Every published object is open access.◆No gatekeeping. All objects published openly.◆PHOTARCH — research infrastructure for archaeology.◆Every published object is open access.◆No gatekeeping. All objects published openly.◆PHOTARCH — research infrastructure for archaeology.◆

Standards

How PHOTARCH aligns with — and differs from — existing imaging and metadata standards.

The need for standardisation in cultural heritage imaging has long been recognised. Institutions such as the Library of Congress and the National Library of the Netherlands have published detailed technical guidelines to ensure the preservation quality, colour fidelity, and long-term accessibility of digital surrogates (FADGI, 2023; Metamorfoze, 2025). PHOTARCH builds upon the foundation laid by these initiatives, while adapting their principles to the specific demands of archaeological artefact photography. Unlike flat document scanning or general digitisation workflows, photographing three-dimensional, textured, and context-rich objects in archaeology requires additional considerations: angle control, physical scale, lighting structure, and a consistent visual logic across entire collections.

What the existing standards govern

The FADGI guidelines (Federal Agencies Digital Guidelines Initiative) outline a multi-star rating system with measurable targets for resolution, noise, colour accuracy and tonal response. Metamorfoze — the Dutch national preservation programme, run by the KB and the Nationaal Archief — sets minimum image quality levels for digitisation in libraries and archives. ISO 19264-1 is the device-quality standard both now lean on. Their technical expectations carry over usefully to photography: ICC profiles, colour targets, control of tonal range.

It is worth being exact about what they do not cover, because the gap is the reason this method exists. FADGI's 2023 edition has no chapter on three-dimensional objects at all: its fifteen material types (chapter 3, sections 3.1–3.15) run from bound volumes and documents to photographs, paintings, X-ray film and microfilm. Metamorfoze addresses flat originals. ISO 19264-1 measures the equipment, against a test target, and says nothing about the photograph of an object. None of the three specifies a background, a light direction, whether a shadow is kept, or how the size of the thing photographed is recorded.

That is not a criticism of them. They were written for books, manuscripts and archival sheets, and they do that work well. But an artefact is not a page.

The three standards, one by one

PHOTARCH does not assess objects by them. It assesses them by its own three levels — see What a photograph is fit for below — and offers the three standards, clearly labelled, for paintings, documents, prints and other flat originals, which is what they were written for. Each has a page of its own: what it is, what it measures, its levels, and how to relate to it when the original is an object.

  • FADGI — the US federal guidelines. Four star levels, numbers per type of material, and conformance as a whole programme rather than a single file.
  • ISO 19264-1 — the international standard the other two align with. It measures the camera or scanner on a test target, in levels A, B and C.
  • Metamorfoze — the Dutch preservation guidelines for paper. Full, Light and Extra Light, and a master that must be able to replace the original.

Where nothing exists yet

0international standards specifying background, light direction, shadow or scale for three-dimensional museum objects
1national object-level standard found — CHIN's, which chooses grey or black and shadowless capture
0studies found that test whether a museum object is read better on a white ground or a grey one

This is not a gap invented in order to be filled. It is the condition the method was developed in, and it is why PHOTARCH specifies things no imaging standard does: where the light stands, what the object lies on, whether its shadow is kept, and how many millimetres one pixel covers. On one point it goes against the standards rather than beyond them: they ask for light that falls evenly across the field, and PHOTARCH gives its light a direction on purpose.

The third figure is an open question, and it cuts both ways. The case for a white ground made under Principles & Setup rests on perception research and on practice, not on a test of object photographs, and the case for grey has no such test either.

One more absence belongs here, because it is PHOTARCH's own to fill. Measuring artefacts from calibrated photographs is established practice — lithic studies have checked such measurements against calipers. What has not been published is how accurate a scale carried in the file is: the board's markers against calipers, across cameras, lenses and institutions. Until it is, the uncertainty each PHOTARCH file states is worked out from the measurement itself, and says what it covers; nobody has yet tested it from outside. The study can be run on any collection photographed on the board.

Where standards are silent, principles are not

No standard says how an object should be photographed. But the field holds itself to principles that bear directly on it, and a method is judged against them before anyone opens an ISO document. These are the ones PHOTARCH is built against.

Collections are documented to a standard, and the records are open

ICOM's Code of Ethics for Museums, revised in June 2026, asks museums to document their collections following established standards and procedures, and to make complete records public, on site and online, apart from what is confidential or sensitive. A method is what lets a photograph be part of such a record: the same light, ground and scale for every object, so that one institution's record can stand beside another's.

