Abstract –This paper shares the experience and learning curve of digitizing the Panorama of Congo, a significant yet rarely exhibited colonial artwork. The immense scale of the painting (115 meters x 14 meters, 1610 m²) made by Belgian painters and commissioned by the Belgian Ministry of Colonies posed both physical and ethical challenges for digital preservation. With limited best practices available, a multidisciplinary team undertook the task of photographing the canvas in December 2022. The method involved volunteers, specialized equipment, and Metamorfoze guidelines to ensure the creation of a digital version for further research and virtual reality applications. Despite the ethical complexities of digitizing colonial heritage, the project aimed to respect the dignity of all peoples and contribute to a better understanding of history. The digital artifact now exists in three forms: individual high-resolution images, an unretouched gigapixel file revealing the canvas’s current state, and a digitally restored file that approximates the original viewing experience. The latter is also integrated in a VR application where it can be seen in 360°.

Keywords – Panorama, Photography, VR, Colonial Heritage, Metamorfoze

This paper was submitted for the iPRES2024 conference on March 17, 2024 and reviewed by Bieke Nouws, William Schlaack and 2 anonymous reviewers. The paper was accepted with reviewer suggestions on May 6, 2024 by co-chairs Heather Moulaison-Sandy (University of Missouri), Jean-Yves Le Meur (CERN) and Julie M. Birkholz (Ghent University & KBR) on behalf of the iPRES2024 Program Committee.

Introduction

On a cold winter day in 2022, a group of Belgian and Portuguese researchers from various fields visited a military storage facility in Ypres, in the Belgian countryside. They had traveled there to see the Panorama of Congo, a massive panoramic painting measuring 115 by 14 meters, created for the World Exhibition in Ghent (Belgium) in 1913. Among enthusiasts and scholars of panoramas, the painting had reached an almost mythical status. Although many had heard of it, few had actually seen it. When the group viewed it in January 2022, it had been stored for nearly 90 years since the last exhibition in Brussels in 1935. They found it rolled up on a large wooden spool, slightly dusty and wrapped in foil. Even without being able to see the painting itself (since the paint was on the inside of the roll and it could not be unrolled without extensive preparations), the object still captured their imagination.

This visit initiated a lengthy process and research project on the Panorama of Congo. The study object was problematic in several respects. Firstly, handling it was an obvious challenge, secondly, the image it concealed was a highly sensitive subject from a modern perspective: Belgian Congo. Moreover, it was created by two Belgian painters under the commission of the Belgian Ministry of Colonies to promote the newly acquired colonial territory of Belgium, making it undeniably propagandistic.

Prior to the digitization, the available historical visual material was very limited, consisting mostly of ephemera from the two exhibitions (1913 in Ghent and 1935 in Brussels, both in Belgium) and a number of digital photos from a very limited unrolling in 2009. A complete view of the panorama was not available. To get an idea of the painting, the research team stitched together a number of historical postcards with faded colours. These did not reveal any details or a real vision of the totality. It also became clear that parts were missing in the ephemera. Therefore, digitization of the canvas was imperative. This confronted the research team with a multitude of questions, not only in relation to the size of the object but also to the charged nature of the depicted subject. This article will focus on the digitization of the Panorama of Congo, discussing the methodology, choices, and solutions, and will also touch briefly on the ethical implications (reinstallation and post production) and considerations.

The Photographic Mission

Imagine, a huge painting was carelessly rolled up on a wooden spool 90 years ago after an exhibition, with no plans to show it again in the foreseeable future, weighing about 4 tons. The visible edges of the canvas show its fragility and crumbling, the subject of the artwork is very sensitive, and it’s stored in an un-climate controlled warehouse too small to even partially unroll it. A photography team specialized in reproduction of artworks might initially think to leave this "skeleton in the closet." However, 60 years after Congo’s independence (1960) and amidst societal debates on the lasting effects of colonization, re-visualizing the Panorama of Congo is undeniably meaningful in many respects1. Bring in the Bergepanzer tanks!

Unrolling and Building the ‘bridge’

The photography team’s first challenge, in collaboration with the War Heritage Institute2, was to move the canvas to a space where photographing and especially unrolling could take place. They opted for a warehouse with a workable space of 15.5 meters wide and about 60 meters long, recalling that the canvas itself is 14 meters wide, with spool almost 15. Two Bergepanzer tanks with cranes would position the canvas for it to be unrolled in various stages.

