OrgRec VPO Collection

Provisional inventory and source availability audit | Version 0.1.0 | 10 September 2026

Scope and interpretation

This reference describes the retained evidence for a historical virtual pipe organ (VPO) discovery collection and supports precise scholarly citation. It does not contain sample payloads or establish complete reconstruction of the collection. The collection listing preserves all 119 rows of the historical acquisition inventory, two additional native-definition directory labels, one reported incomplete acquisition, and one uncataloged SoundFont package: 123 evidence records in total. These are not 123 verified independent organs.

Relationship to the source catalog

The accompanying All_VPO_Catalog.html lists all 3,600 retained discovery-catalog entries, including products, organ definitions, editions, configurations, and stop or rank components. Catalog inclusion does not mean acquisition. Its granularity is 646 dataset or product entries, 223 derived instruments or organ definitions, 46 edition or platform variants, 329 configuration variants, and 2,356 stop or rank components. Derived component labels are discovery metadata, not independently verified downloadable packages.

Historical collection state

A filesystem inspection dated 31 August 2026 reported 551,769 regular files, 582,098,581,740 logical bytes and 235 instrument-definition files. These historical figures were not remeasured for this release. The informal collection label of approximately 591 GB differs from acquisition notes describing approximately 591 GiB and from the inspection total. Scope, exclusions and measurement conventions remain to be reconciled. No collection-wide per-file cryptographic manifest is available in this release.

Acquisition evidence limitations

The inventory selected successful single-entity acquisition-log records and counted 105 distinct sampled-instrument name strings. This is not a verified count of complete sample sets or independent physical organs. One row identifies a website GIF and six identify demonstration MP3s. Legacy grouped records also contain known attribution errors. A missing historical exact path may refer to an archive removed after successful extraction, not a missing extracted dataset. The listed evidence cannot establish exhaustive collection membership without reconciliation against preserved payloads and original logs.

Availability audit method

The 10 September 2026 audit checks 197 unique primary source-page URLs from the entire retained catalog and ten reviewed direct archive URLs from the analytical subset. Secondary source URLs are retained but not probed. Source-page requests use HTTP HEAD. Direct archive requests use HEAD followed by a range request, reading at most 4,096 response-body bytes per request at application level before closing. Redirects are followed only to public HTTP or HTTPS destinations. No authentication, full-package transfer, extraction or cryptographic payload comparison is performed. A positive response is dated evidence of endpoint reachability, not proof of full retrievability, multipart completeness, version identity or permission to redistribute.

Analytical subset

Ten VPO entities contributed to the earlier comparative stop-timbre study and are marked in the listing. Their prior analysis-input lock covers ten selected definitions and 1,112 distinct source-audio hashes, producing 138 ranks and 1,206 pipe observations. This reference includes their reviewed source routes, but is not a standalone reproduction of that study. Cite the study evidence and exact OrgRec software release separately. Jukkasjärvi refers to the AB edition dated 2 December 2024, not the newer NC edition.

MODAVIS identity and organ link coverage

The retained VPO identifiers were joined exactly to canonical.externalIdentifier in a frozen snapshot of the MODAVIS Navigator public projection for Pipe Organ Dataset 1.6.0, retrieved on 2026-09-10. This projection contains 897 independently distributed virtual-instrument identities. Of the 3,600 discovery rows, 897 have unique explicit public mappings and 316 have accepted physical-organ links to 171 distinct organ identities. The other 2,703 rows are not returned in this public projection; this does not establish that they lack identifiers elsewhere in MODAVIS. No names were fuzzily matched and no parent relationships were inherited.

Collection specific MODAVIS results

All 119 historical acquisition-inventory entries have unique public MODAVIS identities; 22 have accepted organ links, representing 21 distinct organs. In the broader 123-entry collection-evidence listing, 122 have identities and 23 have organ links to 22 distinct organs. The additional linked entry is directory-label evidence and must not be treated as independently verified acquisition. All ten analyzed VPOs have public identities, but none has an accepted physical-organ link in this release. An unlinked record is not evidence that the instrument or its samples are unavailable.

Relationship meaning and frozen evidence

Only non-null organ targets with the published resolution state accepted_canonical_organ_relation count as accepted links. Relation type, source confidence, resolution state and evidence hashes are retained without reinterpretation. Links may describe sampling, inspiration, reconstruction or another relationship; they are not all recording-provenance claims. An organ identity does not establish sampled configuration, recording date, package completeness or download availability. Persistent w3id identifiers are clickable in the listings. The compressed API snapshot and its SHA-256, version-labelled source URLs, exact query and replay script preserve the basis for these counts. These are mappings to a published catalog, not independent corroboration of the original catalog facts.

Planned extensions

Future versions are intended to reconcile the inventory with preserved source payloads and original acquisition logs; establish a complete per-file SHA-256 manifest and measured collection size; resolve package identities, duplicates and incomplete acquisitions; document exact versions and all multipart dependencies; expand source-route, access and rights review; and provide a materializer that accepts authorized external downloads or local archives. Verification will distinguish an exact reference match from an incomplete or changed dataset variant. Additional spectral analyses and format adapters are separate extensions, not prerequisites for file-level preservation.

Citation and versioning

Prepared by Dominik Ukolov, version 0.1.0, 10 September 2026. Reserved DOI: 10.5281/zenodo.22693457. This record is an unpublished draft. DOI reservation does not make the package publicly accessible or register a published citation. After publication approval, cite the specific released version. Later substantive changes should be released as linked versions, preserving the evidence used by earlier citations.

Data stewardship and rights

Persistent identification, explicit provenance and machine-readable inventories follow the direction of the FAIR Guiding Principles (Wilkinson et al., 2016), without claiming full FAIR compliance. Public-facing files exclude storage-device names, absolute workstation paths, account details and private acquisition logs. Proposed publication licences are CC BY 4.0 for the authored documentation and compilation, and Apache-2.0 for the checking and rendering code. These licences grant no rights to third-party sample sets; historical licence labels are unverified discovery metadata.

Audit outcome by route type

MODAVIS coverage by scope

Method and interpretation; live Navigator filter. The live view can change; the bundled snapshot is the evidence for this release.

Collection evidence listing

References

Wilkinson et al. (2016), The FAIR Guiding Principles for scientific data management and stewardship, Scientific Data 3, 160018. doi:10.1038/sdata.2016.18.

Zenodo versioning guidance; Zenodo policies. Accessed 10 September 2026.