Shedding New Light on a Renaissance Master

By Doug Ramsey,, 858-822-5825

San Diego, February 10, 2009  -- Visitors to the San Diego Museum of Art can catch a glimpse of the museum of the future. There, alongside the "Madonna and Child" by early Renaissance artist Carlo Crivelli, is an LCD display playing a video that takes museum-goers beneath the surface of the painting.

Maurizio Seracini examines a work by Carlo Crivelli
CISA3 director Maurizio Seracini performs an infrared scan of the San Diego Museum of Art's "Madonna and Child," by Carlo Crivelli. [Photo by Laura Andersen/La Jolla Light] To watch a streaming video about the Crivelli painting and the digital clinical chart project, click here. Length: 24:00 [Windows Media and broadband connection required]

Much of the imagery came from multispectral imaging and other diagnostic tools deployed by scientists and engineers at the University of California, San Diego's Center of Interdisciplinary Science for Art, Architecture and Archaeology (CISA3). The video went on display last week to update the public on the findings of a partnership between SDMA and CISA3. Crivelli's 1470 painting is one of half a dozen Renaissance works in the museum's permanent collection that are being studied in minute detail to create prototypes for what CISA3 Director Maurizio Seracini calls "digital clinical charts" for works of art.

CISA3 is based in the California Institute for Telecommunications and Information Technology (Calit2), and is a partnership with the Jacobs School of Engineering and UC San Diego's Division of Arts & Humanities. The video was produced by Calit2 in conjunction with SDMA. "It's a Madonna and child, the most common subject of Renaissance painting," says museum curator John Marciari.

Carlo Crivelli holds a special place in the pantheon of Renaissance artists. He sits outside the mainstream of painting's development in Renaissance Italy. Yet his creative imagination and his abilities as a painter can rank with those of any of his more famous contemporaries. His Madonna and Child is one of the treasures of the permanent collection at the San Diego Museum of Art, and a high-tech project to document the genesis and state of conservation of  the work now shows that there is much, much more to the painting than meets the eye. And according to John Marciari, the museum's curator of Italian and Spanish paintings, the work can help new audiences appreciate Crivelli's eccentric but beautiful style: "He doesn't look like other Renaissance artists, he's a stranger, and this image is typical of his work. You have this rather severe sculptural virgin, and a frankly strange looking child, and this interesting give-and-take between the old mode of painting, and the new naturalism that we see develop in the years leading to 1500.

John Marciari and Maurizio Seracini
SDMA curator of Italian and Spanish paintings, John Marciari, with Crivelli's masterpiece on display at the museum.
Crivelli died five years before the turn of that century. No images of the artist survive, and details of his life are spotty. Born around 1435 in Venice, Crivelli was an apprentice in the Vivarini workshop. But he spent most of his career in a provincial region called Le Marche, The Marches, far from Venice, which he left after spending six months in jail. "He first comes into written history when, in the 1450s, he is arrested, tried, fined and imprisoned for cohabiting with the wife of a Venetian sailor who is out at sea," explains Marciari.

Crivelli left Venice for good in 1459 - first moving to Padua, where he studied the works of Francesco Squarcione and Andrea Mantegna, then to a port town on the Adriatic Sea, today called Zadar (now in Croatia, but at that time part of the Venetian Republic). Ultimately, Crivelli settled in Le Marche, first in the town of Fermo, and later alternating between Fermo and Ascoli Piceno. But he signed his works as "Carlo Crivelli from Venice" until his death.

Digital Clinical Chart
Sample chart depicting the diagnostic and analytical work being done on Crivelli's "Madonna and Child" and several other Renaissance works in the SDMA permanent collection.
"We don't know a lot about him, because he worked in an out-of-the-way place, compared to what we'd have for an artist of comparable quality in Venice or Rome," explains Marciari. "He has no pupils, so he's a bit of a forgotten man."

Forgotten then, but no longer, thanks in part to the Digital Clinical Chart project, launched by UC San Diego's Center of Interdisciplinary Science for Art, Architecture and Archaeology - CISA3 - and its director, Maurizio Seracini. "You need to use technology, especially multispectral imaging, that can capture at different depths the visual understanding not only of the genesis but also problems of decay that are not recognized with just the naked eye," says Seracini.

Seracini is a pioneer of art diagnostics and is taking it to the next level in a partnership between CISA3 and the San Diego Museum of Art. They are prototyping the clinical chart for art works. The idea: to do for paintings what physicians do before diagnosing or treating a patient - take the patient's history, perform all necessary scans, and even do biopsies if necessary. 

Derrick Cartwright
SDMA executive director Derrick Cartwright
According to Derrick Cartwright, executive director of SDMA: "Maurizio loves to use this analogy of the painting as a kind of patient, this object for study, and we've studied objects in a more or less systematic way in the past. But this is a real advance in being thorough and systematic, and really analyzing the patient, in this case a great painting by Crivelli, in such a way that later scholars can go back and see just how healthy the object was at the time and what we've done to increase its life."

