Creating the composite drawing of the Bark Sanctuary’s western outer wall in the Small Amun Temple at Medinet Habu (Part 3)

Sep 15, 2018

Inking and Post-Processing

Thutmoside lintel reconstruction (detail)


The digital inking process started with upscaling the composite pencil drawing from 400 dpi to 1200 dpi, which resulted in exponential growth in the file size. To keep the drawing manageable while preserving all the layers, the 8-bit RGB file was transformed into 8-bit Grayscale (Image/Mode/Grayscale). Discarding the color information was now possible, since all the different paint episodes were already represented on separate layers. After changing the drawing to grayscale, the new file size became half of the original, about 700 MB. Nonetheless, the 1200 dpi upscaling (Image/Image size/Resolution) resulted a staggering 3 GB of data. To be able to work with such a large-scale drawing, the composite version had to be divided into and cropped to smaller units. The left (MHB 163) and right (MHB 165) wall segments could be inked separately, with the Ptolemaic lintel forming the connection between the two. The door jambs were attached to the wall scenes to facilitate ease of joining between the side walls. The importance of starting with the largest possible documentation area (ideally an entire wall) cannot be emphasized enough. It was now less problematic inking these much smaller units separately, knowing that there would be no adjustments for distortion or perspective correction necessary when rejoining them at the end. Replacing the low-resolution versions of the inked drawings (MHB 163 and MHB 165) with their high resolution counterparts was also relatively easy and straightforward, as it was only a matter of importing the 1200 dpi variants and moving them over their upscaled placeholders.

Switching to 1200 dpi

400 dpi drawing upscaled to 1200 dpi (MHB 163 detail) indicated by blurry lines (left), original drawing scanned at 1200 dpi (MHB 163 detail) with digital scalpel added (middle) and digitally inked 1200 dpi drawing (MHB 165 detail) that has the sharpest edges

(1) The 400 dpi pencil composite drawing’s inked elements upscaled to 1200 dpi (detail).

(2) The original digitally inked version (1200 dpi) imported into the upscaled composite drawing.

(3) Changing the upscaled version’s opacity to be able to match the two layers.

(4) The newly placed 1200 dpi ink layer after the deletion of the upscaled original (detail).

At this point, the facsimile of the left-hand scene (MHB 163) was still a static 1200 dpi scan of the original drawing, which had been inked using Rapidograph pens on traditional photo enlargement. The drawing had to be integrated into the layer system to merge seamlessly with the digitally penciled and inked sections. First, the scanned 1200 dpi version was modified to produce a more homogeneous white background with deeper black colors by adjusting the level sliders on the scan (Image/Adjustments/Levels). Care must be taken when making Levels adjustments in order to avoid data loss, paying attention to the fine details by frequently zooming in and out while adjusting the sliders. Then, in order to change the static scan into a layered Photoshop drawing, it was necessary to duplicate the drawing multiple times in order to produce the desired number of layers (Layer/Duplicate), and to erase the irrelevant portions of these layers accordingly. For example, the duplicated layer with the damage content should not have anything else on it but damage information, etc. This procedure was tedious, but necessary. As soon as MHB 163 was transformed into a multilayered digital drawing, the composite of MHB 163-165 was ready to be finalized.

Digital Inking

A lengthy description of different inking methods and tools to be used has been given in Chapter 5, so only a few peculiarities should be pointed out here. When inking over upscaled (400 dpi to 1200 dpi in this particular case) pencil drawings, one must accept a certain degree of blur when looking at the drawing at full scale. When switching over from inking on a 1200 dpi scan, the artist has to focus on keeping the inked line in the centre of the pencil line and avoid being distracted by the pencil line quality. Nevertheless, the better the in situ pencil line quality, the easier and more precise the inking becomes.

Digital Inking

(1) Line drawing

(2) Line drawing + Trace

(3) Line drawing + Trace + Paint

(4) Line drawing + Trace + Paint + Plaster

(5) Line drawing + Trace + Paint + Plaster + Damage

A special problem occurred while digitally re-inking MHB 163. The newly inked parts had to be blended together with the parts that were originally inked on paper using the traditional Chicago House method. Luckily, MHB 163 had not yet been collated, so the damage texture had not yet been broken by scalpeling, a process that usually happens at the very last phase of inking a drawing. Moreover, aside from the upper and lower sections, certain elements had to be digitally added to the left side of the wall group as well, including the vertical torus, which was not originally planned for inclusion in the drawing. To recreate the exact same damage texture in digital form, a special damage pattern had to be created. Larger damage sections in MHB 163 were cloned to a blank canvas, and this became the base element of the special damage pattern, while further damage strokes were added to create a homogeneous rectangular area. When large enough to hold a variation of darker and lighter areas, this pattern was applied over certain random parts of the damaged areas of the torus.

