Making Of 'Hagia Sophia'

About the image

The idea to make this work come from an old drawing of Hagia Sophia published in a book. I want to make a image ussing imagination, far from my normal technical work, so i dont use any other photographic or techincal references. All modelling, layout and render has been made using Lightwave 7.5 without any speciall plugin. Textures and final compositing was made with Photoshop 6.0. This is a exercise to apply all my knowledge in technical visualization to a fiction work, who is now my professional aspiration.

Modelling

To model all scene i divide global building into small pieces made at arbitrary scale, becouse no technical references were used. I think that big scene is only a joint of small pieces. I use basic modelling tools, a lot of booleans operations, bebels, an so on. I want to add a lot of details to improve realism of image. The key for this is irregularity, i use booleans to add imperfections to wall edges and jitter all small componets of every part of model. I also avoid seams by add small displacements to all repeat elements like windows and so on.

Street has been made from a plane with a lot of jitter and damages to integrate in global style of image.

All modelling is made with polygons ... no subdivision surfaces used. I spent about 40 or 50 hour in all process. Final building model has about 810 K.Polygons.

Texturing

Perhaps more important feature of project. I texture each part individually. I dont use any kind of procedurals, only image texture maps. Usually i use some kind of stone texture base, from my own photograph library. Then i add by layers all dirt, humidity spots, etc, ussing photograph noise as alpha channel to join them.

Second pass of texturing is add horizontal dirt in all hard edges. I use gradients from distance to pivot point to add these layers, up to 10 layers in some elements. Then i add other small details using noise image maps with alpha channels in each position who need it, like all humidity noise in windows, edges, etc.

I use bump, difuse and specular maps too, specially first one, to add old and irregular aspect ... perhaps three or four layers in final stage.

All scene has only 21 image maps ... but some objects has more than 25 or 30 layers to create final material. Only minor work was made in Photoshop, because all layer compositing is created direct in Lightwave.

It was a very complex work ... specially to get all layers in his correct position and to avoid seams. I always add a random displacemnt and scale to base textures in each element to avoid maps continuity. Perhaps i use more than 25 hour in all process.

 

Layout and Lighting

I finish global scene in Lightwave Layout. I load and postion and scale individual part of building to match reference drawing. I add some dark trees and other small vegetation, stones, etc, to increase detail of final image. Final scene has about 1.700 K.Polygons and 200 individual objects.

Background comes from a retouched sun set photopraph take by me here in Spain (beutiful light in this country ....)

Direct lighting comes frome one yellowish area light source that simulate sun from right. Montecarlo sampling is used to add GI solution to final image using 6x18 samples. Ambient lighting comes from a sphere mapped with a sunset photograp (no HDRI). Only minor point lights used to add some datail in certain areas, but effect is near unnoticiable. I think that this is a good example of how well Lightwave render engine works with relative small lighting setup.

Rendering and Compositing

Final image at print resolution (3600x2248) was rendered using default LW renderer in four computers of my renderfarm (P.IV 2.4 Ghz.) who spent about 12 hours each to render (total render time about 48 hours). Perhaps a bit slow render, but i dont use any special optimization in rendering process.

To solve plain aspect of main model and background i render a separate pass with a volumetric lighting solution. To make this pass i delete all minor elements and turn background and objects black with mate color. I use a spotlight with standat Lightwave voluetrics to create image. Finally i compose this fine tunning process in Photoshop retouching final image and volumetric contrast.

Best Regards,
Juan J. Gonzalez Diaz

www.genesisvisual.com
Juanjo@genesisvisual.com

About me

I am a 32-year old self-taught computer graphics specialist working in this fields in the last 14 years. Actually my daily works is about technical visualization and architecture, but i work before in video industry, computer graphics software development and real time applications. Now i am very interested in introducing into fiction projects as environment designer and i am working in other personal projects like "Hagia Sophia". My favorite 3D application is Lightwave with my eyes put over Softimage|XSI and Cinema 4D.

Fetching comments...

Post a comment