649 réponses à ce sujet

[3RavensMore]

It works now.  I've spend that last two hours building with this set.  I'm just floored with it so far.  As it is without any other additions it's already so very useful.  The feel of scrambling around on mountain slopes, the effect of reaching a pass and looking down into a hidden lake in a narrow canyon...it's there. 

[MerricksDad]

Exporting batch 3 of alt tiles right now. They are a lot smoother, and the raised point in the center is 50% more confined to center than before, and slightly lower than before. This makes the difference between the primary and alt tiles slightly less, but still there is variation.

[3RavensMore]

Pretty

 

 

 

Are you channeling PHoD now from his India placeables? 

[MerricksDad]

It works now.  I've spend that last two hours building with this set.  I'm just floored with it so far.  As it is without any other additions it's already so very useful.  The feel of scrambling around on mountain slopes, the effect of reaching a pass and looking down into a hidden lake in a narrow canyon...it's there. 

My wife and child love it as well. I can only imagine what it will look like with decals, rocks and trees. Or with a nice shallow stream heading down a ravine with some stepping stones for it to pass under.

[MerricksDad]

I'm an idiot. I forgot to apply textures of any kind this export. Starting over...

[3RavensMore]

My wife and child love it as well. I can only imagine what it will look like with decals, rocks and trees. Or with a nice shallow stream heading down a ravine with some stepping stones for it to pass under.

 

I... <blinks wide stunned eyes>  am speechless. 

[MerricksDad]

much better variety coming out shortly. I took some of OTR's advice about differing smoothing groups on floor and wall meshes. With wall mesh, I stayed with 10, with floor I went with 20. Because the floor is physically smoother than before, the higher number works better now, but still it magnifies the tile seam while reducing on-tile seams. Not sure if I like the change, but it could be a move in the right direction.

 

As I mentioned, the floor tile is now more physically smooth, which allows more upward facing sections on the tile. Some of the high outcrops over water now look amazing.

 

I didn't change the appearance of primary tiles since the last release, so I am now interested to see if these play well together, or if the primary tiles need to be relaxed and smoothed equally to make a good connection.

[MerricksDad]

Looks like I am all signed back up for classes January 11th, so I need to push hard on this set and get it done. I'll be unofficially taking calc 1 and 2 again in my spare time, so I highly doubt I will have any time for NWN in the spring. I don't expect to take classes for the fall/winter terms, which are actually summer/fall, just because I will need to watch my kid during summer, and probably half of their fall term. So like this year, I will be doing most of my model making during that time, after a possible vacation to some new visually exciting place.

[MerricksDad]

Need to fix some bugs. Two tiles share an exact duplicate of some weird duck beak shape. I'm not even sure how that shape got made. Just weird. Also tile 0001 now has some kind of seam offset again. Why that one tile, and why ever time I do something? Also weird.

 

Otherwise, it looks good, but the primary tiles could use some smoothing group changes to fit better with the alt tiles now. I think by tonight I can easily repack a new release. I'm still tweaking, so I am going to hold off on full quality shadows again. It just takes too long to export, and makes GMAX angry. If I keep it up, I may need to get a cooling platform for the laptop.

 

Many of the tiles also need some kind of texture correction. I had specifically built that into the script, but it doesn't seem to have worked. Instead many of the textures are stretched a bit, making some of the taller hills look like play-doh.

[Tarot Redhand]

@MD, I came across this on deviantArt and thought of you and this tileset.

 

PS. What do you mean by decals? The only decals I know for sure, what they are, are those included with plastic model kits (miss-spent youth - don't ask). In the UK they are known as transfers ('cos they transfer onto the model after soaking in water).

 

TR

[MerricksDad]

I mean exactly that Check out the various splotch textures from OC rural. I'm going to do that with mosses, lichens, and decayed feldspar pegmatite. I might even make a decal for the bottom-most layer of a creeping juniper.

 

Nice pile of dirt there in the pic

 

I was just looking through my pictures from this latest trip and I found some photos from downhill of Breezy Point picnic area with bright orange lichens like pumpkin.

