Art · 7 min read
What Technical Imaging Reveals About Vermeer's Working Process
Element mapping and infrared and X-ray methods reveal revision beneath Vermeer's finished surfaces while leaving intention open.

Looking beneath a polished surface
Johannes Vermeer’s paintings often appear controlled to the point of inevitability. Small rooms, measured light, and carefully placed figures can suggest that the artist transferred a complete design to the canvas and refined it without hesitation. Technical examination gives a different, more material account. Beneath several finished surfaces are broad underlayers, altered objects, shifted contours, and pigment choices invisible to ordinary viewing.
This does not expose a secret formula for Vermeer. Imaging methods respond to particular materials, and not every layer can be reconstructed. A change under the surface shows that the work developed; it does not tell researchers exactly why the artist changed it. Technical evidence becomes strongest when it is combined with microscopy, conservation history, documentary research, and comparison across paintings.
Research for the Rijksmuseum’s 2023 Vermeer exhibition brought macro X-ray fluorescence scanning, reflectance imaging spectroscopy, macro X-ray powder diffraction, and high-resolution photography to paintings from several collections. The National Gallery of Art in Washington separately examined its four paintings then attributed to Vermeer. Together, the projects have changed the picture of his working process and reopened an attribution dispute.
Each method sees a different material trace
Conventional X-radiography records how X-rays pass through a painting. Dense materials, including pigments containing lead, often absorb more radiation and appear lighter. It can reveal changes, canvas structure, and paint distribution, but layers may overlap into a difficult composite.
Infrared reflectography detects infrared radiation reflected from or absorbed by materials. Some dark drawing substances become visible beneath upper paint, while others do not. The method is especially useful for carbon-containing underdrawing, but absence of a visible drawing is not evidence that no planning occurred.
Macro X-ray fluorescence, or MA-XRF, scans the surface point by point and maps characteristic X-rays emitted by chemical elements. A mercury map can indicate vermilion; lead may relate to lead white; copper can occur in several pigments or additives. Element maps are not automatically pigment maps. Researchers must interpret which compound likely produced an element and whether it belongs to an original layer, later restoration, ground, or contamination.
Reflectance imaging spectroscopy collects information across many wavelengths and can help distinguish materials by their spectral response. Microscopic paint samples, taken only where justified, show layer sequences and particles directly but represent a tiny location. No one technique produces a transparent photograph of “what Vermeer first intended.” The result is an inference assembled from complementary traces.
Underpaint was structural, not merely hidden
Technical research supports a working sequence in which Vermeer established composition and light through early paint stages. Broad dark and light masses could organize the scene before small surface effects were added. In some paintings, coarse brushwork and strongly contrasted underpaint sit beneath the quiet finish.
The National Gallery of Art reported that analysis of Girl with the Red Hat found a copper-containing material in black underpaint, probably used as a drying additive. Researchers interpreted the painting as evidence of a transition in which schematic forms and heightened contrasts once confined to underlayers became more visible in the final image. This is a specific account based on paint chemistry and comparison, not a general rule for every Vermeer.
Pigment maps also show economical and strategic material use. Expensive ultramarine made from lapis lazuli is famously associated with Vermeer, but identifying it does not by itself explain patronage, cost, or symbolism. Pigments were mixed, layered, and optically modified. A blue element map can clarify placement while leaving the artist’s intention open.
The polished surface is therefore the endpoint of a sequence, not evidence that the sequence was polished throughout. That finding matters because older admiration sometimes converted visual calm into a claim about effortless execution. Technical images show labor and revision without diminishing control in the finished painting.
The Milkmaid changed during painting
Research on the Rijksmuseum’s The Milkmaid found an underpainting and two objects later painted out: a jug holder behind the standing woman and a fire basket near the floor. Earlier X-radiographs had suggested something in the lower area; newer scanning made the forms clearer. Vermeer also adjusted the edges of the figure and objects as the composition developed.
