Merricksdad's Black Hills Tileset (First Look)

[Legacy_3RavensMore]

               

Is there a road or perhaps a trail planned for this mountain set? 

               
               

               
            

[Legacy_MerricksDad]

               

I want to do a road on the 0 and +1 terrain. I want to do a dirt trail on 0 to +2. Roads will only be allowed on flat tiles (0000) or tiles that rise two sequential corners by +1 (so 0011, 0111). I will also do a road split for three adjacent corner rises. I will do similar for trails going higher (so 0000, 0011, 0111, 0022, 0222).

 

For stream, I am playing with three ideas:

 

The first includes adding a stream only on tiles between 0 and +1 corners, with waterfalls from higher elevations also required to pass between those corner elevations. (0/0, 0/1, 1/2, 2/3, 3/4) This limits the stream tiles required to a reasonable level, but gives streams an ability to have two uneven banks.

 

The second includes having streams located on tile edges, reversing the meaning of crosser placement for streams. For instance, a stream crosser at the top of the tile would place a stream along the top, rather than from top to middle. While this has some drawbacks, it also allows me to place streams in the exact edges where streams would most likely be in a mountain setting, and that is at the lowest point between two higher places. This also means edge streams will only look good on edges which are the lowest two corners on a tile. It may create some confusion when three or four corners are equally the lowest, but I am trying to work out a feasible system for this reversed crosser activity, mostly because I think it can lead to a new building method.

 

Like the second idea, the third idea places streams not via normal crossers, but instead places them as a terrain, similar to shallows. Because streams need to be on the lowest section of a tile, they would only be available on level 0, just like shallow water is. Since any corner can technically be 0 for whatever height plane it is on, that allows streams to be placed anywhere that is not a upward pointing corner. Streams would then be connected via the tile diagonals, or along tile edges. Streams on edges disallows streams to turn a corner in a smooth manner, unless streams are made to turn smoothly over diagonals instead. The tiles required would be reduced if no streams could make 90 degree turns without using a diagonal tile. Having streams on corners also makes it super easy to link streams with water tiles.

 

In addition to the above, I want to include at least one of the wall types I had planned originally. I might just go with an outcrop of a different type of rock for now, at a slant to the ground plane. That way you can use it as a hogback pointing uphill, or an outcrop pointing away from the mountain center, as you would see with many faults. Like roads and paths, I would only include walls for flat areas, or small raised areas. However, a wall-type crosser could also be used to make the jagged spires I wanted for higher areas.

               
               

               
            

[Legacy_MerricksDad]

               

Today I am working on the 2x tile variety. I'm also working on a better and faster method of turning edges to reduce downward facing polygons. Using rayintersect, I can now determine if any edge is hovering over another part of the terrain, and if so, flip it, in a loop checking again and again, until all edges that were naughty have snuffed it. One thing about this new method is that it leaves more face detail, because it prefers to flip edges to erase downward facing polygons, rather than collapsing faces.

 

The second set of tiles, similar to the first, takes the center-most vert and the connected faces out two face selection expansions. Then I flatten them and raise them to the bounding box height of the tile. This creates the diagonal bridges I needed to complete the diagonal variety I usually offer, and gives a walkable outcrop or cul du sac to tiles like 0001, while making more raised walkable area on tiles like 0111. Since the raised center matches the highest level of any given corner, this means the outcrop is going to be as high as it can, especially after noise is added. This also leaves room for more tiles to be created in the future, if I choose to make alternate tiles with the raised center equal to one of the other two heights on a 3 or 4 height tile. This leaves room for creating walkable ledges, or the ability to smooth a tile diagonally, allowing natural looking transition up 4 levels at once.

 

That also leads me to a reasonable walkable angle for this tileset, in the event that I do confine the walkmesh. If I assume that I need to walk up at least +2 elevation changes, then 1200/1000*45 degrees = 54 degrees. So I will be testing some walkmesh confinement on any face elevated over 54 degrees on an edge. Half of that being 27 might show me the face list for grass placement.

               
               

               
            

[Legacy_MerricksDad]

               

I am not entirely sure (waiting on export) but I think this new drawing method makes things softer looking in most places. It might also increase the walkmesh complexity, which is a bad thing. One thing is sure, and that is that it is faster, has less bugs, and doesn't stutter or stammer on places where it cannot flip an edge to get rid of overhang.

 

It has also wiped out the tile edge issues with center-line verts being removed due to heavier optimize calls. This should reduce the seam visibility to near zero. Strangely, that also increased the mesh complexity of more flat tiles, which I hope gets rid of some of the occasional odd shadowing in their centers.

 

I'll report back with pics once I get to actually see the set.

