Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - sfriedberg

Pages: [1] 2 3 4
1
Development News / Re: Dev Silent in Forum and Blog
« on: May 31, 2017, 12:25:17 PM »
jgwinner, did you actually try to download the update?  Remnant extended the update eligibility substantially due to the last year's non-activity, so if you had a license you might get the download.

https://update.world-machine.com/upgrade.php

2
Macros and Plugins / Re: Geoglyph GeoFlow
« on: April 20, 2017, 02:24:24 PM »
byGamersYT, Geoglyph is a paid product from www.quadspinner.com, not a free download.

3
Guest Forum / Re: Facebook Page Closed Group?
« on: April 20, 2017, 02:21:31 PM »
Just a comment.  There are still Neolithic oldtimers out there, like myself, who do not belong to Facebook, and never will.

While I completely understand (and approve of) restricting comments to known group members, hiding the forum content from the public seems unproductive.

4
Plugin Development Forum / Re: Any plan to move off VS2010 Express?
« on: April 18, 2017, 02:50:34 PM »
So it's been three years since I started this thread, and Microsoft has (finally) nuked download access to VS2010 Express.  This adds a certain urgency to the question which was missing 3 years ago.

MS are heavily promoting VS "Community" 2017 as the replacement for VS Express.  It is not clear if VS2015 Express is actually still available.  I found a few links for VS2015 Express on MS' website, but all the ones I tried redirected to VS Community 2017.

5
Those will be difficult in WM.  In the real world they were shaped primarily by glacial erosion and secondarily by fracture (spalling/exfoliation) of exposed rock surfaces.  WM doesn't have a node to do that.

However, if I were trying this, I would start with an Advanced Perlin Noise node using the Sharp Ridged mode and a substantial Steepness.  I'd start with the default Octaves and then experiment with the Persistence setting.  Once you get something approximately to your liking, there are many, many options for fine tuning in the Advanced Perline Noise node.

6
I read that to create different regions/biomes/terrain types, I had to have a landscape generator for each area and draw my land out making sure the regions meet correctly.
Not essential.  It would be perfectly reasonable to have a mask for each area, which could overlap in a sloppy way, and use WM blending to merge the areas together.  Think splat maps, but instead of micro-terrain coloring, macro-terrain WM graphs.
Quote
I take it to do what you are suggesting about the coast to land transition, I use a gradient as a mask?
Yes, but by "modest amount inland" I intended a gradient just around the inside of the mask.  If you use a typical radial gradient, your terrain will hump up in the center  with an obviously conical shape and it won't follow the "coastline".  So a feather or similar boundary-respecting operation would be good.  I don't use Gimp, so can't advise you on the best tool to use in it.  In CorelDRAW (vector), you'd draw the mask outline, offset it to the inside, fill the outer outline with black, fill the inner outline with white, and create a blend between the inner and outer outlines, which auto-interpolates the fill color for each step in the blend giving a gradient effect.  In Corel PhotoPaint (raster), you'd take an inverted copy of your mask, feather the copy to the inside of the area, and fill (masked) with black over the original black/white mask.  Or you could heavily blur a copy of the original mask, invert it, and fill (masked) with black over the original black/white mask.

BTW, you will get a more realistic look if the depth of "modest amount inland" varies substantially as you move around the perimeter of the area.   Long, shallow gradients will slowly build from the waterline to the full WM terrain amplitude, while short, steep gradients will ramp up quickly like coastal ranges of hills (or cliffs in the extreme case).  My advise above was just on getting you to a useful starting point.

7
Let me suggest a different starting point.  Use your mask as a mask.  Don't trace the outline with a layout generator.  Just look at masking a noise as your first step.

You should see something that fairly accurately follows your mask perimeter, but with that nasty height jump at the "coastline" you mentioned.  There are many ways to deal with that.  One is to go back to your drawing program, and modify your mask with a blur or feather or whatever your program provides such that the "coastline" is black and there is a grey scale ramping up to white over a modest distance "inland".  Then apply the modified mask to a noise and you should see the "coastline" at altitude zero, ramping up to the full noisy amplitude "inland".

That should work as a starting point for you, with much less labor and more reproducible results.

Once you get more comfortable with WM nodes, you should be able to do the coastline-to-inland ramp modification in WM itself, without having to modify the black/white mask in your external drawing program.

8
Bueller?

TorQueMoD, what node are you viewing?

9
Bug Reports and Help / Re: Need help working with a heightmap
« on: March 28, 2017, 07:07:10 PM »
The greyscale portion of the image (the landmass) looks like a pretty usable heightmap to me.

To get rid of the blue areas, I would use a FileInput node in bitmap mode followed by a ChannelSplitter to select either the red or green channel, and use the resulting channel as the heightmap.  Set an appropriate altitude range on your project to avoid "flat" in the 3D viewport.  To make it look a bit more like terrain, or at least a map, add a Colorize node.  To make it more realistic, you will probably want to insert a fine-grain (aka high-frequency, low amplitude) AddNoise  or PerlinNoise node before the colorizing, and then you can start applying tutorial lessons about Erosion and so forth.

10
Are you viewing the output of the curves node, or the output of the overall flow?  Select the final erosion node and look at it in the 3D preview.

11
General Discussion / Re: World Machine facebook group
« on: February 20, 2017, 04:50:00 PM »
Same here.  Not joining Facebook for the sake of one application discussion group.

