4/21/2000
This is my first image with Carrara and Canoma, products formerly owned by MetaCreations, now owned by Adobe. Canoma was used to create a 3D object from photographs of my home, while Carrara was used to create the 3D world in which it is placed, including the trees (TreeDruid extension - Pine, Generic Bush, and Elm), mountains, clouds, and the depth of field effect.
A combination of the bark and moss shaders was used for the tree trunk / boughs / branches and the leaves, respectively. The house is texture mapped from several photographs imported into Canoma. Some of the resulting texture maps were updated by hand.
The moss shader was also used for the grass, but due to anomalies in the way Carrara parametrically mapped it to the terrain, it was necessary to flat projection map it from the top and drastically increase the bump map effectiveness (to the maximum possible). The mountains use an inverted Begley Yosemite shader, which offers a fairly realistic dark limestone / snow look for the distant mountains.
The clouds were created above a low Four Elements cloud layer using spherical cloud objects that were flattened and arranged at varying distances between the foreground terrain (a carefully hand adjusted smoothed terrain) and the rough background mountain terrains.
Some of the blur effects on the clouds near the foreground terrain were fortuitous effects generated by a likely bug in the Depth of Field algorithm. However, since they mimic the normal appearance of close clouds so well, I decided to accept them.
Moon placement was extremely difficult. I found that the use of a strong moon halo made it easier to determine when the moon was close to the desired position. My original desire had been to place the moon almost directly behind the mountains, but the granularity of the Four Element Moon placement was such that the moon was either where it is, or below the horizon. To some extent, I think this turned out for the best, as the incomplete image of the moon lends a greater realism than would be produced by a completely planned composition.
Also, it is interesting to note that the moon is considered to be a foreground object by the Depth of Field algorithm. Therefore I had to render a completely sharp image as well as the one above, and paste in the sharp moon.
Finally, one of the most important considerations lay in keeping the sun angle consistent with the requirements of the photographs used in Canoma. Canoma generated objects only look realistic in environments where the lighting is at least similar to that in which the photos were taken.
However, even within that constraint, there are a variety of possibilities. The following early renderings were among those used in the basic lighting decision making process:
 |
 |