"The clinical chart has this purpose: to establish the anatomy, the pathology, and follow-up on the pathology," responds Seracini. "We are in a prototype stage, not only in how we structure it, but in the research and development of the proper technology and methodology to be used in order to create a clinical chart."

"The San Diego Museum of Art is very fortunate in that it has a great collection of Old Master paintings, and the Italian paintings are perhaps its greatest strength. We have work beginning with Giotto from the early Renaissance, all the way to the later Renaissance and singular works by Crivelli and others. The Crivelli, when John arrived here, stood out as a work that could yield a bunch of new information for us all," notes Cartwright.

Javier Rodriguez and Daniel Johnson
Calit2 researchers (l-r) Javier Rodriguez and Daniel Johnson test a new x-y-z scanner they developed to speed the process of capturing high-resolution photography for works of art.
The Crivelli is among a handful of Renaissance paintings at SDMA for which clinical charts are now in the works. Notes curator John Marciari: "We've learned a great deal about the painting and what's happened to it from the moment it was created or during its creation."

The construction of the clinical chart begins by assembling the sort of information that museums have always assembled for their collections: scholarly literature, archival research, provenance records, and so forth. Some of this information is gathered from records and books, and some from the painting itself: a view of the back of the Crivelli panel, for example, reveals old exhibition labels as well as the red wax seal of the collector Oscar Huldschinsky, who owned the painting early in the 20th century.

This information is then juxtaposed with scientific examinations, including very high resolution imaging, ranging from 3D laser modeling, to various types of infrared, ultraviolet, X-ray and other techniques - each of which provides unique insights into different layers of a painting. "Normally the sequence of multispectral acquisitions is to use a wavelength that will help you first establish what's on the surface and then move into the painting all the way to the support," says Seracini.

The Crivelli was first photographed in extreme detail, with cameras and filters deployed on an automated arm. Dozens of images were then stitched together, and the resulting visible-light image shows details that used to require a microscope to see. CISA3 researchers use the world's highest resolution display system, the HIPerSpace wall at the California Institute for Telecommunications and Information Technology, to make sense of the enhanced detail. These brilliant new images help the scientists and the curators study the painting, but they can also be used to demonstrate some of its key points to the public.

John Marciari and Maurizio Seracini
SDMA curator John Marciari (far right) with CISA3 director Maurizio Seracini discuss the Crivelli prior to a LIDAR scan of the work to create a 3D computer model.
"On the one hand, it's a very iconic painting, starting with the gold background, the virgin framed against this cloth of honor behind her," explains Marciari. "Yet you have these modern touches of observed reality, for example in the way the child puts his hand on the virgin's shoulder and leans into her, or the way his thumb grabs her dress. Then, from an artistic view, some very beautiful, naturalistic, almost trompe l'oeil motifs. So the cloth of honor hangs from this rod, it's held on with these laces, and there are these fruits that hang from it. You see these everywhere with Crivelli's paintings, they have both a symbolic and a stylistic function I'd say. The symbolism derives from passages in the Old Testament, specifically in the Psalms, they talk about the fruit of the womb that shall be set upon the throne. This, by the Renaissance, was understood to be a prefiguring of Christ. So Crivelli and a number of other artists associate the Virgin and child with images of fruit. They also derive from classical sculptures, specifically from sarcophagus reliefs. Crivelli studied in Padua when Mantegna was there, and Mantegna makes the great examination of ancient sculpture and begins putting garlands of this type in his works."

SDMA's John Marciari (left) and Maurizio Seracini look at an X-ray of the Crivelli painting.
The painting is on a wooden panel in gold leaf, tempera and oil.  The wood panel was covered with a layer of gesso, and then Crivelli began to draw the composition. Perhaps the most stunning discovery so far in the research on the painting is Crivelli's complete, very detailed underdrawing. Infrared light can pass through the paint layers, allowing the study of a charcoal or ink drawing beneath the paint.  Marciari and Seracini expected to find some underdrawing in Crivelli's painting, for this was the standard practice of 15th century artists, but the completeness of the drawing was stunning. "It's almost an obsessive underdrawing, even though he knew it would be covered by the painted surface," notes SDMA's Marciari. "He begins drawing with dilute ink with a tiny brush, probably in three different passes. He sketches in the basic composition and then uses cross-hatch shading through the entire work, every detail. Only then does he begin painting."

For all the obsessiveness of the underdrawing, Crivelli departs from it in places. Comparing the visible-light image with the infrared, for example, you can see how he first thins, then thickens, the curve of the child's calves, trying to refine that position. The computer-generated overlays of the multispectral scanning process also allows curators, scientists - and now the public - to fade in and out, from one "layer" of the painting to another. Through this process one can easily see, for example, places where Crivelli departed from the carefully drawn pattern on the cloth behind the Madonna. But why did Crivelli do detailed underdrawings that - to Crivelli's knowledge - no one would ever see?

