Everything started with a lonely snowman designed some years ago when Alex was bored in school. We thought about the environment: What is stored in our memories from epic childhood days in the snow and ice? Recalling it led us to the good old Sodium-vapor lamps (λ ≈ 589 nm), an igloo, mountains, a star-bright sky as well as beautiful benches.
Fun Fact: The snow on the top of the Sodium-vapor lamps melted quickly after turning them on, since they emit a high amount of heat due to their bad energy efficiency level.
Modelling was done primarily in Blender. The sky texture is generated with the help of Spacescape, a tool for creating space skyboxes. The trees are generated by a Lindenmayer system in a procedural fashion.
We want our street lights to illuminate not only the floor but also the dust particles in the air.
Since volumes are too expensive, we decided to implement a fake volume using a semi-transparent cone. The cone was setup in such a way that it does not capture those.
We want our trees to match the natural beauty as best as possible.
Since modelling them by hand is a tedious process, we decided to use the grammar-based approach originally suggested by Aristid Lindenmayer. A paper by Sun et al.  made it accesible to us and inspired us.
We want the snowman's hat to look realistic.
In the original Blender model, we used the Velvet shader, which is commonly used to model cloth. After researching on how to reimplement this in our raytracer, we ended up with the Ashikhmin BRDF described in  and .
Moreover, we implemented randomly generated snow and the depth of field effect.
Let's put our effort in numbers.
Rendering Time: 20 min
# Triangles: 1 935 884
# Vertices: 969 832
# Faces: 976 466
 Krzysztof Narkowicz. Cloth Shading. Blog Entry, URL: https://knarkowicz.wordpress.com/2018/01/04/cloth-shading/
 Alejandro Conty Estevez and Christopher Kulla. Production Friendly Microfacet Sheen BRDF. 2017. URL: https://blog.selfshadow.com/publications/s2017-shading-course/imageworks/s2017_pbs_imageworks_sheen.pdf
 Ruoxi Sun, Jinyuan Jia and Marc Jaeger. Intelligent Tree Modeling Based on L-system. 2009. URL: https://www.researchgate.net/publication/216337877_Intelligent_Tree_Modeling_Based_on_L-system
The models for our street lights and benches are taken from Blend Swap.
Basic template for this webpage: https://startbootstrap.com/previews/scrolling-nav
Spacescape sky generator: http://alexcpeterson.com/spacescape/