Car Body - Adding Detail
Tail Fins - Page 3
By now, our tail fins are looking pretty nice, but are just a little bit light on believability. The 'separation' occurring between the tail fin and the shelf for the tail light might be okay under some circumstances, but for the car we are modeling, it is not as accurate as I'd like. There should be a better blend occurring between the two elements. Now we will go about adding this blend.
Note: I have hidden the polygons that will make up the trunk, for viewing clarity in this step only. You have not missed anything!!! (Just in case you were worried!) ;-)
The first thing we will need to do is to get rid of some of our problem polygons.
Since we are manually 'mending' this tail fin area in this step, we will be welding vertices together to make a more efficient cage. However when we streamline the number of vertices, we will end up with extra polygons. To avoid confusion, it's best to just determine which ones are unnecessary and get rid of those.
Now we have a clearer idea of what needs to be done when we look at what's left.
The general idea now is to move all of the points from the tail fin to match exactly the positions of their corresponding points on the shelf area. As you can see, the shelf points are highlighted green in the picture, while the tail fin points are highlighted red. What you want to do is use numeric entry via the structure manager (in point mode) to match coordinates of red points (A, B, C, D, & E) to the positions of the green points (a, b, c, d, & e).
Just so there's no confusion, 'A' should match to 'a', etc.!!! (Like instructions on shampoo... sigh...) :-)
To the left I've shown a couple glimpses of the technique in action as the last couple of points are matched up. In the bottom picture, all are matched except the final z coordinate which is highlighted blue (about to be entered). Luckily my points were close enough on the structure menu I didn't have to go scrolling too far to see my target coordinates. Sometimes it's good to have a piece of paper and pen handy for this though...
Note: I suppose you COULD have matched up points from the shelf to the points on the fin (opposite to what I described), but then you'd have to adjust the flatness of the shelf again by moving all the points back down into position...
So now we can see that we've got a very sharp edge between the two elements, which might be acceptable...
...other than the fact that it creates an unsightly 'kink' right below the edge along the side of the shelf.
...And the fact that the points are occupying the same place, but not joined. They very easily could split apart if we were not careful...
The left picture shows the result right after the optimization. Because now there are suddenly less points & polygons defining the same shape, the shape becomes immediately looser and more amorphous again. Don't look at that as a step backwards! It's not... Merely select the inner polygons of both the shelf faces and the fin back face (leaving the edge polygons untouched), and use the smooth shift tool to create an inset. Just smooth shift a small amount to get a sharp lip... or smooth shift minutely twice in succession to get a really sharp edge.
Notice also that because the edge polygons were left out of the smooth shift operation that there is a slight curving blend between the fin and shelf elements... That was very fortunate, as it brings together the elements naturally and completely reverses that separation effect we were looking to eliminate!
In that case, what we'll do is select all the points (make sure that both points of each 'pair' of points occupying the same coordinates are selected).
Then use the optimize command to make the selected points 'weld'.
Two points within the specified tolerance become one point...
One caveat with this step, is that the optimization of points the way I have described has led to a triangle polygon right on the side of the fin. Earlier, I mentioned that triangulation was bad and should be avoided.
Now I must admit, that sooner or later, I inevitably end up with at least some triangles. I try to keep them off to a relatively harmless location though. If you look at the wireframe, you can see what sort of affect this triangle has on the distribution of the detailed mesh. It is somewhat irregular, but with only one tri, it is fairly minimal, so I'm not overly worried.
I think the main problem involved with triangles is that they can lead to irregular shading properties, and of course the mesh distribution. But also I imagine that this can have a negative impact on the UV of the mesh, making it difficult to texture under some circumstances.
However, the fact that this is a car body and shouldn't require complex UV mapping (most likely it will end up with a basic paint shader covering the entire body) I really don't see this triangle being anything to worry about. Depending on how adamant you are against triangulation in your HyperNURBS mesh, you may decide to rework the point or polygon structure to remove this and stick to quads only... (However, then you are on your own!) :-)
The rest of us will be enjoying the next step!