SourceICOM Code of Ethics for Museums (2026), C.5
A digital copy respects the original, and says what it is

The same code asks that digital copies respect the integrity of the original and are marked as copies, so that nobody takes them for more than they are, and that digital methods keep records accurate. Every picture the archive publishes, and every file downloaded from it, carries a signed statement of what it is and what was done to it, and so does an export from a licensed copy of the PHOTARCH app — see Trust & Provenance.

SourceICOM Code of Ethics for Museums (2026), C.2 and C.7
Show how the picture was made

The London Charter, written for computer-based visualisation of heritage, asks that the decisions behind a visualisation are documented so that its claims can be judged, and calls that record paradata. PHOTARCH borrows the idea for photographs. The scale in each file says how it was measured, when and by which software, and with what uncertainty wherever there is a real figure; a seal lists what was done to the picture; and the app's grades name what was measured and what was not.

SourceThe London Charter 2.1 (2009), §4.6
Findable, accessible, interoperable, reusable

The FAIR principles ask that data can be found, reached, combined and reused by machines as well as by people. In PHOTARCH they map one to one: an ARK identifier for every object; viewing over the open web with no account; LIDO, Dublin Core, Linked Art and IIIF, with terms from the Getty AAT; a licence in every file; and a signed provenance record in every file the archive publishes.

SourceWilkinson et al. (2016), Scientific Data
Communities keep authority over sensitive heritage

The CARE principles for Indigenous data governance — collective benefit, authority to control, responsibility, ethics — give communities authority over data about them, and ICOM's 2026 code asks museums to answer requests to limit access to sensitive material promptly and openly, online as well as on site. In Nordic collections this concerns Sámi heritage first. PHOTARCH does not decide it for anyone: an object can stay unpublished, a find spot can be withheld, and internal fields are never published. The decision belongs to the holding institution and, where it applies, to the community.

SourceCARE Principles (GIDA 2019; Carroll et al. 2020) · ICOM Code of Ethics (2026), E.5

Correction, not retouching

Scientific publishing drew the line between correcting a picture and changing what it shows in 2004, and the line has held since. Brightness, contrast and colour balance may be adjusted if the change applies to the whole image and hides nothing that was there; no single feature may be enhanced, obscured, moved, removed or introduced; and the original data are kept.

PHOTARCH's colour correction from the reference card is of the first kind: one correction for the whole picture, the same for every photograph in the sitting. The master is never rewritten. The white ground is of the second kind, and it should be said plainly. The ground around the object is taken to white through a mask, dust on the ground can be removed, and a few presentation settings act on the object or on the ground alone. The object's own pixels are left out of the whitening; the mask can be corrected by hand; the shadow is kept; and all of it is applied to copies, never to the master, and kept with the photograph as settings rather than baked into it.

The same rule carries a warning that applies directly: a background that drops out completely can take faint detail with it. That is why the whitening stops at the object rather than at a brightness threshold, and why the master — where nothing has been whitened — is kept. The seal on an exported file does not yet name these steps one by one: a whitened ground is recorded in it as a colour adjustment.

Archaeological imaging: Specific needs

Archaeological artefacts are materially complex, often irregular, and typically require multi-angle documentation. They are also tightly linked to context — spatial, typological, and comparative — which places higher demands on metadata and repeatability. A single image of a ceramic sherd is not simply a visual object; it is part of a broader system of interpretation involving typology, chronology, and provenance. PHOTARCH addresses these needs by extending existing image standards into a domain-specific method. It integrates lighting geometry and physical reference tools to allow direct comparison between objects from different collections, countries, and excavations. This supplements — rather than replaces — standards like FADGI and Metamorfoze.

Metadata standards

Every object record in PHOTARCH is mapped to established descriptive standards.

Dublin Core provides the base layer of interoperability. The fifteen core elements — title, type, date, creator, rights, and so on — are populated for every object and available for export as XML. This makes records readable by any system that understands Dublin Core, including institutional repositories and library catalogues.

LIDO (Lightweight Information Describing Objects) is the standard used by European museum networks and aggregators. PHOTARCH generates a LIDO v1.0 XML file per object, covering object classification, repository information, production event, material, technique, dating, find spot, and rights. This export is intended for submission to national heritage portals and cross-collection discovery platforms.