Since the canvas was loosely rolled around the wooden spool, it appeared as though the spool had fractures and was bending at places, which could hinder not just transportation but also the unrolling process. The preliminary research thus primarily faced two major challenges: firstly, finding an efficient and safe way to unroll the canvas, and secondly, constructing a structure that would allow for consistent lighting and photographing of the canvas. These two challenges intertwined and needed simultaneous answers. One (the lighting and photographing process) was dependent of the other (how and where will the canvas be unrolled, what length and what area will remain available for the photographic structures).

The canvas had to be laid out in a military warehouse with limited available space to be photographed from above. After considering several mechanical and automated systems, it was decided to manually unroll it in three phases, despite the estimated total weight (spool and canvas) of 4000kg. Each phase would need to make approximately 40 meters (length) by 14 meters (width) of the canvas visible and photographable.

Due to this a movable structure had to be devised that could scan both the width and length of the canvas. This meant that both the lights and the camera had to be mobile, and the entire 'bridge structure' had to be movable. In collaboration with the Ghent company City Sounds Rent (CSR), a movable truss structure was built. The 'bridge structure' consisted - simplified - of two movable truss columns on which a horizontal truss was mounted. A mobile box was attached to this horizontal truss, which could be manually moved from one end of the bridge to the other with ropes. The lights and camera were mounted on this 'movable' box, with the camera tethered to a laptop and the flashes controlled by infrared. Despite the large span of slightly over 14 meters, the deflection was a maximum of 2cm, which fell within the targeted limits to keep the distance between the camera, light, and canvas consistent in all shots (approx. 3 meters).

The Metamorfoze guidelines

The digitization of the Panorama of Congo was aimed at two main goals: firstly, to create a digital version (preservation master) that would open up the canvas for further research, and secondly, to produce a derived version suitable for use in the virtual reality section of the research (artistic research). Therefore, the photographs were made having the Metamorfoze guidelines (later abbreviated as Mg) as a quality standard orientation, but with an understanding that, due to the singularity of the task in our hands, some adjustments might be necessary in order to complete the reproduction of the panorama.

The Metamorfoze guidelines were developed in the Netherlands by the Koninklijke Bibliotheek and the Nationaal Archief for digital preservation and dissemination, especially for fragile works on paper. The guidelines work with three quality levels with their respective technical criteria and tolerances (cf. colour accuracy, sharpness, sampling rate, gain modulation, etc.). The highest level, Metamorfoze, is used for digitizing original artworks; Metamorfoze Light is for materials where color accuracy is less critical, and Metamorfoze Extra Light is for digitizing books and periodicals. The aim is to create a preservation master that captures all visible information from the original, ensuring complete data transfer.

Sampling Rate and Number of Shots

The Metamorfoze guidelines are very resolute when it comes to the sampling rate: if the work is larger than A2 (420mm x 594mm), it should be at least 150ppi. However, the Panorama of Congo has a surface area of 1610m2, meaning it is 6453 times larger than an A2. It is self-evident that tracking the Mg in this order of magnitude would result in an image size that passes a smooth digital spread. To still stay as close to the guideline as possible, it was therefore opted to work in plots of 240cm x 182cm with the Fuji GFX 100s (11648px x 8736px). An overlap of 25% each was also taken into account to allow for subsequent stitching (about 50cm vertically and 52cm horizontally). In this way, each individual image of the panorama has a resolution of about 125 ppi (only 17% smaller than the recommendation) while each individual image is still 1750% larger than an A2. This sampling rate can be measured with resolution test charts (QA-62-SFR-P-RP). Nine images (9 rows) of 240cm (x 182cm) were taken every 182cm (total 93 columns), bringing the total to over 830 shots of 102mp for the whole canvas.

From our calculations this would allow us to produce a final file that could be printed in the original size (1:1 scale – 14 x 115 mt) without loss in quality, preserving the possibility of a reinstallation of the Panorama of Congo in its entire monumental scale without the need to stress the original canvas.

Exposure and Colour

The importance of consistent lighting cannot be stressed enough when several images have to be merged into one, both for correct light distribution and colour reproduction. In the case of the Panorama of Congo, more than 830 shots were involved, which - once again - required extreme meticulousness to make the subsequent stitching possible in an efficient and economical way.