[MerricksDad]

PAGE_FAULT_IN_NON_PAGED_AREA

 

This application is corrupt and cannot continue

 

LOL! Whatever McAfee. Go away

 

Restart...

 

 

Ok so, I got a new texture installed on half the tiles for the vertical regions. It looks much better and smooths to other faces better without cutting as much. I haven't updated the floor texture yet, other than to adjust placement as was needed.

 

I DO want to reprocess the original tiles because they no longer work as nicely together with the new ones. I'd like to smooth and relax them the same as these new alts, and then apply this new vertical face texture.

 

If I get a chance this weekend, now that all my pre-education stuff is out of the way, I'd like to work seriously on either the vegetation version, or the streams. I must pick, but how...so hard to decide.

[OldTimeRadio]

As you have determined, I pretty much don't want the seams to go away on the rocky section, but I don't mind if the other texture is more smooth, I just need to do it in a different way. As you saw yourself when you smoothed the floor mesh, it looks more smooth on that tile, but adjacent to another tile it will look even more sharp-lined at the seam. Not much more because it was pretty bad to begin with on the '_2 tiles, but still more sharp tile to tile. I really need to fine tune those alt tiles more like the way the previous release was. It still had visible seams, but it was much harder to tell which seam went with which tile.

 

Gotcha and thanks for taking the time to type that explanation!  Things like here and here are just cases of not having a rim of coplanar faces which are shared between tiles, right?  The shadow fragment I described was just a b.s. toolset artifact.  I should put a dollar in the noob jar for even bringing it up.  If I come across the self-shadowing issue with your new tiles, I'll let you know which one.

[MerricksDad]

Gotcha and thanks for taking the time to type that explanation!  Things like here and here are just cases of not having a rim of coplanar faces which are shared between tiles, right?  The shadow fragment I described was just a b.s. toolset artifact.  I should put a dollar in the noob jar for even bringing it up.  If I come across the self-shadowing issue with your new tiles, I'll let you know which one.

 

yup, pretty much.

 

Also, anything the toolset can bring up, I'm convinced the engine can do the same. Something about some face somewhere is wonky enough to make it come alive.

 

As I work on relaxing the mesh, I am seeing less and less flickering shadow faces, but with the physical difference between the shadow casting mesh and the visible mesh, I'm not getting anywhere on that front yet. If I get some time this weekend, I'll do a full shadow render and let everybody see the differences. That is if I can keep GMAX from suffering a stroke.

[MerricksDad]

Version 3 is available for download. Tiles are all softer physically, and softer visually, both in smoothing, and in texture. I also updated the testing module to accommodate slightly higher base level due to smoothing from corners, in that I raised the triggers for area to area teleport by 0.2m. I personally prefer this new texture on the wall faces, but let me know what you think. I left the other one in the hak file so you can dig them out into override if you like and reverse their names (delete the texture hak from the module if you do, so that it can see overrides).

[Zwerkules]

If I get a chance this weekend, now that all my pre-education stuff is out of the way, I'd like to work seriously on either the vegetation version, or the streams. I must pick, but how...so hard to decide.

Vegetation (if it is not a decal) can also be added as placeables while streams can't.

[MerricksDad]

very true

 

I started work on the walkmesh configuration script. It automatically locates proper locations for walkmesh grass and what should be water. I need to go back and redo one of my getFaces.. scripts so it returns bitarrays instead of arrays, then I can move forward again.

 

But first, we're taking the boy to a Halloween store to play.

[MerricksDad]

It grew grass

 

 

If you play this with the grass on, I suggest upping your grass graphics a bit from vanilla. Go in your nwn.ini and find [Video Options]. Under that find "Grass Far Render Distance=" or add it, and set the value to at least 5000. I tried it higher but it gets a bit laggy at 10000 on mine. I think this value is in centimeters rather than meters.

 

Packing and uploading grassed version now. You'll only need wok + dat updated. I've not repacked the "Full Series" file yet until I make sure more stuff is working ok.

 

Edit: I will need to repack the vis portion as well. Apparently only AABB data is pulled from the WOK file. The rest of the walkmesh data is still pulled from within the MDL file... sad panda. But that gives me another idea...

