Vermeer's Camera

By Philip Steadman

Johannes Vermeer van Delft's
Johannes Vermeer van Delft's "The Geographer" (Reprinted with permission by the owner, Städelsches Kunstinstitut Frankfurt)

Portrait of Anthony van Leeuwenhoek by Johannes Verkolje
Portrait of Anthony van Leeuwenhoek by Johannes Verkolje (1686) [Rijksmuseum Amsterdam].
For more than a hundred years it has been suggested that the great Dutch master Johannes Vermeer (1632-75) made use of the camera obscura as an aid to painting. The camera obscura ('dark room') was the predecessor of the photographic camera. It is a simple optical device incorporating a pinhole or - from the 16th century onwards - a lens, with which an image of a scene can be projected onto a screen. The image can then be traced and copied. Some early camera obscuras took the form of blacked-out booths, tents or rooms (hence the term) in which the observer was enclosed. Later versions took the form of small boxes like photographic plate cameras or reflex cameras. Medieval and early modern astronomers used the camera for observing the sun without damaging their eyes. Leonardo da Vinci was the first person to suggest possible uses in art, and by the early 17th century the camera was being described in perspective manuals intended for architects and painters.

It was the American etcher and lithographer Joseph Pennell, writing in the Journal of the Camera Club in 1891, who first speculated that Vermeer might have used some optical aid. Since then the idea has been taken up again and again, notably by the painter Lawrence Gowing in his beautiful monograph Vermeer (1952), and the art historian Svetlana Alpers in The Art of Describing (1983), among many others. This belief does not however rest on any documentary evidence. Very little is known about Vermeer's life other than what can be gleaned from official records, and we do not know whom he studied under or who his pupils were, if indeed he had any. No drawing by his hand has survived. The belief is based rather on certain properties of the paintings themselves.

Pennell drew attention to the 'photographic perspective' of certain pictures such as Officer and Laughing Girl, in which the figure of the soldier is very close to the painting's viewpoint and appears disproportionately large. Other writers have pointed to the way in which Vermeer seems to reproduce in paint some idiosyncrasies of optical images and 'out-of-focus' effects, that would be visible on the camera's screen but not to the naked eye. One example is in his treatment of highlights: reflections of sources of light off shiny surfaces such as metal, ceramics or polished wood. When Vermeer was painting indoors, by north light, the illumination would have come from the windows, whose reflections as highlights would have been rectangular in shape - perhaps distorted when seen on curved surfaces like those of glasses or jugs. However Vermeer uses blobs of light-colored pigment to represent these as perfect circles. It has been suggested that he is copying here from slightly unfocused images, in which small bright patches of light are spread by the lens into 'discs of confusion'. In other places Vermeer renders certain details - the skeins of thread in The Lacemaker, sculptured brass lions' heads on the backs of chairs - in a particularly loose and schematic style, which is 'photographically' true to tone and hue, but nevertheless lacks sharpness or precise outline.

Scholars have discussed who might have been the source of Vermeer's knowledge of optics and lenses - although it has to be said that there are absolutely no documents to support any of these speculations. The names of the painters Carel Fabritius and Samuel van Hoogstraten have been mentioned. Both men were fascinated by perspective illusion and trompe l'oeil, and van Hoogstraten wrote of seeing cameras on several occasions. He had met with Jesuit scholars and tried out their optical instruments at the Imperial Court in Vienna in the 1650s. Vermeer possessed canvases by both men, a fact that suggests he knew them personally. Otherwise the main candidate for the role of Vermeer's 'optical consultant' is Anthony van Leeuwenhoek, the pioneer of microscopy and the first man ever to observe protozoa and bacteria. Vermeer and Leeuwenhoek were exact contemporaries and lived a few streets apart in Delft, not a big town. What is more, when Vermeer died in debt, Leeuwenhoek was appointed by the town's Aldermen as the curator of his estate, a fact that some historians have interpreted as evidence of a friendship. It has even been suggested that Leeuwenhoek was the sitter for Vermeer's two 'scientific' paintings of scholars in their studies, The Astronomer and The Geographer (see illustration). Both bear curious resemblances to known portraits of Leeuwenhoek - although the identification remains controversial.

It is worth mentioning a couple of further scraps of biographical information about Vermeer. He is known to have received two distinguished visitors in Delft, who are usually described in the context as connoisseurs of painting, but who also had extensive connections in the world of European optical science. One was Constantijn Huygens, secretary to the Prince of Orange and father of the astronomer Christiaan. The second was the French diplomat Balthasar de Monconys, who made significant developments in the design of microscopes. Huygens and Monconys were sufficiently prominent in this field to be received together at a meeting of the Royal Society in London in 1663. Both knew about the camera obscura, indeed Huygens himself possessed a portable instrument that he bought in London in the 1620s and took back to show to his artist friends in Holland. (Whether Vermeer was among them we do not know.)