Inking over Pencil Lines

(1) Upscaled pencil lines are a little blurry when zoomed in at 100% but the overall quality is still good enough for precise inking.

(2) Inking over digital penciling is a straightforward affair and changing layer opacity helps a great deal when dealing with small details.

(3) The finished inked line drawing should have the exact same line quality as the already existing parts.

(4) Adding damage from the upscaled B/W photo background shouldn’t be a problem because damage is applied at a 50% scale.

Creating Damage Pattern

(1) Larger, homogeneous portions of existing damaged areas were cut from the original inked drawing.

(2) A custom A4 size new canvas was set up with the regular settings applied (1200 dpi, grayscale, 8 bit).

(3) The damage portions from the original drawing were imported onto the blank canvas.

(4) These damaged spots formed a random pattern of separate islands on the new canvas.

(5) The blank gaps between the islands were filled in by adding more of the same pattern by freehand.

(6) Finally, this process was repeated over and over until the new damage area became satisfyingly large and randomized.

Damage Types

(1) Large crack moving through the cavetto cornice on the left side of the doorway.

(2) Cracks are represented with elongated damage pattern indicating the orientation of the brakes.

(3) Deep hole cut into the unfinished sun disc panel above the doorway left by treasure hunters.

(4) Extreme depth has to be represented by applying intense amount of damage pattern to reach the dramatic effect.

(5) The door socket underneath the lintel became visible thanks to the missing corner but was masked again by modern restoration.

(6) By adding definite edges to the damage pattern, deep features cut into the surface become instantly recognizable.

(7) Chisel marks showings signs of intentional hacking on one of the reused blocks that used to be part of the cavetto cornice.

(8) Chisel marks have a very peculiar pattern of rhythmically changing darker and lighter areas, that has to be carefully captured.

The actual photo background was underneath the damage layer providing the real texture, therefore further adjustments could be applied by hand wherever necessary. The end result was a satisfying blend of old and new damage texture, very similar in style thanks to this careful method. The border between new and old damage was blurred even further when scalpeling was applied all over the damaged surface forming its final appearance. Furthermore, certain types of damage patterns had to be created for representing certain elements: a crack, a hole, a cut-out or a chisel mark had to be distinguished from each other when appearing on the drawing.

Blending Old with New

(1) Scanned multilayered MHB 163 indicating torus outlines by pencil (only for reference).

(2) Torus added digitally with damage indicating where it’s broken off.

(3) Digital damage blended together seamlessly with traditionally inked portions.

(4) Applying digital scalpel marks over the entire damage surface unifying old and new.

The other significant inking challenge that needs to be specifically mentioned here is representing the paint layers. Thanks to the meticulous paint study, the Thutmoside and Ptolemaic phases were already separated out onto different layers, but using a different color for differentiating between them wasn’t an option anymore because the file was in grayscale. It would have been obvious to use a different kind of outlining for one of the layers but inventing special signage for one separate matter would have caused consistency issues with the rest of the documentation material. Furthermore, there is the possibility to publish many different versions of this wall segment where only one of the painted layers are present at a time, therefore both painted layers needed to have a version containing only a single row of dotted lines.

Representing the two separate paint layers

To solve this issue, a non-destructive modification was applied on one of the painted layers, resulting in the dots appearing somewhat fatter and not too intrusive at the same time. To get the desired effect the layer’s style had to be changed by right clicking over the layer on the Layers panel and selecting Blending Options. This called up the Styles panel, where the Stroke style had to be selected. On the Stroke dialog box, while the Fill Type was left as Color, the actual color was set to black, that being the only color used for inking. In the next step, the Stroke Structure’s Size was changed to 4 pixels and was positioned on the Outside, which meant that an extra 4 pixels were added around every dot on the layer. Blend Mode was left to Normal, which made the dots look much larger and overly emphasized. To reduce the effect, Opacity was toned back to 40% causing these new dots to appear transparent and the overall shade look more grayish than black. By reflecting to the actual situation, this nondestructive modification was applied on the Ptolemaic painted layer, that being the topmost and better preserved of the two. Furthermore, the Ptolemaic features were originally painted with thicker, therefore much more prominent outlines anyway.

