A Picture of a Picture of a Flower: On the Limitation of Working with Digitized Images

Fig. 1. A folio from the Herbarium Danicum sempervivum (1656) at the Royal Danish Library.

“The plant is quite large, occupying most of the space on the folio. The yellow colors on the flowers are saturated and vibrant; and the dark green color on the stem of the biggest flower is shinny under light. The volume of the petals and the fury texture on the stems are created with very fine brushstrokes. Pencil or chalk underdrawing is visible throughout the composition of the plant, but most apparent around the petals.”

This is a soberer version of the description I wrote for a folio (Fig. 1) from the Herbarium Danicum sempervivum (1656) when I was at The Royal Danish Library in Copenhagen a couple weeks ago. The album belonged to the Danish pharmacist Christoffer Heerfort and contains more than 130 botanical watercolors.[1] Within the few minutes I had to spend with this particular folio, I jogged down these most striking features that caught my attention.


Most of them have been digitized!

This has been my major task in the past month, and will continue to be for the next few. I travel to museums and library special collections in the Netherlands and abroad to study and describe watercolors of plants and animals in early modern albums. Most of these manuscripts are considered treasures of the collections at their hosting institutions. To enable the public a chance to look at these gems without having them on long-term display for conservation reason, these albums are often digitized in high resolution and the photographs or scans accessible through the institutions’ online collections.[2]

With all these quality digital pictures, it seems that I can conduct my entire project without all the hassle of packing and flying every other week. Yet, that is never how research works. Aside from the privilege and joy to see these manuscripts in person, having close encounters with the watercolors is important for good research. As wonderful and useful these digitized sources are, they do not deliver some crucial information, especially for a project like mine that focuses heavily on the material and materiality of images.


What cannot be digitally captured?

Unfortunately, the photos I have taken so far during my research are for personal study purpose only. To demonstrate my points, I quickly painted a hellebore imitating some of the characteristics I have observed, and digitized the watercolor with my phone camera and a simple scanner. The technology I used is much less advanced, but the limitations of digitized images remain the same.

Multiple perspectives

Digitized sources are mostly photographed or scanned with a top-down view, allowing one perspective to look at a folio. With watercolors, oftentimes the details are much more visible or can only be observed when a folio is tilted. For example, the reflective quality of the paint in Fig. 2 can only be observed when looking from sideway instead of from the top.

Fig. 2. Top: hellebore watercolor photographed from top; Bottom: the same hellebore watercolor photographed at a 45-degree angle. The circles show that the reflectiveness on the leaves is only visible when looking from sideway.

Similarly, while sometimes digitized books do include a side view of the spine, this visual information is not always included. Even when such visualization is provided, it does not tell the whole story. Many early modern books are like a journal that has lots of inserted ticket stubs, making one side thicker than the other. Although additional pictures of all sides of a book can easily be added, they can potentially make the digital presentation of a source rather messy.

Color accuracy

Another issue with digitized sources is their color accuracy. Every institution might have a different setting for their cameras and/or scanners, and hence the digital colors are not always trustworthy. For example, Fig. 3 shows two picture scanned with different settings from the same home scanner I have. The differences in hues and saturations between the two are apparent, and they change the experience of a viewer even though they are pictures of the same watercolor.

Fig. 3. Two scanner settings resulted in two pictures that have quite different colors of the flower and paper from each other.

Scanning or photographing often flattens and de-saturates the original artworks as well, and thus makes a watercolor look duller in its digital form. The flower books I have studied are almost always much more vibrant and impressive in person because the colors catch light in a different way than in the scans/photographs.

Removal of senses

Perhaps the biggest problem with digitized sources is the loss of engagement with several senses. On a computer screen, a big book can appear small, and a miniature can seem like a large poster. I can give the measurement of the watercolor block with the hellebore (23 x 31 cm; 9 x 12 in). With the object in my hand (Fig. 4), however, I can assess the size and weight of the block—as well as the scale of the hellebore to the block—in a much more personal manner, using myself as a measuring device.

Fig. 4. Holding the watercolor block helped me learn that it is rather light in weight and easy to handle because of its relatively small size, and that he hellebore is around the size of my palm.

Additionally, when going though an album, sight is not the only sense that is hard at work to look at the watercolors. My fingers feel the thickness and smoothness of the parchment or paper as I flip through the folios; my nose smell the scent of the materials; my ears hear the (sometimes a terrifying cracking) sound of the turning pages. These makes the viewing experience a very different one from what a digitized album can provide.


Return to the physical objects

Even with all the above-mentioned limitations, digitized sources are still highly useful for my research. They help me know what to expect before seeing an object and better pace myself through my appointments. Combined with the more detailed pictures and descriptions I generate when studying an album, digitized sources enable me to more effectively recall my observations and discoveries.

I have been quite fortunate to get permissions to study many valuable albums in person. Their hosting institutions also recognize the importance of working with the originals for my project, despite having digitized versions in high quality available. So, in the next few months, I continue to visit museums and libraries to study more early modern watercolors in person and write more descriptions, such as the one at the beginning of this blog post. How productive are these hundreds to thousands of descriptions documenting my first impressions of the watercolors? Well, that is a story for the next time.


[1] For more information of the album or see the digitized version, visit http://www.kb.dk/manus/vmanus/2011/dec/ha/object84817/en.

[2] The usefulness of digitizing museum and library collections and how to best do so has been an ongoing debate. See, for example, a recent article by Andrew Prescott and Lorna Hughes that discusses this issue in great extent and provides much literature on the topic, “Why Do We Digitize? The Case for Slow Digitization,” Archive Journal, September 2018, http://www.archivejournal.net/essays/why-do-we-digitize-the-case-for-slow-digitization/.

6 thoughts on “A Picture of a Picture of a Flower: On the Limitation of Working with Digitized Images”

  1. This article makes several fascinating points about the limits of digitization that I had not consciously considered. Of course you’re always aware that seeing it in person is different from seeing it in real life, but you did a wonderful job of illustrating those differences clearly. Having seen the drawing of the hellebore, I would never have guessed it was quite that large! Makes me wonder what other things I’ve been misinterpreting….
    (Also lol’ing at the sometimes terrifying cracking sound of turning pages XD).

    1. I find it particularly interesting that we didn’t give this too much thought when we were taking illustration classes, especially when we had to deal with digital reproduction so much. But now I really wonder why the material qualities change so drastically when digitizing early modern watercolors. There are still slight discrepancies; but with my own watercolor works done in modern paint, I have been mostly able to very closely capture it in digital form. I guess this is the next to investigate!

  2. I have the same experiencr with Balinese Palmleaf manuscripts. When reading the real manuscripts one has to hold them in different ways so that the inscribed letters become much easier to read. Also the feel of the leaves is necessary to decide what they are and where they come ftom.

    1. Precisely so! I would argue the same for handling printed books as well. Medieval and early modern manuscripts and books seem to have a subtle three-dimensional quality that play a big part in the reading experience.

  3. Jessie, I enjoyed reading this blog; Your observations are very interesting. Size is also an important consideration when making work digitally or physically. Digital work looses its scale in relation to the original file potential.

Leave a Reply