Fore Blog Post 5

Final Lines on Fore Abbey Cloisters

Moving towards a model end

The deadline for completion of our 3D project on Fore Abbey’s Cloisters, led to a great concentration of work and a forward momentum that meant our initial concentration on detailed site measurements fell somewhat by the wayside!

We had created our arch array on one machine and worked through our textures and overall cloister area on another and were confident that using the plans of the abbey as a guide would allow us to scale and adjust with some ease.

The Benedictine Abbey remains at Fore in County Westmeath are impressive in their ruined state, both from the sheer scale of the remains, but also from their position in the landscape. Tucked into a valley near Lough Lene the Abbey is a striking visual monument and one can only imagine the awe and sense of wonder it would have cast in its heyday, as it would have been the most impressive manmade object that many of its visitors would ever see.

To do it justice in a 3D rendering we measured what we could of arches, bevels, pillars and physical space. We looked at similar abbey constructions and leant heavily on the OPW plans drawn up by H.G Leask many decades ago. Discussing the project with our supervisor, with some of our fellow students and with an expert in 3D graphics we became aware of the difficult nature of the task but were determined to work towards a model that could form the start of a long term 3D visual project on the Abbey.

Fast forward to the days and nights before the deadline for the project completion and we found ourselves with a mountain to climb in terms of creating the cloisters to a level we would be happy with. One of the most difficult issues was applying the textures that we had intended to use.

 

Applying a texture by loading a bitmap into the material editor we applied it to the main walls of the cloisters and turned on the shaded material with maps viewport setting – but time and again after applying a UVW Map modifier, we lost sight of the material as we utilised the gizmo to rotate and resize the map. After spending hours with no great progress we tried to divert ourselves by creating a compound material using a series of textures we purchased, and which looked very impressive when properly applied but here again while we came close to a result we had spent far too long and achieved very little (which with a deadline approaching is akin to nothing).

We returned to applying a simple texture and went back to the UVW map and after struggling with the Planar/Box selection on the entire element – we tried to apply purely to polygons and set Poly ID’s – applied the material and then the UVW map. But again a problem, as soon as we moved off the Poly, the material seemed applied to the entire element and so with much wailing and gnashing of teeth we selected the Edit Poly modifier and began again. This time the UVW map modifier appeared to stick and we began to work on another part of the walls as a separate poly – disaster – the UVW Map gizmo affected the original poly and we got lost again.

After many fruitless attempts with these settings we managed to apply the the UVW across the full element as a box – and achieved a reasonable result and began to look at the roof tiles. Here we had more success when we altered a setting int he incoming dialogue box from ‘Explicit’ to Planar WXY’  and finally the map applied in a manner in which we could work at a basic level. Yet again it felt like we had stumbled through a maze without having a key to the ‘map’ but we were so close to the end that we kept pressing on rather than going back for directions!

Our final challenges arose when we moved back to the main machines in the Iontas lab and combined all our elements. In spite of our attempts to stick always to architectural plans and to keep our measurements as guides throughout, we had underestimated the effect of a small mistake when amplified across 14 cloisters. The arches were slightly off scale and the more we attempted to adjust the cloister square, the further askew they seemed to go. We debated how to proceed and were at a loss, until we remembered that the one area we were totally reliant on guesswork was the height of the cloister roof pitch. Our initial creation was constructed on photos which showed the remains of a connection seal or gutter high up around the cloister square walls – a point which we had never measured, just extrapolated from the photographs. By re-adjusting this point, which we were estimating to being with, we were able to adequately scale the arches array and press on with the model.

Once again however the texturing would not apply correctly – perhaps we had used the wrong shaders, perhaps our rendering was badly set, but as the night wore on into morning it became clear that the texture we had chosen would not display as we would have wished and we were unable to resolve it. Nonetheless the cloisters looked like cloisters, our roof construction had solid beams under the texture for support and although not quite up to modern roofing standards our guttering was based on similar abbey designs and gave the roof a more finished look. The grass in the centre of the cloisters was a combination of a purchased texture and a patchwork using free textures adjusted in a photo layering app – again we needed more time to apply a bump map and create a more pleasing look, but we contented ourselves with the look achieved and the addition of a single fallen stick gave a sense of realism to an Abbey which would have been surrounded by forests in its heyday.

Looking back on the entirety of the project, both myself and Marianna, who worked with me, feel that we lost a lot of time trying to create a very accurate pillar for our arch array and that the time to cut our losses on this part of the project was missed. We spent far too long working and re-working combinations of cylinders, splines, and modifiers and in the end we wound up working on a version that was as close to the photos as a version we had worked up weeks before.

Our enthusiasm for the project never waned and in spite of our not being entirely happy with the end result, I think we both feel that it is a very good point for us to move on from as we look to create a piece of work for the community to use in their promotion of the Abbey as an important historical and spiritual centre that deserves to be preserved and researched to a far higher level than at present. Throughout the project we talked to local people on the ground, several tourism co-ordinators and historians connected with the Abbey – these connections have convinced us of the need to create something that can further the future of Fore Abbey and perhaps restore some of the awe and wonder that would have been felt by those who saw it in its heyday.