Tweaking the Layer Style by modifying the Stroke Styles settings in Photoshop. Changing the values effects the entire layer and can be turned on and off by clicking on the eye icon on the Layers panel.

Assembling the Composite Drawing

All three inked drawings (MHB 163, 164 and 165) were based on the same composite pencil drawing, therefore their proportion and features were an exact match. Nonetheless reassembling the three drawings to a unit had some challenges, because once the three files were merged it was impossible to interact with this significant amount of data due to the large file size. There was a simple solution to avoid too much correction when joining the three drawings.

Assembling the Composite Drawing

(1) When inking MHB 164, the connecting edge of MHB 163 was copied over as a help layer to provide the exact layout information.

(2) When inking was finished, MHB 164 (in black) was moved to the enlarged MHB 163 canvas with the help layer (in grey) included.

(3) When the help layer was moved over its counterpart’s exact position on the inked MH-B 163 drawing, it could be deleted.

(4) At the end MHB 163 and MHB 165 joined perfectly thanks to the help layer, and there was no need for further adjustments.

(5) As a final step, the relevant layers on both drawings needed to be merged with each other to keep the layer system simpler.

When the inking of MHB 163 had been finished and studio work on MHB 164 had started, an approximately 10 cm area of the right edge of the merged MHB 163 drawing was copied over to the MHB 164 canvas as a guide for the exact layout. Technically the doorjambs belonged to MHB 163, but they were inked as part of the left wall panel’s canvas to minimize the joining surface on the two drawings. Luckily the large architectural elements, such as the lintel, cavetto cornice and torus moulding provided a natural divider between the two separate drawings. Needless to mention that the fewer connecting elements are the better when inking separately such drawings that have to be joined later on. Before the two drawings could be joined, MHB 164 had to have enough background space to accommodate the lintel drawing (Image/Canvas Size).

Adding space to the right side of the canvas

Importing all the layers from MHB 164 onto MHB 163 was done by selecting all the layers on the layer panel and simply dragging them over MHB 164 using the Move tool. One has to keep in mind that newly imported layers always placed right above the most recently selected layer, therefore one has to be very careful with choosing the right place for all the new information. When moving this much data on multiple layers it’s very important to keep the new layers’ relative position to each other, therefore it is advised to create a temporary folder to drop all the new layers into. When this folder is transformed in any way (moved, scaled etc.) all of its contents transforms with it, even if not all the layers are selected. It’s also strongly advised to lock these layers once they are dropped onto the new canvas to avoid any misalignment happening by mistake.

MHB 163 (left), 164 (middle) and 165 (right) prepared for publication (Click each photos to discover in full resolution)

MHB 163, 164 and 165 prepared for publication - detail (Click photo to discover in higher resolution)

As soon as MHB 163 and MHB 164 were joined, the appropriate layers on both drawings were merged to avoid duplicate versions of the same type of information (e.g. Line Drawing, Trace, Paint etc.) and last but not least to decrease the file size. Once this was done, the entire process had to be repeated when adding MHB 165 to the group. The end product became a 1200 dpi, 1:4 (106 X 78 centimeters, 50079 x 36850 pixels) replica of the entire wall section that contained 2,73 GB (11,6 GB at editing size) of data represented on more than 20 layers.

As a final cosmetic intervention to the appearance of MHB 164, the doorjambs, which were inked as part of MHB 163 and MHB 165 were included with the lintel to make it more relevant for the drawing’s separate publication format. Furthermore, architectural lines were added at the bottom on a separate layer indicating the contemporary ground line and giving the drawings’ overall appearance the right balance.

The composite drawing prepared for publication (Click photo to discover in full resolution)

The composite drawing prepared for publication - detail (Click photo to discover in higher resolution)

Reconstructing the early phases

Normally this is the point as far as a regular Epigraphic Survey drawing is taken before it is being published. However, as it was already mentioned at the beginning of this chapter, this wall section has two reused blocks used as part of the doorjamb that originally belonged to the Thutmoside section above the doorway. To reconstruct the original appearance of this area, certain preparations had to be done on the composite drawing itself along with studying the many parallels of contemporary doorways coming from the same monument.