 

What if the walkmesh node on the MDL file is open, but the AABB data on the WOK file is closed? I wonder what would happen?

[Zwerkules]

What if the walkmesh node on the MDL file is open, but the AABB data on the WOK file is closed? I wonder what would happen?

What do you mean by the 'node is open'?

The walkmesh information in the mdl files is used in single player games. The wok file is send to the clients by a server and used instead of the walkmesh data in the mdl file. This is to prevent PW players from changing the walkmesh data in mdl files to be able to walk through walls, etc.

[MerricksDad]

Right now the walkmesh has no non-walkable faces, or deepwater faces, so everything is free and open for visuals and for physical movement. If I "close it", you won't be able to walk from any level 0 to 2, 3, 4, or deepwater.

 

I was thinking it might be interesting, since I noticed just by having updated WOK files in my directory with grass and water turned on, it was not enough to render grass or water effects. I had to also update the AABB element in the visual mesh. I was hoping that the actual pathing was derived from the WOK file, and the other details were pulled from the one in the MDL file. If that were the case, I could supply different WOK files than MDL AABB and possibly achieve all-over movement, but have pathing mechanisms override if you only click your location once.

 

The reason I was thinking to try this is because normally I walk around in drive mode, with my pointer always clicked down and I just move the camera with my keyboard, which leaves my mouse pointer in the forward center position. If I need to adjust direction or speed, I change my mouse position in addition to the camera angle. This type of movement overrides walkmesh pathing while the mouse button is down because the location is almost always within the current MDL range, or is so near inside the next MDL coordinates that it doesn't do much pathing. Even on more distant locations, the mouse button held down seems to prevent most path correction from taking place until you let up the button.

 

By combining the intended walking area and offering an open environment fully walkable area at the same time, it might have been possible to do some stuff I would otherwise have to do with pathnode settings in the SET file. I was hoping for alternatives, or to play with some quirks in the engine. I like quirks, as long as they give me at least one beneficial thing to do.

[MerricksDad]

Hey, I don't know if anybody noticed this back when I was doing the gems and gem-like weapons, but if you set an area to interior, even if it exterior, you can get things which are transparent or glowy to look better. I was trying to figure out if that is based on the module-area settings or the tileset SET file settings. I need to play with this some more, because I am trying to figure out some other shadow-related stuff related to interior settings as well.

 

Edit: Oh sure, now I can't find what I was playing with before. Anyway, I had made it so items with self-illumination were more vivid underground or in places defined as interior. I could go up to a cityscape and the items self-illumination were totally turned off. How did I do that.

 

Nevermind. What it appears I was seeing was shadow being cast on self-illuminated objects. Simply turning off sun-shadows fixed this. The appearance of self-illuminated objects is much darker than it should be if the object is actually illuminated. Basically shadow should not affect self-illuminated objects, but it does, and it is applied AFTER the illumination on the object, rather than the object being re-illuminated after shadow. This generally is another hit against environment shadows in my book.

 

Turning off sun/moon shadows in the environment setting panel of the area, I was able to see the difference for this tileset with and without shadows, and I am thinking more and more that it look better without. Furthermore, when you hold a light casting object against environment shadow casting objects, it does not get drawn out like placeable shadow meshes, or characters. This is a very unfortunate, and visible, setup in my book. Just another eye-sore. Perhaps it is best to leave shadows out of tilesets? I was really hoping to find some way to make the tileset objects cast shadows as placeables, but I may have to resort to adding features as placeables to accomplish such things.

[MerricksDad]

OTR pointed out the possible need to clear up some toothy looking parts where the tiles kind of part their hair. If you look at the current tiles, while it is taken down a bit from the first release, you can see at the edge of one terrain texture and the next, there is a visible jagged line. I intended to leave that and then put a hairy shaggy grass texture extrusion over it, as most tilesets for the out of doors tend to do.

