Review of Traces of Vermeer by Jane Jelley
Oxford University Press, NY, NY, 2017
336 pp., illus. 85 col.
Ever since Lawrence Gowing examined x-ray images of Johannes Vermeer’s Girl with a Pearl Earring (c. 1665) in 1950 and was struck by the lack of pentimenti, preliminary lines, and other marks routinely found in underdrawings in comparable paintings of the Dutch Golden Age, there has been speculation about how Vermeer created his works, which appear “fully formed.” This x-ray evidence, geometrical perspective, and distinctive “optical” effects in other paintings, such as blur spots resembling those of an out-of-focus portion of an optically projected image, led a number of scholars to propose that Vermeer employed optical devices during the execution of his works—either directly for guiding the placement of his marks, or indirectly, as a visual referent. Such optical theories have attracted the attention of scholars in the history of art, history of optics, rigorous optical analysis, computer image analysis, and even a few intrepid experimenters—and of course the art-loving public. Curators such as the late Walter Liedtke of the Metropolitan Museum of Art, who argued forcefully that Vermeer did not use optics directly, and Arthur K. Wheelock, of the National Gallery Washington, who countered that he did (at least in part), have both presented a range of material, visual and historical evidence to bolster their conclusions. Liedtke stressed, for instance, that many scenes were in large part “fictive,” with no direct source and hence no corresponding optical image, while Wheelock highlighted the appearance of blur spots.
Such debates were expanded by scientists and technologists who performed rigorous optical analyses, experiments, and putative “re-enactments.” The perspective in works such as in Officer and Laughing Girl (1655-60), which is “accelerated” (i.e., the center of projection is close to the scene), also suggests to some that Vermeer traced a projected image on his canvas. Most notable among these technologists was Philip Steadman, a professor of architecture who analyzed such perspective as well as the image-forming capabilities of candidate optical devices. The evidence he felt was most compelling was that the image sizes and perspective in six paintings—all executed in the same studio in Vermeer’s mother-in-law Maria Thin’s house in Delft—were consistent with the artist using a camera obscura to project images onto canvases, where the canvases (“screens”) were placed against at the same back wall. He argued that this fact was unlikely to arise by chance or coincidence had Vermeer not used a camera obscura.
Another experimental effort to create “a Vermeer” was documented in Tim’s Vermeer, a film that showed technologist Tim Jenison’s herculean five-year effort to “duplicate” A Lady at the Virginals with a Gentleman (early 1660s). Jenison built or purchased full-sized objects (e.g. virginal, furniture) and fashioned a room with windows matching the dimensions estimated for Vermeer’s tableau. He then spent 130 days painting on canvas using a somewhat complicated optical device and procedure—neither of which documented from Vermeer’s time. Jenison’s proposed method relied on a plane mirror on a pivot that he could rotate back and forth to visually compare the paint he applied to the horizontal canvas against the image of the scene projected onto a wall. The resulting painting, by this admitted non-artist, certainly was geometrically precise, and in many ways resembled a genuine Vermeer. Jenison points to a very slight warp in a line that appears in his painting as well as Vermeer’s original due, he claims, to a slight aberration in the optics.
It seems rather unlikely, however, that Vermeer would have used an undocumented procedure and have expended such an extraordinary effort and endured such crushing tedium to execute his works. Jenison admits on-camera that he would have stopped out of boredom and tedium had he not been making the documentary. Moreover, this method provides no explanation for the key evidence discovered by Gowing that initiated the speculations about the use of optics. Incidentally, Steadman appears throughout Tim’s Vermeer and surely realized that Jenison’s theory is also incompatible with his own explanation; elsewhere and later he says as much. (As we shall see, Jenison’s theory is incompatible with that proposed in the book under review.) Alas, Jenison’s work has never been presented in sufficient detail in a rigorous open forum, such as peer-reviewed journal or conference, that might allow detailed, independent scholarly analysis.
Note that one cannot render colors accurately while painting under a projected image because the color of pigments changes greatly under the different colored lights. For this and other reasons, no serious scholar contends that Vermeer painted the final, visible layer (with its subtle colors) under a projected image. Jenison avoids that limitation by optically switching back and forth so he sees the pigments he is applying and then the separate optically projected image.
An alternate, more modest, proposal is that Vermeer painted directly under optical projections but only in broad areas of light and dark to produce an underdrawing for guiding the design and his subsequent marks. Here the color need not be accurate, just the design.
Enter Jane Jelley, a realist artist working in Oxford who brings her practical experience to bear on her theory about Vermeer’s purported direct use of optics.  First, it should be noted that her writing is elegant and clear, and she accurately summarizes much of the scholarly background on Vermeer, his home life, the Dutch Golden Age, and more. She has deep historical and practical knowledge of pigments, supports, and mixing practices likely used by Vermeer.
The distinctive physical properties of Vermeer’s paintings that led her to her theory are the same that Gowing identified over 60 years earlier: the underdrawings in several of paintings contain broad regions of generally dark pigments that show little or no evidence of modification or development—they seem “fully formed.”
Jelley’s theory is that Vermeer built a simple camera obscura to project a real image—not onto his canvas directly, but instead onto a white board. Then he placed an oiled sheet of paper atop the board and, guided by the projected image, painted an “under drawing” onto the sheet. Finally, he took this sheet, still bearing wet oil pigment, and pressed it against his canvas, transferring the pigment design onto the canvas in a way loosely related to that of a monotype. Such an underdrawing would likely be monochrome and most useful for capturing the design and perspective, not the color. Later, presumably without the use of a projected image, Vermeer would paint the final, visible layer.
