new models - castle towns

LinkyNStoof · Post by **LinkyNStoof** » Tue Aug 31, 2004 4:19 am

With some help from AutoFlip, I'm finally relatively ready to release this model of a castle town.

It is however rather large (15x13 area-wise, and height wise I think it reaches nearly 7 tiles high at certain spots), so perhaps it's not the most practical custom model to use in a level. And I have to admit it looks mediocre, but sorry, that's basically how the model was originally.

There are 2 models, because in the original model, there's both a front and a side entrance to the castle town. So I made two orientations, one where the front entrance faces you (model1.3ds), and the other where the side entrance faces you (model2.3ds). For some unfathomable reason, despite only having rotated the model 90 degrees, the second model has much brighter hues than the first. Go figure.

The dead spot for both models have been set to where that well/pond thingie is. You'll have to play around and see where to lay out the X floors. It is a good idea to allow Stinky to enter the buildings, or at least to step into the entrance of the various buildings (and then maybe have a teleport whisk Stinky away or something).

If you really like the models but want to try out a smaller size, PM me and I'll look into it.

[EDIT: over the next few days I'll try to edit the colors so they are more consistent. The first model's colors are too dark and the second one's too bright, even though they come from the exact same original model]

Jersey · Post by **Jersey** » Thu Sep 02, 2004 1:57 am

These look great! Very medieval! And a darn sight better than what I could come up with!!!

I'm gonna try them out when I get up the nerve to add custom models to my levels...never mind the nerve, I need to figure out how to do it.!

I'm not the sharpest tool in the shed, if you catch my drift.

majic1 · Post by **majic1** » Thu Sep 02, 2004 5:10 am

These could be FUN! We should get some VERY interesting levels with these!

popo · Post by **popo** » Mon Sep 13, 2004 3:39 pm

Hey LNS, these are awesome.

I can really appreciate all the hard work that you've put into these, and they look great, especially the second one. I've just spent all day reading all the posts that have been put up while I've been away and I see there are some levels already with these in them. I can't wait to see them in the flesh as it were. It's gonna take me a while to get up to speed with all the new levels, and I can see that I'll get no housework done for at least a week.

I also noticed that you're up to version 1.8 on your Autoflip, what happened to the ones in between?

LinkyNStoof · Post by **LinkyNStoof** » Tue Sep 14, 2004 1:47 am

popo wrote:I also noticed that you're up to version 1.8 on your Autoflip, what happened to the ones in between?

Well I was going to call it 2.0, but then decided I shouldn't do that until it has been further tested (hint, hint). Also I was planning to make a few minor tweaks to the "new feature" I had for version 1.8 so that's why I set the version number a little away from 2.0.

The "new feature" is really just a way to (hopefully) have AutoFlip get more of the triangles facing in the right direction, so there are less for you to flip yourself. It's based on the assumption that usually, there are more parts of a surface that has an outward curvature (like a sphere, or adjacent faces on a cube) then an inward curvature (like a dent), and so, each time it finishes making the vertex order consistent across a bunch of adjacent triangles, if it finds that there are more inwardly-curved surfaces compared with outwardly-curved surfaces, it flips the entire bunch of triangles, and thereby hopefully getting those faces all correct.

This works fairly well in the limited number of models I've tested, namely the ladybug, the frog, and your bench (the bench it did a perfect job on, btw; it got everything right). It does however have a defect in that when you have a flat surface composed of many triangles, it might confuse AutoFlip into doing the wrong thing. I had ways to deal with that, but I'd rather have the program more thoroughly tested first and see whether it really needs extra tweaks.

It worked okay for the castle, but I still needed to do quite a bit of manual flipping in the end, but this is partly because the castle model itself has a bunch of actual errors (examine it carefully in AC3D and you'll find things like actually missing pieces of back walls etc.)

It'd be great if you can test it on the plant-generating program and see how well or terrible AutoFlip does.

popo · Post by **popo** » Tue Sep 14, 2004 7:26 am

The trouble with the plant program is that the surfaces are the right way for the models, but, because you see the backs of some of the leaves and petals they appear invisible. I'm not sure that Autoflip would be able to differentiate between the ones that are facing the right way, i.e. forwards, and the ones that are facing the wrong way, i.e. backwards. What I've tried to do myself is turn just the parts of the plant that you can see the backs of, but it's more trouble than it's worth because the plants have very high polygon counts. If Wonderland could cope with double sided surfaces it would be much simpler but that would double the polygons too and make the models too big to use.

I'll have a scout through the models I've got and see if there are any to test the latest Autoflip on. The trouble is a lot of the models look ok until I try them in the game, so that means trying them all in the game first to see if there are any holes. It may take some time so be patient.