 

But this might be a bit too conspicuous, even with such a thing done. So what I am proposing to do is change my tile division script in a way that it examines the number of edges shared by verts on a triangle. If the triangle shares only one vert with any other triangles of the same texture, then it will be transmuted to the other texture. If it shares two verts, it then needs to determine what to do by the number of other faces sharing the same verts. It it determines that the number of shared verts puts it more in touch with the other texture, it will flip, otherwise it will ask the surrounding faces to possibly flip to make a more smooth line across the mesh.

 

Hopefully, this will be sufficient in creating a useful line from which I can extrude (via script of course) that entire unbroken edge. Given a specific angle and width of the extrusion, I should be able to duplicate the grassy shag from the OC rural tileset. It will also shore up some obvious geometry issues in other places.

 

Another thing I would like to do is purposely keep a few of the loose triangles around, so that there are splotches of grass purposely growing right out of the stone faces. If you view a few of them now, you can see there are single triangles, or larger polygons, which the above mentioned script methods might erase. In these cases, it might be more beneficial to detect if they are pretty much floating far away from that intended boundary line. If they are, they can be visually rounded by tessellating the single triangle into 4, selecting the center triangle, and then scaling that triangle in such a way as to make a hexagon from the original triangle, and then relax the whole thing back down to nearly match the underlying walkmesh. That is much closer to a circle, or oval, and will be more accepted by the eye, I think. It also makes the shape conform to the other script idea if it runs twice on the same tile.

 

Just to list what this set still needs, for my own use later:

 

• script to pick decal locations, clone them off, raise them, and apply a nice decal texture
• script to determine which faces are over 27 degrees and which are over 54 degrees, for better walkmesh control
• script to locate hard edges which could benefit from a decal, shrub, or alternate smoothing
• If walkmesh is made to use non-walk surfaces, a script to save a dummy node defining the path-node selection for the SET file.
• script to slice water planes into the walkmesh (optional)
• script to slice water planes into the visible mesh, separate the mesh(es) at that point, and color anything X cm over the water level with a wet stone/sandy grass, while coloring everything below water with either mucky rock, or wet gravel debris
• script to detect faces which use tverts at or near 45 degrees from one of 6 cubic face coordinates. It should then retexture those faces in a 45 degree tilted cube. This ultimately textures the object in the shape of a 26 sided die (look 'em up, they're multipurpose). The benefit over spherical or shrink wrapping a texture is the lack of stretching caused by radian math. The drawbacks are the same: faces textured to something they should not be. I've tried the custom wrapping method in Gmax which seems to use face normals in a spherical grid, and I don't like the unattached feeling of that one. I may need to make my own wrapping method using face normals.

[3RavensMore]

Initial observation of the addition of the grass.  To my eye, the big wavy stalks give the impression that this less a rugged black hills / mountain set, and more of a rural--if a bit hilly--environment.  Before, the eye could be fooled that a canyon was expansive, but the addition of the grass makes everything just seem a lot smaller.  Does anyone else have that impression?  Maybe it's just the grass texture itself, or the scale of it, that looks a little bit out of place to me for this set... 

[MerricksDad]

Without the other plant types on the hill, and the more brown texture for pine needle covered hills on slope 2 and 3, it definitely looks weird to me. But its just an intermediate stage. It reminds me more of the summertime up in the bighorns where you have the combination of tall grasses with tiny tundra things, plus the rocks showing. Right now, you can still grab the v3 without grass if you reinstall the "Full" zip.

 

The plant covered variety is going to be a lot more confined feeling, which is why having both is a plus. Right now the grass covers faces which are up to 45 degrees above a zero plane. When I get this vegetated one done, grass will only be on faces up to 27 degrees over zero. Any slope higher than that will only have odd plants and pine needles under ponderosa pines and spruces.

 

From 28 to 54 I plan to have walkable steep faces, covered at least to 45 with pine needles, with 45 to 54 being rocky debris, probably in slate-looking forms, or boulders taken from the surrounding granite. I'm thinking less flat boulders should be placeables rather than on the tile. Flat ones can be on the tile if you can walk on them.