Removing the jug holder simplified the wall. Replacing the fire basket with floor tiles and a foot warmer changed the lower-right area. Curators have argued that these decisions concentrated attention on the woman and strengthened the composition. That interpretation fits the visual result, but the scan does not record Vermeer’s reason. The objects might have been removed for spatial, narrative, symbolic, or practical reasons that cannot now be separated.
The finding also complicates claims that Vermeer always built scenes from a fixed optical projection. Scholars have long debated his possible use of a camera obscura, drawing on optical effects such as softened focus and bright disks of light. Technical evidence of changes does not rule out optical aids; an artist could use an optical device at one stage and revise freely at another. Nor does resemblance to optical effects establish that a device was used. No document settles the matter.
What the underlayers do establish is that at least some compositions were actively edited on the canvas. A theory of method must accommodate that revision rather than treating the finished geometry as a fully predetermined transfer.
Technical evidence can reopen attribution
Washington’s Girl with a Flute had long been attributed to Vermeer. After imaging and microscopic analysis, National Gallery researchers concluded in 2022 that it was made by an associate rather than by Vermeer himself. They cited differences in handling and layer structure alongside material similarities that suggested someone familiar with his practice and supplies. The museum proposed the possible existence of a studio, a consequential idea because no documentary record has established Vermeer pupils.
Other specialists have not treated the conclusion as final. Attribution combines connoisseurship, materials, technique, provenance, and comparative judgment. Shared pigments can indicate a shared time and place rather than a single hand. Different handling can reflect condition, experiment, chronology, or another maker. The National Gallery’s “studio” language is an inference from a cluster of evidence, not the discovery of a payroll or apprenticeship contract.
This case shows both the power and limit of technical examination. It can reveal that two surfaces were built differently and expose features unavailable to unaided vision. It cannot assign authorship by instrument alone. The meaning of a layer sequence depends on the comparison set and on assumptions about how much one artist’s practice can vary.
Condition and restoration complicate the image
Paintings have histories after the studio. Varnish yellows, pigments alter, surfaces abrade, canvases are lined, and restorers add and remove material. Vermeer’s use of ultramarine and yellow pigments can look different after centuries of chemical change. Dark passages may have lost modeling; highlights may be worn.
Technical teams distinguish original paint from later intervention through microscopy, imaging, records, and experience, but boundaries can remain uncertain. A digital reconstruction of an altered color is a hypothesis constrained by measurements. It should not be confused with the appearance witnessed in the seventeenth century, which also depended on lighting and viewing conditions we cannot reproduce exactly.
Large collaborative projects reduce the risk of interpreting one scan in isolation. The Rijksmuseum worked with the Mauritshuis and other lenders; Washington combined conservation science and curatorial analysis. Open publication and comparable methods allow disagreement to focus on evidence rather than on the authority of a single eye.
What the instruments establish
Technical imaging has established that several Vermeer paintings contain underpaint, compositional changes, and carefully differentiated materials. The Milkmaid once included objects later removed. The Washington pictures do not all share the same construction, and those differences bear on attribution.
What remains unknown is the full sequence of every painting, the reasons behind most changes, the exact role of optical devices, and whether material differences indicate chronology, experiment, condition, or another hand. Instruments widen the visible evidence; they do not eliminate interpretation.
The strongest reconsideration is modest. Vermeer’s finished calm was achieved through material decisions that could include broad setup, revision, deletion, and minute adjustment. Technical images make that work visible. They replace the myth of an untouched perfect plan with a better-supported account of a process—without pretending to watch the artist think.
Sources and further reading
4 sources, every one with a link you can follow.
- Rijksmuseum, “Vermeer Technical Studies” and research on The Milkmaid.
- National Gallery of Art, “New Findings Suggest the Existence of a Studio of Vermeer” (2022).
- National Gallery of Art, “Reconsidering Vermeer’s Perfectionism”.
- National Gallery, London, Vermeer and Technique.
