How Ballarat's Digital Archives Ended Up Full of Duplicate Images — And What It Cost to Get Here

Ballarat's civic image library contains thousands of duplicate photographs — some files stored three or four times across separate servers — after more than a decade of disconnected digitisation drives failed to produce a single unified archive. The problem, long acknowledged inside the City of Ballarat's heritage and communications teams, has resurfaced as the council works toward a new digital asset management system expected to go live before the end of the 2026–27 financial year.

The timing matters because the duplication issue is not merely a storage headache. Cultural institutions across Ballarat, including the Art Gallery of Ballarat on Lydiard Street and the Museum of Australian Democracy at Eureka on Eureka Street, have spent portions of their own operational budgets maintaining parallel image collections that overlap significantly with council holdings. Every redundant file costs money to store, to license-check and to serve to the public. As regional arts funding tightens, those costs are harder to absorb.

A Chain of Competing Projects

The duplication problem did not arrive suddenly. It accumulated across at least four separate initiatives between roughly 2010 and 2023. Early digitisation work at the Ballarat Historical Society — based at the Ballarat Library precinct on Doveton Street — produced a catalogue that was never formally integrated with City of Ballarat records. Sovereign Hill's own photographic archive, which documents more than fifty years of living history programming, similarly developed in isolation, using proprietary metadata standards incompatible with state government frameworks.

When Heritage Victoria rolled out its state-wide Places database upgrades in the mid-2010s, participating Ballarat organisations uploaded images independently rather than drawing from a shared pool. The result was the same 1850s goldfields photograph appearing under different file names, different copyright attributions and different resolution specifications across multiple databases simultaneously. Staff at the Ballarat Heritage Office have described the situation — in general terms in council reports — as a legacy of good intentions outpacing coordination.

Ballarat Health Services encountered a related problem inside its own administrative systems, though on a smaller scale. When BHS digitised its historical facilities records as part of capital redevelopment planning for the Drummond Street North campus, image files from three different project consultants were folded into a shared drive without deduplication protocols. That process, completed in late 2023, reportedly produced more than 1,400 redundant image files from a collection of roughly 4,000 items.

What Deduplication Actually Involves

Fixing the problem is neither quick nor cheap. Industry benchmarks for cultural sector digital asset management suggest that a mid-sized regional archive — one holding between 20,000 and 80,000 image records — can expect deduplication and re-cataloguing work to run between $80,000 and $250,000 depending on the degree of manual review required. Automated tools can flag likely duplicates based on pixel-hash matching, but heritage collections frequently include near-duplicate images — slightly different exposures of the same scene — that require human judgment to classify correctly.

The City of Ballarat's draft 2026–27 budget, released for community consultation in May, allocated funding toward a digital infrastructure review that encompasses the image library problem, though the council has not publicly itemised a specific line for deduplication work. The Ballarat Regional Tourism body has separately flagged the issue in its own planning documents, given that Sovereign Hill's image assets underpin a significant share of regional promotional material distributed to state and national media.

For organisations waiting on a resolution, the practical advice from archivists is consistent: stop adding to the problem before starting to fix it. Any new digitisation project in Ballarat — whether at a school, a community hall or a sporting club — should adopt the Public Record Office Victoria's standard metadata schema from day one, and check for existing holdings before uploading new files. The cost of good filing habits at the front end is negligible. The cost of undoing five years of duplication, as institutions across this city are now discovering, is not.