Walkthrough of Spectral RTI in Mirador Illustrating Hans Christian Andersen Cutouts

Transcript

This is Todd Hanneken, Director of the Jubilees Palimpsest Project. This screen capture shows the capabilities of Spectral Reflectance Transformation Imaging and the International Image Interoperability Framework using Mirador as the primary viewer. Rather than our usual ancient erased manuscripts, we will be looking at cutouts by Hans Christian Andersen, the nineteenth century Danish author famous for children’s stories and movies based on them. Spectral RTI is not just about recovering lost writing. Accurate color and visualization of texture can be valuable for appreciation of artistic beauty.

If we begin from the main page of the Jubilees Palimpsest Project, the first link will bring us to the Mirador viewer. From here we can scroll or filter to find the Hans Christian Andersen cutouts. This collection, shelfmark Laage Petersen 659 at the Royal Library of Denmark, consists of six cutouts and three pages of written word. We can select a page by clicking on a thumbnail or an item in the index on the left. Once we are in page view we can move between pages using the index, the arrows on the side, or the tray at the bottom, which can be minimized. Hans Christian Andersen made these cutouts for a girl named Marie, who died when she was four-and-a-half years old.

Mirador has many features that are not directly related to Spectral RTI. It is easy to pan and zoom. We can open up additional slots if we wish to compare images side by side. For example, if we wished to study consistencies and inconsistencies in how Hans Christian Andersen cut flowers across projects, we could open two flower-based images next to each other. We can see thematic similarities, but also some very different ways of making the individual flowers.

This implementation of Mirador has an annotation tab. All pages list the WebRTI images, to which we shall return. The pages with text offer transcription and translation. The annotations can also be edited, so if we found an error in the translation we could click on the annotation icon. If we hover over a blue box we get the annotation and an option to edit it. We could also add new annotations in various shapes and colors.

Near the annotations icon is the image manipulation icon. For example, some cutouts are ambiguous as to which way is up. We could rotate the page accordingly. The other image adjustment controls may be helpful, but should not be mistaken for the powerful processing possible with multispectral data.

The layers tab shows additional images of the same page. We can turn on raking light from any of four directions. These show the condition of the paper, which is much more warped than is obvious in diffuse light. If we zoom in we can see the texture of the layers of paper and even the texture of the paper itself.

If these raking angles are not enough, or if we want more interactivity to change the lighting, we can go to the WebRTI images in the Annotations tab. Clicking on one of these will open a new browser tab. Now if we click or tap the light-bulb icon on the left so it turns yellow, our mouse or finger on a touchscreen will move where the light appears to be coming from. From the way the highlights and shadows change as the light position changes we can perceive the texture and come much closer to simulating first-hand experience.

The interactive relightable images are what puts the RTI in Spectral RTI. The other part is the multispectral imaging, which gives us options for Accurate Color and Extended Spectrum (possibly others) in the RTIs and layers. Many people think that multispectral imaging is only for illegible writing or other indecipherable features. As far as I know there is nothing indecipherable in the cutouts, but we can demonstrate the utility of spectral imaging in two ways.

First, a conservator might be interested in traces of materials on the cutouts. Extended Spectrum is a simple multispectral visualization that squeezes invisible light in ultraviolet and infrared into the range we can see, violet and red respectively. This visualization shows spots that are much harder to see in the accurate color, although once they are pointed out we can see the slight browning in accurate color.

Second, and more importantly, spectral imaging allows us to achieve a much higher degree of color accuracy than we get from conventional digitization. Accuracy is important for conservation because the color can be compared accurately across time and imaging systems. It is also important for appreciation of art to see the colors as they really are, not worse or artificially better. Conventional digitization captures three colors at once, while spectral imaging uses several captures to measure color across the visible spectrum. Conventional photography compensates for certain intrinsic shortcomings by enhancing the picture to look prettier than the captured data, but artificial enhancements are inaccurate even if they are pretty. If you have a screen with good color quality, we can see the difference between conventional and multispectral digitization. First we go to the information icon in the upper right of Mirador. From here we find the related links, including the conventional digitization by the Royal Library of Denmark. This is a very nice image as far as conventional digitization goes. But if we compare it side by side with the multispectral accurate color, it is clear that the conventional digitization has more saturated and less accurate color. Again, it doesn’t look bad. The problem is it does not look the way the cutout looks in real life.

For more information about Spectral RTI and the Jubilees Palimpsest Project visit jubilees.stmarytx.edu and follow me on twitter, @thanneken.