Doorways at the Small Amun Temple

A regular doorway at the Small Amun Temple has 1-3 vertical columns of inscription accompanied with 1-3 rows of horizontal text that appear below a winged sun disc. The most common scenario is a 2-column doorjamb paired with a 2-row lintel, but the Small Amun Temple’s doorways have a wide variety of the above combinations.

Doorway representations from MH IX

The outer bark sanctuary doorway has only one vertical row of inscription on both sides and its context indicates that text started just a few centimeters above the remaining parts. Therefore, it’s safe to assume that the original lintel was placed right above the last surviving doorjamb block making the Thutmoside doorway much lower that its later counterpart. Furthermore, the doorjambs’ inner sides were cut back during the late Ptolemaic repurposing of this space giving room for a much wider door, the door socket of which is still visible on the underside of the new lintel.

Single door socket cut for late Ptolemaic door

Unfortunately there are no direct parallels to the original Thutmoside doorway’s appearance. The main entrance to the Bark Sanctuary on the east side was - although not replaced, but - heavily modified, while the inner side of the doorway in question was elaborately rebuilt by Ptolemy VIII. However still the closest parallel related to the outer lintel on the west must be the one facing inwards on the east side of the bark sanctuary. It has one single column of inscription on the doorjambs and one single horizontal line of text on the lintel, right below a winged sun disc.

Doorway lintel facing inwards the bark sanctuary on the east recarved in Ptolemaic style

According to the measurements of the space available above the doorjambs, it is impossible to fit in two or more horizontal rows of text below the kheker frieze, which ran across the entire wall according to the reused fragment depicting the winged sun disc. It’s important to mention that there are other blocks that possibly originate from the Thutmoside doorway. Some of them were reused in Ptolemy VIII’s elaborate gate, as a few peculiar profiles suggest on the other side:

Cavetto cornice placed sideways by Ptolemy VIII as part of the new gate on the western interior of the Bark Sanctuary (OIC 17618 - detail)

When creating a visual reconstruction, a number of possible solutions were considered. The simplest way to indicate the earlier versions of the section above the doorway could have been done on a separate drawing. However, there was an attempt from the Survey to keep the original drawing as intact as possible and represent the reconstructed versions on separate layers directly above the inked composite drawing. Enhancing the already complex file with more layers of information was a real technical challenge and required careful planning. Once again, the file size had to be temporarily reduced to be able to work with the additional layers, therefore the composite drawing was cropped to focus on the central area.

Repositioning the Reused Blocks

(1) As they appear on the wall, the reused blocks are part of the recreated doorway’s heightened doorjambs. The one on the left is used in its original orientation, while the one on the right is upside down.

(2) Large gap in the center shows how many of the original wall blocks had to be deconstructed and rebuilt to make room for the elaborate doorway constructed by Ptolemy VIII.

(3) The new placement of the reused blocks indicating their original position on the wall. One block depicts a winged sun disc, while the other bears the remains of torus moulding and cavetto cornice.

In the first step the two reused blocks were selected, duplicated (preserving their separate layer structure) and new groups were created for their layers. The two groups were named as Reused Left (2) and Reused Right (6). Both groups received their own white background layer (5) by painting a white panel (white outline painted by the brush tool filled with white color) behind each reused block. The inclusion of these white background panels was necessary to block out the reconstruction layer without deleting any of its segments, so they can be revealed again if necessary for refinement. In the next step the entire new addition to the Thutmoside doorway had to be blocked out in the same manner, adding a much larger white background panel above the central area (19).

Once all the necessary obtrusions were covered up, creating the reconstructed central decoration started with adding the vertical and horizontal architectural features, such as outlines for the torus molding and cavetto cornice (17). To create the actual drawing, a single weight 6-pixel wide brush stroke was used which is the equivalent of the regular trace line weight. Kheker, torus moulding and cavetto cornice templates from both the Thutmoside and the early Ptolemaic phases were created to fill in the architectural space (8-16). Regarding the lintel reconstruction, many contemporary doorways were looked at within the temple building and a mixture of the most similar ones were used for this template. With just a single row of horizontal inscription to be reconstructed, the textual context had to be very condensed.

