Johann Sebastian Bach Chorales in Symbolic Music Notation

Johann Sebastian Bach – The Chorales is a dataset from the DCML corpus initiative, originating from engraved Capella files compiled by Ulrich Kaiser. The data has been converted to MuseScore format and parsed using the ms3 Python library, though the conversion introduced incorrect accidentals. The dataset is released under a CC0 1.0 Universal license.

Use Cases

• Musicological analysis of Bach's four-part chorale harmonization based on the symbolic notation.
• Training or evaluating music transcription models based on the parsed MuseScore data.
• Comparative studies of chorale datasets by using pitch class vectors for correction, as suggested in the description.

Strengths

• Dataset originates from a known scholarly source, the DCML corpus initiative and Ulrich Kaiser's work.
• Provides a list of 50 chorale files with associated measures and labels, as shown in the overview snippet.

Limitations

• The conversion to MuseScore format resulted in many incorrect accidentals, requiring correction.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file size are unknown, which may limit suitability assessment.

Provenance

Source

DCML corpus initiative, based on Capella files engraved by Ulrich Kaiser.

Collection Method

Original Capella files were converted to MuseScore 3.6.2 format and parsed using the ms3 Python library.