Ten Songs of Tagore, In Data
Ten Rabindrasangeet encoded from akarmatrik swaralipi into open symbolic notation, with derived formats, synthesized audio, provenance, and an open verification path.
- Context
- NeelVerse Lab
- Status checked
- Public role
- Built by Indranil Banerjee
- Topics
- Knowledge systems · Memory and provenance · Open cultural data
Why it exists
This NeelVerse Lab artifact makes a small part of Rabindrasangeet computationally legible without substituting performance recordings for the written composition. Ten songs are encoded in a documented symbolic notation, then transformed into readable text, MIDI, MusicXML, and synthesized reference audio.
What can be inspected
- The canonical song data and schema.
- A listening page where audio is synthesized directly from the notation.
- An interactive method walkthrough.
- Provenance, licensing, integrity checks, and a public correction path.
- A blind continuation experiment comparing a model’s continuation with Tagore’s composition.
What makes the evidence useful
The artifact does not ask the visitor to trust a claim about digitization. The data, derived files, audio, method, verification record, and limitations are all separately inspectable.
Open the sources, not just the summary.
What this record does not claim.
The repository identifies this as a ten-song micro-dataset and records partial print-scan verification. This page does not generalise its findings to the full Rabindrasangeet corpus or publish a broader unapproved NeelVerse business thesis.