New layers created to accommodate the reconstruction data within the Reference Only layer group

Blocking out the Background

(1) Reconstruction layer showing the area where the reused lintel fragment has to be placed.

(2) Multiplying the reused block’s layers helps finding the right alignment within the system.

(3) Painting a white panel behind the fragment blocks out the background.

(4) Blocking the background is non-destructive and it can be revealed again if necessary.

Luckily the available space allowed us to be fairly certain about the exact layout. Vertical spacing and hieroglyph size on the doorjambs defined the correct measurements for the horizontal signs as well. A short version of Amun’s name was used next to the cartouches at both ends of the symmetrical inscription. This was based on the name’s appearance on the opposite doorway leading into the inner sanctuaries (see: The Epigraphic Survey - Medinet Habu IX. The Eighteenth Dynasty Temple, Part I: The Inner Sanctuaries - With Translations of Texts, Commentary, and Glossary, Plates 14-15). The vertical text panels next to the lintel that belong to the wall scenes on the left and right were left blank, with only the corners of the sky added for reference.

Cropping the composite drawing to a much smaller size was necessary to be able to accommodate further layers representing the reconstructed areas.

Reconstructing the Original Layout

(1) Hypothetical architectural lines representing the horizontal dividers of the upper section. The fragment with torus and cavetto remains doesn’t have a precise horizontal position within this space.

(2) The repetitive elements were added to the upper section by using the same templates that were originally created to help with penciling the wall scenes.

(3) The lintel representation and text layout are based on the available space and certain parallels within the Small Amun Temple.

(4)The upper section’s painted decoration went through a serious design change during the first Ptolemaic phase. Only later paint traces of the reused blocks are shown with this version accordingly.

Based on the Ptolemaic paint scars to be found on the winged sun disc fragment, it’s safe to assume that the lintel itself was heavily modified in paint during the early Ptolemaic phase. However, there are neither direct parallels nor other known fragments that belong to the lintel indicating the exact nature of this modification. As far as we can tell from the wall scenes around the doorway, there were no modifications carried out on the carved decoration itself, therefore the early Ptolemaic reconstruction appears on the drawing with the same lintel layout as on the Thutmoside version.

Comparison of the late Ptolemaic (left) and Thutmoside (right) doorvay (Click photo to discover in full resolution)

Comparison of the late Ptolemaic (on the left) and the Thutmoside (on the right) doorway - detail (Click photo to discover in higher resolution)

Once all the reconstructive elements were in place, the entire group of these new layers were imported back to the original file and became part of the Reference Only layer group. Moving and placing the new layers onto their right place was carried out in the same way as was done earlier when MHB 163, MHB 164 and MHB 165 had to be joined.

Finally, the size of the doorway had to be changed as well according to the late Ptolemaic modifications to accommodate a larger door the signs of which are visible on the doorjambs. As part of Ptolemy VIII’s elaborate redesign, approximately 10 centimeters were shaved off of the left doorjamb and about 8 centimeters were taken from the right one, while significantly raising the hight of the doorway, as can be seen on the comparison drawings. To make the reconstruction drawing reflect this change the architectural lines of the existing doorway were covered with a white panel in the same manner as was done earlier, and new hypothetical architectural lines were created on the layer containing the existing architectural reconstruction (17).

At the end, the opacity of the entire reconstruction layer group was changed to 70% to give it a grayish hue with a more discreet appearance as opposed to the existing decorative elements’ solid black. As a result of careful planning the entire drawing from the first sketch to the last of the reconstruction layers could be kept as a single file, allowing us to display this enormous amount of data in numerous different ways.

The Western Outer Wall of the Bark Sanctuary in the 18. dynasty - finished reconstruction (Click photo to discover in full resolution)

The Western Outer Wall of the Bark Sanctuary in the 18. Dynasty (finished reconstruction) - detail (Click photo to discover in higher resolution)

The Western Outer Wall of the Bark Sanctuary in the early Ptolemaic period - finished reconstruction (Click photo to discover in full resolution)

The Western Outer Wall of the Bark Sanctuary in the early Ptolemaic period (finished reconstruction) - detail (Click photo to discover in higher resolution)


0 comment(s)

Leave a comment(We'll keep your email address private)

Want to stay in the loop and hear first-hand about digital EPIGRAPHY's latest articles and future plans?