Hong Kong's Digital Archive Push Forces Hard Choices on Duplicate Image Replacement

Hong Kong's major public repositories are facing a reckoning. Libraries, museums and government record offices holding digitised photograph collections have accumulated years of duplicate image files — redundant scans, re-uploaded assets and legacy conversions — that are clogging storage infrastructure and complicating public access. The question of what replaces those files, and who decides, is now urgent.

The pressure has sharpened because several institutions are mid-cycle on technology upgrade contracts that expire before the end of 2026. The Hong Kong Public Libraries network, which operates 70 branches across the territory including the central branch on Harbour Road in Wan Chai, is one of the bodies understood to be reviewing its digital asset management platform. The Hong Kong Museum of History in Tsim Sha Tsui, currently running public programming tied to its permanent galleries, is separately managing a collections digitisation project that has produced overlapping image records requiring resolution.

Why the Duplicate Problem Compounds

Duplicate images are not simply a storage nuisance. When a photograph exists in three or four versions inside an archive — each tagged with slightly different metadata, different resolution specifications or different rights classifications — every downstream decision about publication, licensing or public display inherits that confusion. A researcher accessing the Hong Kong Film Archive's digital portal on Hoi Shing Road in Sham Shui Po, for example, may retrieve multiple versions of the same still without any clear indication of which is authoritative.

The problem is compounded by migration history. Many Hong Kong institutions transferred analogue holdings to digital formats during two distinct waves: one in the early 2000s following government efficiency drives, and a second accelerated round after 2020, when public health restrictions sharply reduced physical access and institutions rushed to expand remote availability. Each migration wave used different scanning standards. JPEG files created in 2003 sit alongside TIFF masters from 2021, sometimes tagged as equivalents when they are not.

Storage costs are real. Commercial cloud archiving for uncompressed image files at institutional scale runs at rates that make redundancy expensive. According to publicly available pricing from major enterprise cloud providers as of mid-2026, storing one terabyte of infrequently accessed archive data costs roughly US$23 per month. An institution holding 50 terabytes of duplicated assets is paying for capacity it does not need — money that could fund conservation work or public programming instead.

The Decisions That Cannot Be Deferred

Three choices are now unavoidable for any Hong Kong institution managing a digitised image collection. First, they must decide whether to pursue automated deduplication — software that identifies and flags redundant files — or manual curatorial review, which is slower but preserves human judgment about which version of a contested image holds the most archival value. For collections covering politically sensitive periods, that judgment call matters considerably.

Second, institutions must settle on a master file standard going forward. The archival community broadly favours uncompressed TIFF at 600 DPI for photographic originals, but that standard demands more storage than JPEG 2000 alternatives that some Hong Kong government IT procurement frameworks have historically preferred. Getting procurement and curatorial staff aligned is not automatic.

Third, and most consequential, is the rights question. When a duplicate is retired, the surviving master file must carry clean rights metadata. Given that significant portions of Hong Kong's photographic heritage were created by news organisations, some of which have restructured or closed since 2020, tracing copyright ownership for replacement-file clearance is a legal task, not just a technical one.

Institutions that defer these decisions will find the problem harder to solve after the current upgrade cycle closes. Collections managers at bodies operating under the Leisure and Cultural Services Department should expect interdepartmental guidance before the third quarter ends. For private archives and university libraries — including those at institutions in Kowloon Tong and Pokfulam — the timeline is self-imposed, but the technical logic is identical. The files need a decision, and the window to make it cleanly is narrowing.