Originally Posted by
Conscars
I purposely chose to model it low-poly until I was finished the general shape. The reason was that I had no completly accurate reference images to work from until today. I don't want to waste time and effort modeling in details when it could be all wrong. Don't mind the random poly's, they're 'artifacts' of my deleting rampage when I discovered a bunch of stuff was wrong. I was discussing my modeling method with Phopo on AIM, and I have come to the conclusion that it should not be used unless your reference point has more than 1 axis that you can use to align with your reference images. Valhalla is a bad map for this. The terrain is very ambiguous, I don't know anything for sure about it's shape and angles. I chose the bases as a reference point because they were the only things in the map that had straight lines and definite angles. The problem is, my base still ended up inaccurate from the lack of proper reference images from it. Besides modeling a map from an inaccurate reference object, the base itself does not extend along many axis except for the vertical one. This doesn't help me at all. To like up my cameras with it to model from, they need to be aligned to the base's rotation. It's hard to tell if your image in the background lines up with the base model because it doesn't extend close enough to the camera. This has lead to the map ending up a bit warped. I do not reccomend this method to anyone who wants to model an OUTDOOR map. Camera-triangulation modeling is great for any map that has perfect angles and flat surfaces to reference, but not outdoor ones where the base is all that you have. That being said, I have the layout comparison below.
(click to enlarge)
[wshot]http://img508.imageshack.us/img508/8216/accuracyeb1.jpg[/wshot]
As you can see, some parts of the map ended up quite accurate while others did not.
By the way, does anyone know how I'm going to get grass, stone, AND snow detail maps into my ground bitmap? Or is there something in H2 that will let me do that.
edit: thought of a solution already; since these 3 surfaces don't all meet together anywhere, I can just have 2 ground bitmaps. One will have grass/stone for the center, the other can have grass/snow for the edges/cliffs.