Ancient Spaces Blog

For my extra work this week, I worked on many parts of the Intiwatana. However, none of them are at a state to make a wiki-post.

First of all, the textures. As always, textures could become a pain when a part or parts of a wall is not visible. In the Intiwatana structure 2 for example, the destroyed walls took forever to "reconstruct". Furthermore, if it's a realistic model we are aiming for, it means that the rock creases of each edge must be seamless. Now, that's the hard part. By trial and error in Blender and Photoshop, I was able to create a somewhat decent texture for it. Nevertheless, there's room for improvements. Another detail added to the realism is that, the rocks must have the same tint if they were to be tiled. Some previously rendered pictures do not have that, thus it looks weird. In the coming version, that issue should be addressed.

The rest of the landscape is now almost done. In order to export to Ogre properly and efficiently, I joined all the landscape objects as one object. Yes, joining them may be easy, just select all and press Ctrl+J, but to actually optimize it, ie, to get rid of the faces that are never seen, or welding the vertices together could greatly reduce the polycount. Yet, time-consuming. This is also difficult because the sundial has blocked parts of the wall that's essential to the complete model! With the plan inaccurate, I had to approximate the distance and size of the blocked wall.

So far, the polycount is adequate. It does not slow down the rendering and we still have a decent FPS. Matt and I discussed about this. We should make a realistic/photographic texture for the stairs or even build the stones rock by rock. This way, it would greatly reduce the time in texture editing and improve realism drastically. Rock by rock method is not a problem here, because, we can handle a lot more polygons than we thought. Therefore, with this method, we should be able to complete the entire Machu Picchu site within the deadline and with great quality.