 

I'm going to leave 0 to 27 completely free of trees, but offer up the same trees from higher slopes as placeables, that way you can personally cover flatter areas with pines, spruces, oaks, and aspens, as your area needs. Areas from 28 to 54 will have trees growing right out of the walkmesh surface. I'll then leave 54 to 90 treeless except for placeables you can jam in cracks, or make L-shaped arms coming out of the rock faces. I plan to offer at least 100 trees combined, if not doing 101 trees for each species. Variety is good, and script-built variety is super simple. Offering that many in a single package may sound overkill, but allowing people to look through a visual catalog and then pick, by filename, which ones they really want to use in their placeables.2da, is a much better goal. So if I do a package of 50-100 each: black hills spruce, scrub white oak, Colorado aspen, and ponderosa pine, that should help me get the variety filled black hills environment I am looking for. I'll do another half that each: beetle-killed pine, and fire-killed pine. A portion of the black hills spruce options will be moss covered, or mossy and mostly dead at the same time.

 

For ground cover placeables, I have these picked out to cover the plain but noisy ground: dirt path, crushed granite like from a mine, tailings piles also from a mine, wildflower collections and singles, sections of stacked limestone and stacked mine chunks for use as walls, but not as an on-tile wall.

 

Instead of putting them on-tile, I've also decided to do some red sandstone and yellow limestone outcrops as placeables which cover 1/3 to 1/2 of the base of a cliff. I'm going to make them wedge shaped, so that the majority of the smaller wedge section is inside of the hill, but if you use it away from a hill, it makes a slanted wall shape. I'll probably end up doing schist and gneiss outcrops the same way, instead of adding visible options for the cliff walls.

 

I've also decided to go with placeable granite spires of various widths. If I do about 25 variety, they can be used up the side of a slope, placed directly out in the open, or buried any distance into the ground and serve multiple purposes.

 

Tomorrow I may start work on the road tiles. They'll be the easiest to start and finish, and will give me an idea of what I may need to change for streams.

 

If the ideas from roads don't pan out for streams, I will start up a test on streams using only tile diagonals. It will be a system of streams much different than people are used to, but I think it can be very interesting looking, and will save a lot of time for the builder to just plot points this way rather than using a crosser. What I am thinking is to allow only any two corners on a tile to have a stream terrain. If only one corner is stream terrain, it knows the stream has only touched on that location, and should not show much stream. If two corners are used, the stream must go from either 0 to 0, or 1 to 0, or maybe even 2 to 0, in relative corner elevations. This substantially cuts the number of tiles which can have streams, and cuts the ridiculous work of making all-things-possible down to a workable collection. I'll then release stream combo group tiles, so you choose exactly where a stream comes together, or flows underground, or comes out of a crack, or whatever, instead of letting the stream be open-ended while drawing, or accidentally draw to just where ever you drag your mouse. My guess is this adds less than 250 more tiles total, and hopefully even less. The script gods will be able to tell me. Below are the combos I'm coming up with so far for trails and streams using this method:

where x is >0:

00xx: two options x-x right and left (11, 12, 13, 14, 21, 22, 23, 24, 31, 32, 33, 34, 41, 42, 43, 44) [=16]

 

where x is >0:

0xx2: one option 2-0 (11, 12, 13, 14, 21, 22, 23, 24, 31, 32, 33, 34, 41, 42, 43, 44) [=16]

 

where x is any number:

0xx1: one option 1-0 (00, 01, 02, 03, 04, 10, 11, 12, 13, 14, 20, 21, 22, 23, 24, 30, 31, 32, 33, 34, 40, 41, 42, 43, 44) [=25]

 

where x is any number:

0xx0: one option 0-0 (00, 01, 02, 03, 04, 10, 11, 12, 13, 14, 20, 21, 22, 23, 24, 30, 31, 32, 33, 34, 40, 41, 42, 43, 44) [=25]

 

82 or less with duplication from previous sets? I can live with that.

[3RavensMore]

Love your plans for the above.  You're even doing beetle-killed pines?  Those damn things went through our trees like a buzz saw.  Now, a horde of giant beetles to kill among the pines may just be therapeutic for their little cousins killing more than a hundred of our trees. 

 

When you do your final release, are you planning an optional "no trees" set?  It would work great for an above tree line set as well.