Jelley’s procedure overcomes one challenge attending the direct use of optics during painting: preserving the left-right orientation of the final image. An inverted image projected by a traditional camera obscura, when traced and turned right-side up, will be reversed left-to-right compared to the direct view of the scene. Jelley’s pressing of the oiled paper against the canvas adds another reversal, thus ensuring the left-right orientation of the final painting. One could instead introduce a plane mirror or use a concave (converging) mirror to ensure this orientation, but each of those more-complicated techniques has several practical drawbacks.
Jelley’s theory addresses the underpainting but provides no mechanism yielding the visible features that many feel are the hallmarks of the optical look of many of Vermeer’s works, in particular the “blur spots,” such as in Lacemaker (1669–70). These resemble the blur spots arising from an out-of-focus bright highlight. Such “optical” effects would arise in subsequent stages of the development of the painting, not the optical step Jelley promotes.
It is clear, too, that Jelley’s printing method would not work for Vermeer’s large works, such as Art of painting (130 cm x 110 cm, 1665-68), Christ in the House of Martha and Mary (160 cm × 142 cm, 1655), as no lens could produce an undistorted, sufficiently bright image that large; furthermore tiling a patchwork of projections and images would be exceedingly awkward and prone to producing large perspective and image inconsistencies. If we grant that Vermeer did not use optics directly for any of his large works, we are led to ask: would he have used a projector for just his small works?
Jelley’s demonstrations for this early stage are at least credible as they comport with known material culture and optical expertise of the time (but see below). They are not particularly compelling evidence for Vermeer’s use of optics, however, as they address a small portion of the evidence and (crucially) were not integrated into other, more relevant evidence. At root, Jelley seems not to care sufficiently about optics nor perhaps realizes the role of optics in her demonstrations.
She describes the optics of her setup vaguely at best, and incompletely at worst, and much of what she does explain is not sufficiently related to a realistic setup that we can draw strong conclusions. She never gives a clear photograph of the full optical setup, never states all the relevant optical properties such as focal length, aperture, type of lens (convex, plano-convex, etc.). In some experiments she even uses modern computer projector to project a digital image of Vermeer’s final painting onto her oiled paper! Such an approach is inappropriate and misleading for judging a claim about Vermeer’s potential use of optics for these (and other) reasons:
1. Her source image is flat, but Vermeer’s scene was three-dimensional and this leads to numerous optical effects associated with depth of field (certain parts of the image will be in sharp focus, others not), aberrations or imperfections in the image (it might be sharp at the center and blurred and distorted at the periphery, or vice versa), and so on. She acknowledges this weakness or limitation of her argument (p. 201), but does nothing to dispel it, leaving the reader wondering the relevance of her experiments.
2. The lens in a modern optical projector is compound and extremely well designed through modern computer methods to give a high-quality projected image; a lens that would have been available to Vermeer would have had numerous imperfections, aberrations and limitations. It is likely that some of these aberrations would have been apparent in the projected image and affect any tracing. Jelley thus has lost the opportunity to find such image evidence that might have supported her tracing claim.
3. The overall light luminance associated with a computer projector is much higher than would have been available in a projection of Vermeer’s studio by a small lens. I am extremely skeptical of Jelley’s claim that the luminance associated with the projection in her photograph 9.3a (p. 194) corresponds to what Vermeer would have seen.
4. A typical indoor scene will have dynamic range (range of light intensities) as large as 10,000:1 or more, but the dynamic range from a computer projection will typically be far less, roughly 200:1. This reduction in the dynamic range would likely affect what Vermeer would or would not have seen and hence traced.
Most disappointing is that this professional studio artist does not paint atop her traced layer to complete a full painting (much as engineer Jenison had done). In completing a full painting, Jelley would certainly have confronted many challenges that would have had bearing on the optical projection claim, and perhaps identified other sources of evidence that might have had supported or contradicted the optical claim. Perhaps she would have found that the underdrawings provide less help than one might imagine. It is hard to understand why this painter manifestly spent so much time in background reading, traveling to Delft and writing this book but not completing a full painting using the technique she proposes.
There is other important physical evidence in Vermeer’s paintings that is not directly supported by Jelley’s theory, such as pinprick holes at the central vanishing point, as is found on the Art of painting and other works. Such pinprick holes were certainly used in the service of perspective constructions, for instance floor tiles. Jelley acknowledges that her theory does not explain such evidence, but states that her theory is at least compatible with such evidence: Vermeer might have employed her proposed projection/tracing method and then separately placed a pin at the (approximate) center of projection and use traditional perspective constructions. If Vermeer indeed worked this way, we might find a mis-registration between the optics-based printed layer and a layer constructed using traditional geometrical perspective. No one has found such an image discrepancy or mis-registration.
In short, Jelley makes a quite plausible case that Vermeer could have used optics to produce an image, paint over the image, and “print” it onto his canvas as a first step in his method but falls short of making a convincing case that he actually did so. A more scholarly book might have presented counter-arguments to optical claims from other authors. If Jelley wants to convince scholars that Vermeer exploited optics, she will simply have to show a careful understanding of the relevant optics, perform the relevant optical experiments (including completing a full painting), and analyze how such optics makes her tracing/printing theory correct or at least highly plausible and competing proposals unlikely. Until—and perhaps even after such analysis—the precise method behind Vermeer’s masterpieces will remain a sublime mystery.
 Philip Steadman, Vermeer’s Camera: Uncovering the Truth behind the Masterpieces, Oxford University Press (2001).
 Tim’s Vermeer, 1h 20m (A Penn and Teller Film) (2013).
 Jane Jelley, “From Perception to Paint: the Practical Use of the Camera Obscura in the Time of Vermeer,” Art and Perception 1(1-2): 17–47 (2013).