MIDI Linked Data

MIDI Linked Data

MIDI Linked Data is the representation of symbolic music in MIDI format following the Web data publishing principles of Linked Data. Join us and contribute to extending the MIDI Linked Data Cloud!

Or browse a sample MIDI file as Linked Data.

The MIDI Linked Data Cloud

The MIDI Linked Data Cloud connects all MIDI files in the world in a giant graph of interconnected MIDI statements. To do this, we use Semantic Web technology, like RDF and SPARQL, to express MIDI information in the form of triples. These triples convey as much as human language in a way machines can process meaningfully. For example: Stairway to Heaven starts with an A, or that A is played in an accoustic guitar. Therefore, a MIDI pattern (song) is just the RDF graph containing all musical statements it consists of. Songs can also be connected among themselves through their shared resources (e.g. same instruments, same notes), and to the rest of the Web through common metadata (e.g. same bands, events, styles, etc.).

The purpose of the MIDI Linked Data Cloud dataset is to support researchers investigating interoperability of various music notations using Web standards; and to provide a challenging heterogeneous dataset for Entity Linking. But besides these, we expect the MIDI Linked Data Cloud to help research in a variety of tasks. For semantics: what does it mean for a song to convey sadness? What do all blues songs have in common? For music information retrieval: can we conceal the various music annotation methods following Web standards? Can we search music combining both text-based metadata and musical features? For interlinked musical databases: can we make MusicXML, MIDI, MEI, incipit databases interoperable?

About this last point, check out these for related efforts on making music notation interoperable through Linked Data:

Online MIDI to RDF conversion and linkage to the cloud

If you want to convert your MIDIs to Linked Data, and optionally integrate them to the growing MIDI Linked Data cloud, you don't need to install any software. Just use the online converter here!

brws it!

Surf the MIDI Linked Data Cloud using our brwsr instance. Just like humans like it!

SPARQL endpoint

Send your SPARQL to the endpoint, and query 10,215,557,355 triples of 308,443 interconnected MIDI files (check out the MIDI sources to see where those come from).

Check our SPARQL tutorial for further instructions and lots of examples!

API

An API on top of the SPARQL endpoint with basic search operations is accessible via grlc here.

Dumps

You can find the RDF dump archives of the MIDI Linked Data Cloud here.

midi2rdf tool suite

Convert your MIDI files to RDF, your MIDI-LD back to MIDI, or stream your instrument directly as RDF triples with the midi2rdf suite of tools.

SPARQL-DJ

Mix music using your favourite technology stack to mix data on the Web. Try out our prototype for MIDI mash-up generation, the SPARQL-DJ.

Learn more

Take a look at the project documentation and our tutorials. Don't hesitate to report bugs, and get in touch with us.

Contribute!

Join us now and make your voice (or another instrument) count!

If you want to report a bug or have a suggestion, please use the project's issue tracker (if you want to file an issue about a specific component, for example midi2rdf, please use their repo-specific issue trackers).

Publications

If you use the MIDI Linked Data Cloud in your research, please cite the following paper:

Albert Meroño-Peñuela, Rinke Hoekstra, Aldo Gangemi, Peter Bloem, Reinier de Valk, Bas Stringer, Berit Janssen, Victor de Boer, Alo Allik, Stefan Schlobach, Kevin Page. “The MIDI Linked Data Cloud”. In: Proceedings of the 16th International Semantic Web Conference (ISWC 2017), to appear (2017). (PDF)

Here's a list of other publications that deal with generating, enriching or using the MIDI Linked Data Cloud:

  • Albert Meroño-Peñuela, Rinke Hoekstra. “The Song Remains the Same: Lossless Conversion and Streaming of MIDI to RDF and Back”. In: 13th Extended Semantic Web Conference (ESWC 2016), posters and demos track. May 29th — June 2nd, Heraklion, Crete, Greece (2016). (PDF)
  • Albert Meroño-Peñuela, Rinke Hoekstra. “We’re Not Gonna Take It: Putting the Web at Musicians’ Service”. Digital Humanities Congress, University of Sheffield. September 8-10 (2016) (PDF)