community

Slovenian Folk Song Ballads 404 Slovenian folk song ballads with family themes, collected between the years 1819 and 1995. This dataset is an archive of the “Slovenian Folk Song Ballads” corpus (scores, measure maps, analyses, recordings, synchronizations, metadata). It provides both raw data and data for integration with the Dezrann music web platform: https://www.dezrann.net/explore/slovenian-folk-song-ballads. Family songs are folk song ballads, i.e., songs with relatively short but repetitive melodic sections through which the singer narrates a longer story through the lyrics. The melody and lyrics are known not to be fully dependent on each other, meaning, the same melody could be adapted for another lyric and vice versa. Thematically, they fall into the category of those that sing about family destinies, including several motifs from Slovenian, Slavic, and broader European themes. The content often focuses on describing the natural course of family life, from courtship to childbirth. The themes and motifs of family ballads (feudal, rural environment, the time of Turkish invasions, pilgrimages, etc.) revolve around socio-legal relations and both immediate and broader family matters in historical periods from which individual ballads originate. The collection of Slovenian folk song ballads contains transcribed field material collected by Slovenian ethnologists, folklorists, ethnomusicologists and various colleagues of Glasbenonarodopisni inštitut ZRC SAZU spanning from the years 1819 to 1995. Categorized thematically as family ballads, this collection features 404 folk songs, and includes the initial verse of the lyrics, extensive metadata, and musical analysis, encompassing contours, harmony, and song structure (melody and lyric) (see (Borsan et al., 2023) and (Borsan et al., under submission)). 23 of these songs have historical recordings. License: CC-BY-NC-SA-4.0 Maintainers: Vanessa Nina Borsan vanessa@algomus.fr, Mathieu Giraud mathieu@algomus.fr References (Borsan et al., submitted) https://dx.doi.org/10.57745/SINZFK https://www.algomus.fr/data Dataset content slovenian-folk-song-ballads.json. Main archive catalog, with metadata, as described on https://doc.dezrann.net/metadata score/: Scores measure-map/: Measure maps as defined by https://doi.org/10.1145/3625135.3625136, analysis/: Analyses, in .dez format, as described on https://doc.dezrann.net/dez-format audio/: Audio recordings synchro/: Sychronizations between musical time and audio time, as described on https://doc.dezrann.net/synchro

Explore the historical Whois records related to funny-music.net (Domain). Get insights into ownership history and changes over time.

The net cash of Indigo Books & Music with headquarters in Canada amounted to **** million Canadian dollars in 2022. The reported fiscal year ends on March 27.Compared to the earliest depicted value from 2018 this is a total increase by approximately ***** million Canadian dollars. The trend from 2018 to 2022 shows, however, that this increase did not happen continuously.

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.

Please find further information and a corresponding Python package on this Github page: https://github.com/CPJKU/msmd

If you use this dataset, please cite:
[1] Matthias Dorfer, Jan Hajič jr., Andreas Arzt, Harald Frostel, Gerhard Widmer.
Learning Audio-Sheet Music Correspondences for Cross-Modal Retrieval and Piece Identification (PDF).
Transactions of the International Society for Music Information Retrieval, issue 1, 2018.

The net income of Indigo Books & Music with headquarters in Canada amounted to ****** million Canadian dollars in 2022. The reported fiscal year ends on March 27.Compared to the earliest depicted value from 2018 this is a total decrease by approximately ***** million Canadian dollars. The trend from 2018 to 2022 shows, however, that this decrease did not happen continuously.

Dataset

This dataset was created by OCR Team Riad

Contents

Explore the historical Whois records related to c-music.net (Domain). Get insights into ownership history and changes over time.