When unrolling the canvas, however, it quickly became clear that manually laying a canvas of this size and weight perfectly flat was an impossible task. Moreover, it turned out that the canvas itself - perhaps caused during the roll-up - showed irreparable deep creases. These bumps, some up to 20 cm high, made it impossible to photograph the canvas with a crossed (45°-35°), polarised exposure, which is common in 'normal light' reproduction of artworks. Working with this exposure would create multiple shadows and also seriously complicate subsequent stitching as a result. For this reason, it was decided to work with two 1 meter by 1 meter square softboxes facing downwards with only a slight inclination of up to 5° to generate a balanced spread. The flashes and softboxes were mounted 75cm from the camera to counteract possible lens flare. This setup provided a uniform exposure on which a Lens Cast Calibration (LCC) profile was created that would allow additional controls and adjustments. To also optimise aperture and dynamic range, it was chosen to work with two 1000Ws flashes. To avoid further deflection of the horizontal truss, we opted to work with the compact and relatively light (2.35kg) Elinchrom Compact ELC Pro HD 1000 monoblocks.

In any task of photographic reproduction, creating color correction profiles is essential to obtain an image that establishes a relationship as close as possible to the material reality of the object and, above all, is measurable. For the panorama, a photo was taken at the beginning of each column to measure exposure and tonal response, which led to the creation of color profiles via the reproduction of the color target ColorChecker Digital SG (the ColorChecker Classic 24 was also used to maintain constant redundancy throughout the process).

This quality control on each of the 93 columns eventually allowed for the selection of a profile that was applied to all images so that the next task, merging all images, could be performed while maintaining the Mg for the entire panorama. After several experiments, the color profile was chosen that allowed a more balanced experience across all images of the panorama, at the same time promoting a color experience very close to the original in each of the images and allowed the merging of all images so that we can visualize the entire panorama in a single digital image. For the creation of the color profile for all the files we used the Basic Color Input 6 Pro software, which was letter applied to all the files in the interpretation process via Capture One pro 22 Software.

Stitching and the Individual Files

After the colour profile was applied to all images, the resolution had to be reduced to allow the merging of all images into one file of about 1400x11500 cm. This reduction in resolution still provided remarkable quality in the final file (about 50 ppi, a 66% reduction in Metamorfoze recommendations). The PTGui software was used to paste all the images. This was a learning process. The PTGui is a quite complex software that is parameterized in its origin to create panoramas that have been shoot with a a camera that rotates in the nodal point of the lens, for the panorama we had a camera that was moving in two axis (x and y) “scanning” a flat surface (instead of picturing a landscape). In order to use to software to produce a stitched file from all our images we had to introduce specific parameters in focal depth or alternate the orientation of the “stitching” path (from horizontal to vertical). We also found out that PTGui worked better when using files that had gone through a light JPEG compression (compression 12 in Photoshop) instead of the full TIFF images. This was due to the well known creation of the color clustering that JPEG algorithm applies to the RGB pixels. Due to the weight and quantity of the images, it was necessary to make this stitching in vertical portions of several individual columns (remember that there are 93 columns divided by 9 rows each). The stitching software makes minor perspective and colour adjustments during the process. These adjustments are necessary and would be made regardless of whether the process was manual or automatic, but still introduce small changes to the original proportions. This difference is invisible in the final file and is only noticeable when comparing an individual image with the final file. It should also be noted that the resolution used in creating the final file allows the panorama to be printed at full size for future installation. Like we referred previously, we can print a Panorama of Congo on canvas and reinstall it without disturbing the original.

Which basically achieves the Mg with respect to the purpose of the preservation master. A second argument to support this way of working is that the canvas was painted with the aim of being viewed from a distance of more than 10m from the visitors’ platform3.