EXIF and IPTC. The master is never altered: its camera-generated EXIF data (capture date, focal length, exposure settings) stays as the camera wrote it. The descriptive fields (title, caption, rights, photographer) are written into every copy when it is downloaded or exported. This ensures that every file that leaves the archive carries its own documentation regardless of where it ends up.

Scale carried in the file

A photograph only shows an object's size if the viewer knows how large one pixel is on the object. PHOTARCH writes that into every exported image of a measured object as Image Scale Metadata (ISM): a few XMP fields giving millimetres per pixel, the pixel grid the scale applies to, and how it was measured — with a ruler in the picture, with calipers on the object, or from a target. Any program can read the fields and draw a true scale bar, also after the file has been resized or turned; a copy cropped to another shape is recognised, and its scale is not used. The TIFF resolution fields are set to match, and the archive's IIIF manifests carry the same scale through IIIF's physical dimensions service.

The design is not novel, and it is not meant to be. Medical imaging has carried its scale as a field since the 1990s: DICOM records the spacing of pixels on the detector and the spacing in the patient as two separate numbers, forbids the one from standing in for the other, and has a field for how the calibration was obtained. ISM asks a photograph for the same three things — the number, the grid it belongs to, and how it was arrived at — in a handful of fields rather than an industry's worth. What is missing is not the idea but a place to put it: no imaging standard for cultural heritage carries a scale at all.

ISM is an open format, independent of PHOTARCH, and any software may write or read it. The draft specification (CC BY 4.0), a reference library (MIT), test files, a validator and a viewer that draws the scale out of a file are published at image-scale-metadata.github.io, in an organisation of its own without PHOTARCH in the name.

Content Credentials

Since 2026 every published picture, and every file downloaded from the archive, carries a C2PA manifest: a signed record of how the picture was made — photographed, colour adjusted, cropped, published — and of the scale it holds, countersigned by an independent timestamp authority so that it keeps proving itself after the signing certificate expires.

The scale therefore appears twice in the same file. In the XMP it can be read by anything and changed by anything. Inside the manifest it is signed: it cannot be altered, only broken, and a reader is told which happened. That is the difference between a statement and an attestation.

Until the product has been through C2PA's Conformance Program the certificate is PHOTARCH's own, so a validator will report a valid signature by an unknown signer. Everything else in the manifest is already correct.

Linked Art and controlled vocabulary

Object records publish as Linked Art — CIDOC-CRM expressed in JSON-LD — which is the form museum systems and aggregators read. Material, technique and period are mapped to the Getty Art & Architecture Thesaurus, so that a word in a PHOTARCH record means what it means elsewhere rather than only here.

What a photograph is fit for

The PHOTARCH app assesses every photograph by PHOTARCH's own three levels. They say what the picture can be used for, and each requirement is measured, not declared:

Record shot Study Reproduction
Resolution at the object ≥ 4 px/mm (≈ 100 ppi), when there is a scale ≥ 12 px/mm (≈ 300 ppi): prints at life size at 300 dpi ≥ 24 px/mm (≈ 600 ppi): prints at twice life size
Scale not required set, with its method and uncertainty from calipers or the board, not a ruler; uncertainty ≤ 1 %
Colour not required a ColorChecker in the sitting; mean ΔE2000 ≤ 5 mean ΔE2000 ≤ 3, and the 90th percentile ≤ 6
Exposure — ≤ 1 % of the object clipped ≤ 0.1 % of the object clipped
Capture and file any a raw, or a lossless file such as a focus stack in TIFF or PSD; ≥ 10 bits; a lossless master a raw or a lossless file; ≥ 12 bits; a 16-bit TIFF in Adobe RGB or wider
Sharpness — ≥ 80 % of the object in focus ≥ 95 % of the object in focus

The two figures for sharpness are provisional. Sharpness is measured as a share of the object: how much of it is in focus. Areas with too little texture cannot be judged either way; they are left out of the share, and the app reports how much was left out. Before the picture is taken, the Sharp, Exposure and Board lights in Capture say whether the middle of the frame is sharp, whether highlights are clipping and whether the board reads flat and square.

Two things stay outside the assessment, and are said to: evenness of light, which is not a goal, because the method lights an object from one side so that its form and shadow read (see Principles & Setup); and the direction of the light, which the photographer records with the pictures.