Crivelli infrared prior to image stitching
To produce a very high-resolution infrared scan of Crivelli's "Madonna and Child," CISA3 engineers photographed more than 220 images and stitched them together - then repeated the process at several different wavelengths.
According to Marciari, "it has more to do with Crivelli satisfying himself, rather than with satisfying any necessary artistic practice. That's the most satisfying bit of the creation of the painting we've learned about from the multispectral scans so far."

With the underdrawing complete, the gold leaf would then be applied to the painting. Special punch tools were used to create patterns in the gold. Only then would painting begin. Here too, Crivelli used an elaborate technique, working in multiple layers.

"Then we move into the range of the infrared band," explains CISA3 director Seracini. "Here we operate at 1 micron, especially in the pseudo-color infrared."

Pseudo-color infrared helps with identifying pigments, he adds: "Since most of the pigments and binding media are transparent in this wavelength range, it is possible to gather information from the infrared light sent over the painting and reflected at different depths."

Most of the painting was in tempera, which involves color suspended in protein binding media (such as egg white). "You can see the very tiny brush strokes typical of this technique, with pseudo-color infrared we were able to identify the specific colors used by the artists, for example, the red edge of the robe of the Madonna, which becomes yellow in the pseudo-color IR due to the presence of vermillion."

Crivelli's painting was made around 1470, a time when Venetian painters were beginning to use oil paint. Ultraviolet light produces a very bright, yellow fluorescence on oil paint, so scans in that spectrum make it easy to detect which parts of Crivelli's painting were done in oil, and which were done in tempera. "For as meticulous Crivelli is with little strokes of tempera paint, or his obsessive cross-hatched design underneath the tempera when he gets to oil paint., he paints in what in closeup photography look like sloppy, broad strokes. He was just at this point learning to use oil paint, the consistency of which and properties of which are radically different from tempera paint," explains SDMA's Marciari.

Multispectral scans of the Renaissance painting
Eight high-resolution scans of the Crivelli painting (l-r) in visible light, raking light, ultraviolet fluorescence, reflected UV, pseudo-color infrared, near IR, reflectography, and radiography (X-ray).
In fact, Crivelli painted the entire panel in tempera, then used oil in a few key places, for the highlights in the white veil, the veil's lace fringe that has a seethrough, transparent quality to it, and the trompe l'oeil device that seems to be both part of and separate from the painting, carrying the seal of Crivelli's hitherto unidentified patron, seeming almost lifelike in three dimensions to the naked eye.

Notes Marciari: "You will see in some of the images of the X-rays and infrared photography how differently the oil is handled from the tempera paint."

The ultraviolet scans are also an essential tool in documenting what has happened to the painting since Crivelli painted it. The greenish glow over the surface indicates more than one layer of varnish. "Then we see restorations, some of which look dark, because they have absorbed the UV light. That is usually indicative of recent restorations, and we don't mean only cleaning or overcleaning, but in the case of the Crivelli, I should say re-painting," says the CISA3 director..

Seracini points to thin lines of repainting to fill in some of the major craquelure lines, as well as black spots, indicative of over-painting.

"Then we move to the X-ray, which is much more penetrating obviously," adds Seracini. "It is a projection of the radiopacity of different materials, in this case the panel."

Previous attempts to fill holes in the wood panel with gesso appear clearly as bright, white dots on the X-ray. Explains Marciari: "We know there was some event which caused a major cracking of the paint surface that is visible to the naked eye, but especially in microscopic photographs. This is probably the moment when the panel warped and when it lost its original frame."

Also visible: the extent to which wood worms have eaten into or out of the panel. But the study of wood worm holes can also confirm that the painting was not part of a larger work. "If you trim a panel, you expose trails, the size of trails, whereas the worms only ever go directly in or out. So as long as the holes are nice clean round entry holes, you know you have the original edges," says Marciari.

In time, pigments fade, varnishes darken, and wood panels warp - all of which generally leads conservation work. Today conservation work is done in pigments that can easily be removed, but in the past, much more aggressive means were used in attempts to freshen up old paintings.

"One oddity from the X-ray is a series of incisions carved into the painting's surface, but they seem to have been made not before it was done, as you'd often find, but long after the painting had already cracked. So it may have been by a restorer who may have cut these incisions to give the contours more relief or more depth," explains SDMA curator John Marciari. "We also see three or four earlier campaigns of restoration work on the panel that date, possibly one from the 16th century, and there are at least two or three more in the 19th and 20th centuries - the sort of thing you'd find on any Renaissance painting. Certain pigments were harmed by cleaning solvents, especially the green and the brown shadows of the child's legs that are very much damaged and slightly remade, and the blue of the virgin's robe seems to have been completely repainted at some point, but perhaps early in its history."

