[C4:AC] Fungus graphics options

[Kable]

And submerge some wheat and you've got a passable kelp farm

Oh, btw getting the "remove xenofungus" to work at sea was pretty trivial, probably not worth even posting the changed files... just clone the BUILD_REMOVE_JUNGLE element and call it something else, change the FeatureType to ice/whatever the new nifs are used to replace, then add the new function name to the UNIT_WORK_BOAT build options list, and voila.

If anyone does want me to post the changed files let me know.

[Rubin]

Kable, do you have any suggestions on land based fungus graphics? Would it be viable to modify the jungle feature .nif files?

[ForesterSOF]

What about replacing one of the bonus items and then in the map making make that bonus a big percent.

[Kable]

Here's a quick jungle replacement. You need to drop a texture named icepack_1024.dds in the folder with it, I didn't include it to save size.



I'll see what can be done about hills using this technique.

[Kable]

Terrain tracking in the forest/jungles might have something to do with the values in the UserPropBuffer from the NIF:

OFFSETCOUNT = 25
OFFSET1 = -91.4581,87.773
...etc...
OFFSET25 = 92.9705,-44.9156

I'll let you know when I figure it out.

[Rubin]

Kable, this is looking quite promising!

The submerged fungus may appear a bit too solid, but actually looks good.

The land version causes transitions between land and sea (coastlines) to look broken and only flatland tiles show correctly (which is an already known issue).

There is still something wrong with the transparency which I hope we are able to fix. Transparency would make the fungal patches look less solid and could add some extra detail to the sea fungus (if elevated slightly).

Here are two images of how the new fungus looks:

[Kable]

The offset stuff in the userdata area of the .nifs does seem related to the terrain tracking, as adding these values to the fungus files makes it jump up and down all over the screen, and it tracks the terrain, even if it's about half a mile off the ground :P. There are only two dimensions on the offset, so if each one is attached to a single tree, I assume they must share a common plane.

If I can prove this theory to be correct, I'll see what happens if I chop the fungus into small overlapping but separate pieces, each with a corner on the same plane as the vertexes referenced in the jungle file, and each with a 'skirt' that will avoid gaps when they move up and down, then put in offset information for the corners. Hopefully it will then track over hills and down coastlines. If it's possible to change 'sea floor level' somehow it might even allow a single fungus object to be used across land and sea.

Failing that there might be some SDK hackery options to trick the land and sea fungus into recognising each other as the same type of feature. This, combined with the terrain tracking land fungus, should mean they form a continous mass.

_________

On the semitransparency, I'm at work and I don't have civ installed, but I'd not noticed the alpha parameter which is available through nifscope when I was comparing the blender output to the originals.



The problem I'd had where it was either completely transparent or solid sounds exactly like it had defaulted to 'binary' alpha, so perhaps changing this to smooth would re-enable the semitransparency in the models. I'll try it when I get a chance, and post an update. If you get a chance to try it before I do, let me know if it worked.

[Rubin]

Kable, I have tried changing the alpha to smooth in NifSkope. This does not seem to have any effect on the transparency issue.

I then replaced one of your new jungle feature nif files with a corresponding nif file from the original icepack feature. The transparency shows fine on the original.

When I open the original file in NifSkope I see this property, which I cannot find in your modified nif files:

NiAlphaProperty

The tooltip says that the shape must be linked to this property in order to enable transparency.

Edit: It seems I am able to transpose the default ice feature using NifSkope without any loss of transparency.

Edit 2: Here's an image of an attempt at elevating the ice feature Z=5:

[GeoModder]

Wow!
Now, if you guys could manage rivers be visible at all times...

[Rubin]

GeoModder, you cannot see the rivers?

[murielwasser]

That looks awesome.

[GeoModder]

Originally posted by Rubin
GeoModder, you cannot see the rivers?

I meant to see the color of the water, blue, instead of the pink parts when it passes through the fungal plots.

[Kable]

Originally posted by GeoModder


I meant to see the color of the water, blue, instead of the pink parts when it passes through the fungal plots.

Now that might be harder, as they appear to be painted on. However, there are some custom variables in the river nif which are used to keep jungle/forest out of the rivers. They seem to just show up as cubes in NifSkope, so they are almost certainly using some kind of custom coding in the engine to handle it.

___

The terrain is done with the offset list in the userdata field on the nif (the offset appears to be in blender units X10 (approx), and in this example I've scaled down the fungus to 10% of it's original size and added offset information at the center of the new square (hence the large gap between tiny squares):



It's slightly more stable than when I just added the offset list from the jungle to the file, but squares still sometimes disappear, and occasionally jump in height. But it does track with terrain, as I've shown in the diagram.

I'm starting to think that using a forest as a starting point to model the overlapping blobs for a terrain tracking fungus floor might be the quickest route to a really nice looking land fungus. The evergreens have less polys and are kind of round and blob like. I'll knock up a quick test.

It's possible to alter the heightmap for the ocean so that forest appears, but it would seem as though anything below sea level gets removed. However when the ocean is brought above sea level, it looks rubbish, but forest will appear.



On the bright side, it appears to be textured.... perhaps if this texture could be identified and made transparent or removed, a single land/sea xenofungus feature could be implemented. Has anyone come across the deep blue dirt texture which appears on the ocean floor?

[Rubin]

Kable, there are quite a few water .dds files. You can try deleting (or brighten) the custom oceanblend.dds that I included in the 0207 base. Perhaps this is deep blue dirt texture you are looking for?

If we use a submerged fungus feature for water tiles (replacing one of the forest features) and a modified forest feature for land tiles (replacing the jungle feature), we'd probably have decent coastline transitions. Edit: Well, actually, there'll still be a transition issue, unless we are able to allow for multiple features on the same tile.

This is very interesting!

(Also, if we are able to have two features displayed on the same tile we wouldn't have to worry about if the fungus is on a water tile or a land tile.)

Edit 2:

Originally posted by Kable
It's possible to alter the heightmap for the ocean so that forest appears, but it would seem as though anything below sea level gets removed. However when the ocean is brought above sea level, it looks rubbish, but forest will appear.

Do you know why anything below sea level gets removed?

[Kable]

I assume somewhere in the unexposed code for dealing with forests/etc. there is an "if (z<0) visible = false", unfortunately.

I managed to get the ocean floor transparent, (don't know which file, I replaced every terrainblend with the noterrain.dds) but I don't think it will do much good, as the trees are too fragile. I cannot find any way to even export and reimport them with either blender or nifskope. Aside from editing them vertice by vertice inside NifSkope, I'm out of ideas. I'll think on that some more.

Two features to a plot is a no-go as far as I can see. Quite simply even if we added it to CvPlot, the method which would create the draw-list (RebuildTileArt) is a method provided to the SDK, not defined within it (it's in CvEngine), and can't be modified (well, probably not without breaking the EULA) and thus this is out of the scope of the SDK. Maybe someone who's spent more time looking at the code can confirm this.
___

This is useless for any practical purpose as far as I can see, but a transparent terrain texture gives the appearance of a smooth join...



....but all the ingame models are floating in midair and when you rotate the camera is looks awful. No advantage of using a feature instead of a terrain (in this example).

Anyway, I'm not giving up by any means, although I'm busy for the rest of the week on a training course so I'll have to put everything on hold until then.