For a raw, the bits are those the camera recorded, read from the raw file's own header, not those of the file they were written into. An ordinary 14-bit raw can therefore reach Reproduction, and an iPhone's 10-bit ProRAW can reach Study. A finished picture brought in as a lossless file — a focus stack exported as a TIFF, or kept as a PSD with its exposures as layers — is judged like any other on the bits it holds: a 16-bit stack can reach Reproduction, an 8-bit one is a record shot. A PSD's master is written as a TIFF of its composite. A picture is exported at the depth it was photographed in: 8 bits stay 8, and anything deeper is written as 16, with the recorded number in the file.

The imaging standards, for flat originals

FADGI, ISO 19264-1 and Metamorfoze are not shown by default. A setting for flat originals — paintings, documents, prints — turns them on, each clearly labelled, and the app then reports the level each standard reaches on what it can measure: the file format and colour space, the resolution at the object from its measured scale, and colour accuracy, white balance and tone from the ColorChecker photographed in the same sitting. In this mode bit depth follows FADGI's own rule and is judged on the file, with the bits the camera recorded shown beside it. Flat originals are lit evenly from two sides; the app does not measure how evenly.

The app also says what it does not measure for these standards: their sharpness on a test target, sharpening, noise, evenness of light, channel misregistration and gain modulation. Two of the levels are indirect, and are labelled so. The ISO level is derived through FADGI's correspondence, not from a test target analysed under ISO 19264-1; and the Metamorfoze colour figures are indicative, because Metamorfoze measures on a ColorChecker Digital SG and PHOTARCH reads a ColorChecker Classic.

A green light therefore means nothing measured stands in the way. It is not a claim of certified conformance, and the application says so rather than letting a level be mistaken for a certificate.

Persistent identification

Every object in PHOTARCH carries an ARK persistent identifier (Archival Resource Key) under NAAN 73195. ARKs are resolvable, stable identifiers that remain valid even if the URL structure of the archive changes. The identifier appears on each object record and resolves via the PHOTARCH resolver at app.photarch.com/id/.

ARKs were developed by the California Digital Library and are used by major institutions including the Bibliothèque nationale de France and the Internet Archive. They carry no licensing fees and require no institutional membership.

File integrity

Every master file in PHOTARCH is assigned a SHA-256 checksum at the time of upload. The checksum is displayed on each object record and included in bulk CSV exports. This allows institutions and researchers to verify that a file has not been altered since it was ingested — a basic requirement for archival trust.

Access and export

PHOTARCH supports the following export formats for all object records:

  • Dublin Core XML — standard descriptive metadata, suitable for OAI-PMH and library systems
  • LIDO XML — structured museum metadata for aggregation with heritage portals
  • CSV (bulk dataset) — every published object with its metadata and image URL, open to anyone at app.photarch.com/dataset.csv and rebuilt nightly
  • ARK identifier — persistent URI for citation and linking

Toward a machine-readable future

One of PHOTARCH’s key ambitions is to make archaeological photography machine-readable without sacrificing human interpretability. By structuring camera angles, embedding metadata, and ensuring visual consistency, the method supports not only comparison and education, but also AI-driven categorisation and automated analysis. As cultural heritage data becomes increasingly digitised, consistent visual documentation grows in both technical and ethical importance. A photographic method followed the same way every time ensures not only preservation and accessibility — but also the ability to see, study, and share artefacts more clearly across generations.

References · imaging standards

  • guidelineFADGI (2023). Technical Guidelines for Digitizing Cultural Heritage Materials, 3rd ed. Library of Congress.Flat, bound, paintings, film. No chapter on three-dimensional objects.
  • guidelineMetamorfoze (2025). Preservation Imaging Guidelines, version 2.0 (first edition 2012). KB / Nationaal Archief.The Dutch national preservation programme, for flat originals.
  • standardISO 19264-1:2021. Imaging systems quality analysis — reflective originals.Device metrics on a test target: OECF, ΔE, SFR, noise, uniformity.
  • standardISO/TR 19263-1:2017. Best practices for digital image capture of cultural heritage material.
  • institutionalRijksmuseum (2017). Manual for the Photography of 3D Objects.One of two object-level studio standards found, and the most authoritative.
  • institutionalCHIN. Canadian Museum of Civilization digitisation standards.The one national object-level standard found; grey or black ground, shadowless.

References · principles

References · metadata and provenance

← GuidesMetadata →