At some point long ago, the painting lost its original frame, which was nailed directly onto the surface of the panel, in the now-unpainted margin. The frame moldings were usually also covered in gold leaf, but in this case, the researchers found traces of red paint all around the margin of the painting, and determined that at least the inner edge of the frame was vermillion red. This is unusual, but just another of the many facets of the painting that the recent research has revealed.

Still ahead for the project are a series of diagnostic tests that will hopefully provide answers to some of the questions raised by the multispectral imaging. "We still have not done pigment analysis or cross-section analysis. There are some places were we are curious about the layers of paint, and which are original and which are retouching, and when that retouching may have been done," says Marciari. "If there is any significant time lapse, in a cross-section looked at sideways, you'd see a layer of dirt just from the dust in the studio."

Says SDMA executive director Cartwright: "It's again this marriage of John's expertise with Maurizio's, which brought the Crivelli forward as an object that was ripe for this kind of study, because John knew from his experience looking at other works by Crivelli that there would be very interesting underdrawings that were not necessarily part of past studies of this painting. Indeed, that's what we found out, that that's the case, and there is a lot of material there not visible to the naked eye, which with Maurizio's tools and the insights of the Calit2 scientists were able to bring that out to show it to the public in a very compelling way."

"It will allow us to call up the infrared image, or X-ray and visible spectrum, and to manipulate them to study the painting and to store new images for comparison in the clinical chart. And ultimately this will be linked not only to our own database but to those of other museums that have such charts," notes Marciari. "For example, in the case of the Crivelli, it would be very interesting to know when he starts to use oil paint. And that work just hasn't been done."

Meanwhile, Marciari and the team from UC San Diego are developing clinical charts for several other key works from the San Diego Museum of Art's collection. Like Crivelli, Cosme Tura is known to have made complete underdrawings for his paintings, so Tura's "St. George" was an obvious candidate for multispectral imaging. Also included: a work by Giorgione, the masterpiece of the museum's Renaissance collection. And because Giorgione and Vicenzo Catena shared a studio, the project is developing a chart for Catena's "Holy Family with Saint Anne". "Some works kind of at the margins of Giorgione - some disputed attributions by Giorgione - might turn out to have been made by Catena, and we hope that information gathered here and compared with those in other museums' paintings may help clarify the work of the two artists," says Marciari.

All of the paintings hang in a new permanent exhibit space at SDMA for Renaissance Italian paintings. "San Diego has an astonishingly good collection of Renaissance art and I don't think most people in San Diego know that," says curator Marciari. "So by concentrating it here and having important works at the center of each wall, the Crivelli, the Giotto, the Giorgione on the opposite wall, I am hoping to emphasize the strength of this collection in this area, just as the next room emphasizes our strengths in Spanish painting."

Researchers are also scanning the Crivelli to produce a 3d mathematical model of the painting, and the model will become the basic reference onto which multispectral scans are draped to provide future curators and museum goers with virtual reality versions of the "Madonna and Child". Also ahead for the project, researchers will do point-by-point, non-invasive identification of inorganic materials, using both X-ray fluorescence, and Raman spectroscopy. And they'll use a new imaging technology that has not yet been used on paintings. It's called terahertz time domain spectroscopy:
"It is a new revolutionary technology that will be able to create an image at the interface of each layer of paint to make a spectroscopy analysis to determine which binding media is present. This will give an outstanding, incredible support to the restorer," explains Seracini.

Currently, only the wealthiest museums can afford to own equipment of the sort used by CISA3 to develop digital clinical charts. But as Seracini and his colleagues prototype a standard methodology and portable technologies - like the automated x-y-z scanner already developed at UC San Diego - Seracini says the cost of the equipment will come down: "As in any new science, we have to see that this is the prototype, the early stage. But I truly believe that any museum out there will be able to afford it."

According to Derrick Cartwright, the executive director of SDMA, "The clinical chart, well applied, will be a standard reference for these objects in the future, and if we can develop a useful model that can be exported to other institutions, I think you'll find that a lot of museums will be able to use these techniques to help them in their conservation efforts."

Concludes CISA3's Seracini: "I can see how museums can finally join together in tackling these crucial problems related to conservation and also dissemination of new understanding of these works of art beneficial for the public at large."

Seracini and his colleagues in CISA3 hope their work at the San Diego Museum of Art will lead to a worldwide cyberinfrastructure, accessible via the Internet, to allow museums and researchers to upload and share their clinical charts - one day paving the way for new collaborations, new findings, and a more scientific way of monitoring and safeguarding the health and safety of great works of art.

Media Contacts

Doug Ramsey, 858-822-5825,

Related Links
San Diego Museum of Art