The Post Production

After the final version of the stitching of 830 images has been completed, we can observe an image that allows us to assess the material conditions of the panorama. This first image speaks of the painting as an object, not hiding its marks, cuts, dirt, wrinkles, folds, holes, etc. allowing observation of the life of an object. It is this materiality that opens the doors to a drift of history through the panorama, the holes made by a soldier’s knife during the war, the wrinkles of its age, the fading of the pigments etc. From this file we then carried out the final stage of our project, the digital retouching/restoration of the panorama, trying to promote a visual experience closer to the state of the panorama in the first half of the 20th century. This work is carried out in a similar way to the material restoration of a painting, taking samples of color and texture from places close to the one to be restored and applying it there. Without the burden of manipulating the original object, we can carry out deeper actions and on a larger scale without worrying about altering the original. So in the end we have three instances of reproduction of the panorama:

1. Individual images of the panorama at a resolution of 125 dpi, each covering a 240 x 182 cm portion of the panorama, allowing a detailed view of the painting and its state.

2. The entire panorama in one file of approximately 7 gigapixels without restoration, allowing to assess the current state of the panorama in a single image.

3. The entire panorama in a digitally restored file that allows a visual experience of the panorama close to the original experience, but far from its current material reality.

Ethical Considerations

In an era in which both scholars and laypeople alike are making considerations and asking fundamental questions about removing colonial heritage from the public realm, the reinstallation of a propaganda object like the Panorama of Congo is an extremely delicate matter. It raises a number of ethical questions revolving around the interpretation of history, the legacy of colonialism and the impact on contemporary societies. We are aware of the fact that this is beyond the scope of this article, but we still want to at least address the questions that relate to it. The first question to ask is that of 'representation': can the reinstallation of such works reaffirm or even reinforce outdated, biased or offensive representations of colonised peoples and territories? And if so, how can these works be displayed in a way that does not endorse the colonial ideologies they originally sought to promote? In other words: what is the appropriate context for displaying colonial propaganda? How can we provide sufficient educational context that acknowledges the historical harm and misrepresentation, and stimulates informed discourse?

A possible answer to these questions should also certainly not ignore the question around the 'who'4. Who should have the right to own, display, or make decisions regarding colonial artifacts and art? What role do the descendants of the colonized communities play in this?

How does the reinstallation of such works affect collective memory and the healing process for societies that have been harmed by colonial practices?

When the canvas was unrolled, all researchers were greatly impressed by the beauty and the extreme artistry that the painting displayed, so how can we separate the artistic value of such works from their propagandistic content? Is it possible, or even desirable, to appreciate the artistry while condemning the message?

Our last question in this non-exhaustive series is the following: Is not displaying such works a form of censorship or erasure of history?

All these questions are complex and have no easy answers. Addressing them requires a multidisciplinary approach, involving historians, ethicists, community leaders and representatives of the regions affected by the content of the propaganda. The ultimate goal is to deal with such artefacts in a way that respects the dignity of all peoples and contributes to a better understanding of history. What we are absolutely convinced of, however, is that offering rigid answers in this ignores our approach in which we want to develop new narratives in an open and dynamic way.

Conclusion

Digitising a work of the scale of the Panorama of Congo with limited resources and a limited vision of its current state presented the group of photographers with a very large number of challenges. These challenges ranged from questions of handling, exposure, sampling rate, colour reproduction, stitching, post-production to a number of ethical concerns. Nevertheless, these led to an efficient way of evaluating, adapting and developing a number of best practices. For example, the Metamorfoze guidelines, which were initially developed for the digitisation of artefacts no larger than A2, required re-evaluation and adapted application due to the extreme scaling up (by more than 6000 times). For the digital colour reproduction, it was opted to make it truthful (to the current state of the canvas) and measurable by creating a large number of colour profiles. In turn, working with the well-known software PTGui required some adjustments that made the stitching process easier, but this with more manual processing. The digitisation resulted in three parts: 830 separate high-resolution images, a true-to-life stitched image and a stitched image to which summary post-production was applied.

ACKNOWLEDGMENTS

This research was developed in the context of the FilmEU – European Universities Alliance for Film and Media Arts work and supported in part by funding from the FILMEU_RIT – Research | Innovation | Transformation project, European Union GRANT_NUMBER: H2020-IBA-SwafS-Support-2-2020, Ref: 101035820 and the FILMEU – The European University for Film and Media Arts project, European Union GRANT_NUMBER: 101004047, EPP-EUR-UNIV-2020.

1. Website: https://congopanorama.filmeu.eu/

2. Head researchers: Victor Flores and Leen Engelen

3. The photography team: António Alberto Coelho, Lennert Deprettere, Rodrigo Peixoto, Tomas Vandecasteele, assisted by Oleksander Lyashchenko and José Fadolla