               
               

               
            

[Legacy_T0r0]

               

Haven't had a chance to DL yet but this is an instant masterpiece bro !

And the fact that it is all script based is incredible. What I'm really hoping for is to somehow put all those scripts (shadow doctor, pathnode finder, etc) to use and improve older tilesets. Will that be possible ?

               
               

               
            

[Legacy_MerricksDad]

               

It could be some time. Right now I am fine tuning them for my own needs, but I think they will work ok on other tiles. Shadow doctor should work on any type of tile, no matter what. The pathnode finder works only on tiles named like mine and expects similar walkthrough patterns. The rest of the stuff might work on other tiles if they are not concave on the x/y plane, but not sure. Out of 1200 tiles, only one tile failed to optimize with the new edge optimizer, and it tries multiple times the first run. At least it flags those ones red it cannot do. After two more tries, it got a noise pattern it could work with and was able to complete.

 

So I'm having some trouble today. I keep running Gmax out of memory, or trying to do batch work on 2000 or more objects. Gmax is unhappy with me, so the simplest things have taken hours to complete for a batch of meshes. I was able to export about 20 tiles earlier, and found I had made a mistake in the shadow mesh, simply because I had forgotten to add a trimesh modifier and turn off render. Now all hell has broken loose because I need to manage 10000+ objects and apply that modifier, and it just doesn't want to do it in a batch, or individually by script. I also found that somewhere along the lines today, the texture on the floor and wall faces got smudged, so I need to fix that.

 

What a pain.

               
               

               
            

[Legacy_Empyre65]

               

I wonder if you could do multiple Gmax batches using and old-fashioned DOS-style batch (.bat) file. To do this, use start /WAIT and then your Gmax command, or else it might try to run them all at the same time. It would look something like this:

 

cd \folder\path

start /WAIT gmax.exe some parameters here

start /WAIT gmax.exe parameters for the second batch

etc.

 

               
               

               
            

[Legacy_CaveGnome]

               

<cut>So I'm having some trouble today. I keep running Gmax out of memory, or trying to do batch work on 2000 or more objects. Gmax is unhappy with me, so the simplest things have taken hours to complete for a batch of meshes. <cut>

What a pain.

 

Have you tried to expand your system virtual memory ? Allocating more hard disk drive space in the paging file can alleviate heavy program use (on XP or Win7: system, advanced tab, under performance & settings, virtual memory. Double the max limit for ex.). Another way to boost program perf is to allocate this virtual memory file only on your fastest drive (system create it by default on C:, but that's not always the best unit, specially if you own an SSD).

               
               

               
            

[Legacy_MerricksDad]

               

This morning after making a few changes last night, the current file is not able to be lit properly. There are over 16000 objects in the scene at the moment. What I'm going to do is move the new half into a separate file and go from there. On the previous computer, there was no way I could run half that many elements in a scene and still be able to rotate my view. And on the machine before that, just having 200 trees in a scene extracted from OC Mdl files was enough to crash the viewport.

 

Hopefully this fixes it, or I will have to revert to an older file and see if something else is wrong with it. I know one of my scripts made a pile of meshes with no verts, and applying certain modifiers to that crashes gmax without notice. I cleaned those up and got some access back last night.

 

While I cannot select all the meshes I want to move to another file by name, because there is almost 600 per subset, with up to 14 child meshes, I did make the good move of separating the new set and old set by 15000cm. I'm running a delete loop on everything over 15000, and then will reverse the function on the previous file to separate the other half. That should put me back where It was yesterday morning.

 

Edit:

Wow, that worked. Now something that was taking hours is now taking 30 seconds.

 

Exporting...

               
               

               
            

[Legacy_MerricksDad]

               

Export of the new originals shows that their polycount on the visible faces is a bit lower. I can see more faces in the sun. But, something about it is multiple times faster to export. Instead of 45 minutes, this only took 7.

 

 

I'm exporting the alternate tiles right now, and then will update the SET file with the alts. It should look more varied with 2x variety, and some of the areas will be 'fatter' and have more walking area which is flatter. This can also make the ravine shapes more tight feeling.

 

I'm also coming up with some texture options. Something more white on the rock, and something less moss and lichen covered for the ground might look more sharp and bright.

               
               

               
            

[Legacy_MerricksDad]

               

Nobody caught that the SET file in the download was mangled? Lol! I had apparently changed the set name in the old set editor and it reset all the tile offsets to 0. This new one I am uploading right now will fix it.

 

Ok, so this next version has an updated SET file which includes 2 tiles for every possible combo. Some of the tiles are really similar, but not entirely duplicates, but most of the tiles represent an "innie and outtie" variety. The new outtie files are all suffixed by _2, and sometimes have taller knobs at high places. They also have more reasonably walkable area on knobs and at wider flat sections.