They say there are ‘Seven Wonders of Fore’ – the eighth wonder is perhaps that so little is made of this gem in the middle of Westmeath and we ourselves are left wondering if someday we can make a model to bring this amazing site once more back to life in a digital realm undreamed of by the monks who first textured its walls with the rough stone of the valley and breathed life into its cloisters with their prayer and song.

Textures, Wraps and Scales

Fore Abbey Cloisters

Wrestling with Wraps, Maps and Texture Aps

Our efforts to model the cloisters at Fore Abbey have moved from the large scale idea of the whole abbey as a loose model, towards a more concentrated look at the cloisters which sit in the centre of the abbey space. The cloister area has several columns and arches still standing and as such there are visual references as well as our measures to work from.

Throughout the process, the most difficult task has been to correctly figure out the start point to construct from. We have drawn and extruded from splines, we have taken basic shapes and used modifiers to morph them towards our pillar shapes and we have used boolean operators to unify sets of cylinders – none of these efforts were entirely satisfactory but they did lead us back towards a spline based object to work from. The principle issue with our pillars was a massive poly count and so we removed modifiers especially the Turbosmooth modifier that had beautifully bevelled our pillar caps and edges.

Redrawing from a spline, we used the CapHoles modifier to fill out a new pillar and then reused our original bevel base and tops but added a chamfer rather than Turbosmooth.

Returning to the overall model we used boolean operators to extract door spaces in the outer cloister wall and began to look at our arch shapes. For these we used a basic box shape and drew cylinders over the plan of the arches and then used a ProBoolean subtraction to cut holes through the box to make our arch. These we then took into a separate 3Ds Max project to combine with the pillars and began to create an array.

Having worked together on the one computer throughout – we began to press further into the project on two fronts. Marianna took the arches and began to reshape and tweak them to create more realistic looking shapes as I began to seek out textures and looks.

Through conversations and our own research we decided that aged concrete might give a good look to our pillars, arches and the walls on which they stand.  We also sought out medieval textures online and scanned many websites for comparable buildings that might give us further visual cues to work from.

The outer walls still standing are composed of rough stone and for this reason we chose a texture that, when combined properly, would give a sense of the rough, solid stone that has stood for the centuries.

As the cloister began to assume some kind of shape,  we worked on a schedule to try and finish the model in so far as time and our skill-sets would allow.  The texturing and the scaling into a finished form were the areas we hoped would push our flat looking construction onto a different plane and as we found a wonderful interior roofing texture we entered into our final weeks work on a hopeful note.

Fore Abbey Cloisters

Inch by Inch Stone by Stone

Trying to build a virtual stone wall

In taking on Fore Abbey as a 3D visual image project, we always imagined that there would be some hard labour involved. The Abbey was a vast space, added to and amended through the centuries as it grew from its original 7th Century wooden structure through to the fortified stone towers and vaulted arches of its high point in the 13th Century. But when we decided to rescale our efforts and work solely on the cloisters we thought it might be a simple task.

There are several columns and arches still standing as part of the cloisters on the site – with free access to the site I took a lot of photographs to use as visual guides and we consulted the plans that had been drawn by OPW architect H.G Leask for a pamphlet in the 1930’s.

By comparing the plan and the photograph we struck on a plan to create multiple cylinders on top of the plan but to begin to create the cylinders using the measurements from my trips to the site as a starting scale. In this way we believed we would find an acceptable level of accuracy in the construction, mindful that although our brief was to create a visualisation for a television documentary, we still wanted to stay as close to the real measures as possible.

Our first attempts led to a very confused looking series of shapes that became far to complicated when we tried to extrude them to height. We had measured the primary rectangular base and constructed a 1inch box, we then separately constructed what we hoped would be the next three beveled bases that sit on the rectangular base. The measures from the site were accurate but not professionally done, when I measured a beveled edge I did not have the architectural or archeological nous to produce a proper diagram of my figures and consequently there was a confusion in the construction.

Nevertheless we brought the cylinders to a point and then tried to combine them into one element using the boolean union tool. In this we were somewhat successful but it took far to long to create our basic bases and somewhat disheartened we turned towards the column themselves as they seemed like a simpler proposition and offered the chance to create a structure that would reward our hours of labour with a solid piece of geometry.

During the work, we discussed how we might incorporate the cloister model into a documentary style piece on the Abbey. We talked about shooting some more videos on the site and using these, allied with some pre-existing drone footage we could license, to incorporate our construction. As we talked we decided to look at two possible video integrations, one was to overlay the cloisters into the drone footage, the other was to license the plan model from Leask’s drawings, or to create our own and add the cloister model on top using a simple crop or transparency option to reveal the 3d Image.

Returning to the model as it stood, we struggled to create a low polygon unit out of our cylindrical assembly. We found the align tool and moved and measure our way to a series of objects that we could work with, but found that we were, once again, hours behind our hoped for schedule and parked the model where it was – hoping to reconvene having done more research into the areas of the software that were taking us so long to overcome.

I suggested I might try and model some of the stone wall upon which the cloister columns stand and that we would try and talk to others with more knowledge of the software or the 3D visual field in order to find simpler ways to progress the project.

As ever I am amazed at how difficult it is to find the right way to start into a model even with multiple tutorials to follow on the web. The old Irish line ‘well I wouldn’t start from here’ comes to the Fore far too often!