I have approached this old subject from a new direction: via the perspective geometry of the pictures. Vermeer produced more than two dozen paintings of domestic interiors with one, two or three figures. A few of them depict distinctive and unique spaces. But the majority seem to show just a few rooms repeatedly, with sitters and furniture rearranged. It is possible to draw some tentative conclusions about how many rooms might be involved, by making an inventory of their architectural features: floor surfaces, wooden ceilings, some very characteristic decorative patterns of leading in the window panes. As many as ten pictures appear - with a few anomalies - to represent the very same room. I have tested this possibility more rigorously, by reconstructing the geometry of the spaces of the paintings through a process that, in effect, reverses the conventional method for setting up perspective views. It is possible to do this for paintings, like Vermeer's, that show rooms whose floors are tiled. The actual dimensions of the spaces, and the shapes and sizes of pieces of furniture, can all be calculated with some precision. This is because Vermeer depicts a number of recognizable objects - chairs, Delftware tiles, musical instruments, wall-maps and paintings by other artists - which all survive in museums today and whose sizes are therefore known. It turns out that the ten paintings do indeed depict a single room whose dimensions are broadly consistent throughout.

In one painting, The Music Lesson, Vermeer includes a mirror that reflects his own vantage point and a small part of the back wall behind him, not otherwise visible in any picture. This evidence provides a measurement for the overall length of the room. The dimension is nicely compatible with the width in plan of both of the houses in which it is believed Vermeer may have had studios: the inn called Mechelen on the Market Place owned by his family, and a house nearby owned by his mother-in-law Maria Thins. It is my belief that the room in the pictures in question was in fact in Maria Thins's house.

How does all this relate to the camera obscura? The perspective reconstructions make it possible to plot the positions in space of the theoretical viewpoints of the ten relevant paintings. Everything that can be seen in each picture is contained in a 'visual pyramid' (a pyramid on its side), whose apex is at the viewpoint. Suppose that the sloping edge-lines of this pyramid are extended back, through the viewpoint, to meet the room's back wall. They define a rectangle on that wall. In six out of the ten cases this rectangle is the exact size of Vermeer's painting. The geometry is that of a booth-type camera, with its lens at the painting's viewpoint and the screen on the wall. The image is the same size as Vermeer's canvas because he has traced it. The image would have been upside down, and might have been reversed left-to-right. But there are several ways - including optical ways - in which Vermeer could have reversed the image, all of them reasonably practicable and straightforward. Some writers have argued that Vermeer constructed his perspectives mathematically, or that he used mirrors or other mechanical aids to painting. But it is difficult to see how these could account for the very curious results shown by the geometrical analysis, that are so readily accounted for by the camera hypothesis.

I have had a one-sixth scale physical model made of the room, with a photographic plate camera in the place of Vermeer's camera obscura. Using this it has been possible to create photographic simulations of the paintings, and test phenomena such as the patterns of light and shadow, all of which Vermeer reproduces - in general - with great fidelity. Recently a team in California led by Jonathan Erland has been building a 3D computer model, that should provide a more flexible tool for experiment. The BBC built a full-size reconstruction for a television film, with a booth-type camera incorporating a simple bi-convex lens of 10 cm diameter. This cast an image of The Music Lesson, at full size, onto a translucent screen, that was quite bright enough to film. These experiments have served to test the basic geometrical findings of the perspective reconstructions, and have shown the feasibility of casting images at the sizes of Vermeer's canvases using very modest optical technology. What has long been suspected - that Vermeer was a camera user - is now confirmed.

Philip Steadman is a professor of architecture and town planning at University College London. His book Vermeer's Camera, the product of twenty years' fascination with the painter, will be published by Oxford University Press in February 2001. More details will be available at

APS encourages the redistribution of the materials included in this newspaper provided that attribution to the source is noted and the materials are not truncated or changed.

Editor: Alan Chodos
Associate Editor: Jennifer Ouellette

December 2000 (Volume 9, Number 11)

Table of Contents

APS News Archives

Contact APS News Editor

Articles in this Issue
APS Launches New Web Site for the Public
Three Budding Young Physicists are New Apker Recipients
2000 Nobel Prizes Announced in Physics, Chemistry, Medicine
New Scientific Coalition Targets Climate Change on the Hill
Better Tabletop Accelerators, Fusion in a Beer Can Featured at DPP
First RHIC Results Highlight 2000 DNP Meeting
Assessment Tests Can Marginalize Science Education
Palestinian Visits: Irresponsible or Just Ill-Timed?
Physicists Honored at Annual DPP, DFD Meetings
New Fellowship in Washington Office
2000 Year-End Gift Ideas: Help Physics Programs
Editorial Cartoon
Zero Gravity: The Lighter Side of Science
This Month in Physics History
Meeting Briefs
Members in the Media
The Back Page