 

The walk for each is a bit less complex than the last release, but not by much. Shadows are cast with this mesh shape.

 

Visual mesh is about 1/3 less detailed from what I can see. This makes export/import and loading a little faster. The visual mesh is also less collapsed, and more rounded due to picking more edge turning and less face collapse. I'm not sure which one I like better, but this version has less errors and far less manual cleanup after running the script, so that makes it my go-to choice. In the previous version of the edge turning script, it often left 1 in 10 tiles processed with a missing or flipped face, and those were ALWAYS along the tile seam.

 

Speaking of seams, the previous version's tile seams were often off by one vertex missing per edge in about 1/2 of the tiles. That was a lot of seams. This version should have exactly the same vert count and position per tile edge from tile to tile. If you do spot a seam, let me know. One tile to watch is 0001. Certain tests are reporting it may have two ridiculous fraction verts which may hang off the edge. I don't see it yet so I'm letting it ride.

 

I set the opacity on the water texture to 50% but it didn't seem to take, or may be related to the TXI animation. I'll play with that later.

               
               

               
            

[Legacy_OldTimeRadio]

               

Strange.  I can't login to your test module.  You know what's up?  I went ahead and made a new character and even then I couldn't login.

 

?

 

Edit: I guess I'm having the same problem that 3RavensMore is having.  (puzzled face)

 

Edit-2: I can make a new module with the haks and it works but you might want to genericize that test module so people can quickly jump in and check it out.

               
               

               
            

[Legacy_kalbaern]

               

I've just been adding the hak to a test module of my own and it seems to work fine.

 

I really like how it looks so far. I'd like to see the grass option actually have grass though and was hoping to see some of those hoodoos shown way back in the first pages here. '>

               
               

               
            

[Legacy_OldTimeRadio]

               

Feedback: I have tried to do a lot of things through automation that I eventually realized couldn't fully be done in an automated way.  Or, put another way, I could get like 80% of the way there with automation and the remaining 20% had to be done by hand.  Which led me into sort of an 80/20 "rule" paradox: Getting 80% done only took 20% of the time.  Getting the remaining 20% took 80% of the time.  After making a scratch area (looks like this) and turning the shadows on and off in the toolset I was seeing very few shadows.  But what really stood out were the smoothing discrepancies.  I understand that given the overall nature of the material itself, there are going to be "smoothing discrepancies" because it's granite and it cleaves in such and such a way and it's just going to look like that.  But the discrepancy between tiles (if you look on the right side of that pic) is really pronounced.  To take a sample tile: tmd07_0011_2.  It looks kind of wonky on the slope and the tile seams are really visible.  The part of the tile near the top of the picture looks freaking fantastic but the bottom part and the edge seams maybe deviate away from looking like granite and more just like a low poly model with not a lot of smoothing applied.  From reading upthread it looks like you might be doing an autosmooth set to 10 (?) over the whole tile. 

 

I'd just kick out two pieces of feedback which I admit are of dubious utility because I'm no tile expert:

First, would it be possible to apply a higher smoothing to faces closer to the edge and less in the middle 3/4 (or 7/8?) of the tile to make the issue I describe above "go away" while still retaining the automated nature of the process?

 

Second, I kept ticking shadows on and off and saw very few shadows appearing- it could obviously have been my choice of tiles.  However, there was a self shadowing (I think that's what it was) on one tile and another which was close to a body of water was throwing a weird shadow fragment on it.  If you had 10 "zots" to throw at the tileset between shading and shadows, my vote would be more zots in the smoothing and getting the tiles to look "just so" and not spend so many zots on the shadows, which are (I'm assuming) way more time consuming per fix.  You had sort of floated a question earlier about how many of us used shadows.  For me, if something comes with shadows I'll probably play around with them on.  But if they aren't included, no big deal.  I just don't miss them that much.  But what really stands out is the smoothing, because that's always on unless the ambient sun/moon is set so light that it blows it away.

               
               

               
            

[Legacy_OldTimeRadio]

               

Amendment to first suggestion: Just checked out the actual tile in Max.  If tmd07_0011_2 is representative of the rest of the tiles, where the visible mesh is broken into several different meshes by material, maybe just apply a different smoothing formula to those portions of the ground mesh which are supposed to be rockier versus ones which are less rocky?  I.e., if it's got the rock texture, smooth it at 10 (or whatever) and if's got more of a ground texture, smooth it at something higher?

 
Current and with non-rock "ground" in one smoothing group.  One smoothing group might still be overkill for the texture that's being used though (raises eyebrow) a fat grassy texture in contrast with that granite could give it a look similar to the hills in the Giant's Causeway in Ireland.