Whenever I hit one of these "social" sites where I have to create an account to access nominally public postings, Facebook, Pinterest, etc, I just close the page and walk away.

12
Feature Requests / Re: Allow simple math in number fields
« on: May 19, 2016, 10:01:30 AM »
He means in the field itself.  So instead of being limited to numeric literals, fields would accept simple arithmetic expressions, evaluate them and use the result as if you had entered that literal value.  Not talking about "live" expressions.

13
Feature Requests / Re: Sense of Scale
« on: May 10, 2016, 02:00:36 PM »
A couple of excellent approaches were discussed back in 2013 in this thread.  The one I especially like is overlaying a tiling nested-square texture where you know how much "model space" each size of texture square takes up.  See ivanhawkes's post in that thread.

14
Geology and Terrain / Re: Ideas on making caves
« on: March 14, 2016, 12:52:11 PM »
I've never attempted this, but I think I'd start with a layout generator to define the "plan view" of the caves.  Without some way to force both the floor and roof to meet consistently, you'd have an awful patch job to do.  So given a definite outline, you'd want both floor/roof to blend to a fixed height at the outline.  Further from the outline, more of the "raw" floor/roof graphs gets used.  I'd probably start with a blend to 100% roof at whatever WM scale corresponds to 10 real-work (or game-world) feet, then add a bit of noise to that scale distance.

For the floor and roof, besides general rockiness, you can obviously toss in stalactite/stalagmite spikes.  If you want a Hollywood cave, keep the floor and ceiling amplitudes fairly small (so things are flattish).  For more realism, I'd add in some substantial amplitude low-frequency noise, so the top and bottom (can) come close enough together to force a crawl or at least a crouch in places.

Getting back to the fixed height.  You may get a more realistic look by letting that "fixed" height be some low-frequency noise with amplitude a healthy fraction of the cave interior height.  That way the floor/roof seam isn't so obviously horizontal.

15
Bug Reports and Help / Re: Undo is always greyed out
« on: February 23, 2016, 03:40:32 AM »
The Dev version (River) has undo support, although I haven't exercised it very much.  If we ever see Remnant again  :(  I'm sure the feature will migrate eventually into the main line version of WM.

16
Geology and Terrain / Re: Realistic Mountain Ranges (Very small scale)
« on: February 08, 2016, 03:44:59 PM »
That real-world example is defined by heavy erosion of a relatively featureless mountain uplift (orogeny).  You may or may not be able to reproduce the effect in WM directly at the final scale you want.  You may have to do your WM modeling at a larger scale for the WM erosion node to give comparable results, then scale the results down afterward.

17
WM cannot create overhangs, so you can't reproduce the areas where the slate sticks out over the sand.  That's inherent to the heightmap format.

For the slate, the terrace node is the place to start.  The slate looks like a cap layer over some other softer rock, so you might limit/mask the terrace effect to just the higher elevations.  After getting something that looked reasonable, I think I'd throw in a "spidery" or "ridge-y" noise to break up the masses of slate.

For the sand, you have a couple of approaches to try.  You could crank up erosion really high and use the sediment deposits as the sand.  Or you could use a separate low frequency terrain node and blend/mask that in to the rockier terrain.  Take the max of the sand and rock terrains, maybe.  Might identify the slate areas first, then taper the sand amplitude down to zero as it approaches slate (via masking and feathering/making-gradual the edges of the masks).

I think I would apply the terrace node after the sand/rock blending, but that might that's just an initial guess at what would work.

For the rubble, add in a very high frequency, low-and-clamped amplitude noise in selected regions.  In that photo, those regions would be elevations just under the bottom of the slate threshold.

That's all off the top of my head.  Be very interested to see what you come up with!

18
General Discussion / Re: Using World Machine to create rock topology?
« on: February 02, 2016, 05:55:15 PM »
Two different uses of the word "tile".

A World Machine tile is a square chunk of terrain, where the overall terrain computation is cut into tiles.

A tiling texture is one that repeats.  People often mean seamless tiling, where adjacent "stamps" of the texture merge smoothly into each other.

World Machine can export a large terrain as a set of smaller adjacent tiles.

World Machine does not compute heightmap patterns that tile seamlessly.

19
General Discussion / Re: Multiple splatmaps possible?
« on: January 20, 2016, 04:28:32 PM »
Again, I know nothing about Unity, but a little Googling suggests that when you add more than 4 terrains to a material (in Unity), Unity automatically creates another control texture (i.e. splatmap slot).  Presumably, you can then load a 2nd 4-channel splatmap generated by WM into that 2nd slot.

Here were a few links I found that touch on the subject:
http://forum.unity3d.com/threads/splat-map-shader-with-more-then-4-textues.267268/
http://forum.unity3d.com/threads/custom-splat-map-terrain-with-more-then-4-textures.266990/
http://answers.unity3d.com/questions/831621/splatmap-limits.html
http://answers.unity3d.com/questions/138903/splatmap-performance-question.html
http://www.terraincomposer.com/how-to-faq/
http://www.bitsalive.com/how-to-create-awesome-terrains-in-unity/

20
Guest Forum / Re: level river bed
« on: January 19, 2016, 10:37:12 AM »
That would be a hasty assumption.

Try copying your layout generator, shrinking it in width, using it as a mask for a clamp operation.

Pages: [1] 2 3 4