Classic Erosion plugin

Started by Daniil, November 03, 2017, 05:14:19 AM

Previous topic - Next topic

WAS

Quote from: blinkfrog on June 26, 2018, 01:17:01 PM
Classic Erosion v1.0.6

- Added maps export (this feature was requested several times)
- Fixed bug in the Rivers mode that led to strange points in flowmap and some strange pits

Maps are exported into .png-files. You specify the folder and general name for maps, and they are exported into multiple single-channel grayscale 16-bit pngs (except for streamline map which is exported to 8-bit) - this is a good compromise between size and quality and preserves the dynamic range of maps so you can adjust white and black points or dynamic curve in the graphic editor. See attached images - I used GIMP to modify original flowmap (0.png) to two variants.

As for final filenames - they are generated automatically. For example, if you specify "abc" as filename, it results in several files: "abc.flow.png", "abc.deposition.png", "abc.wear.png" etc.

Daniil

Omg! I'm so excited! Thanks so much Daniil! This sounds awesome!

And this fix for the rivers mode, do you think this would fix the weird issue I was having where rivers weren't actually starting at depression points and seemed random?

Daniil

Quote from: WASasquatch on June 26, 2018, 01:23:28 PM
Omg! I'm so excited! Thanks so much Daniil! This sounds awesome!
And this fix for the rivers mode, do you think this would fix the weird issue I was having where rivers weren't actually starting at depression points and seemed random?
I am afraid this won't fix this issue. Rivers headwaters are placed RANDOMLY indeed, not in depression points. Headwaters are static sources of water, such as underground springs that can be located in any places.

Thank you for the website page.

As for the problem with "cut" and "paste" - I tried it several times but had no any problems.

Daniil

bobbystahr

Thanks for the update Daniil
something borrowed,
something Blue.
Ring out the Old.
Bring in the New
Bobby Stahr, Paracosmologist

Daniil

Quote from: Mid-Knight Acchan on June 14, 2018, 03:08:11 AM
https://www65.atwiki.jp/terragen/pages/144.html
This site is made without permission.
I am glad if you give me permission.
And.. Japanese only... :P
Use google!

Hi again,

So, I've reviewed this page (thanks to WASasquatch), and all seems ok (if you asked Ulco and Martin for permission). Great work BTW!

Daniil


Dune

@ Mid-Knight Acchan: I can't remember if I ever gave this permission, and any query of yours may have ended up in the spam box, but I herewith give you permission to use the images. No problem at all.

Mid-Knight Acchan

WASasquatch, Thank you for making an offline page.
Daniil, Thank you for remembering.
Ulco, I sent an email, but it does not seem to have arrived, thanks for permission.

BTW,
Under what conditions will the "Water depth map" be output in map export?
Although it is touched, only four maps are output.
I started 3D landscape with "Bryce" and am currently editing Japanese Wiki as a Terragen user. I am riding the Kawasaki ZEPHYR1100. I am a reader.

Daniil

Quote from: Mid-Knight Acchan on June 27, 2018, 04:34:47 AM
Under what conditions will the "Water depth map" be output in map export?
Although it is touched, only four maps are output.
Water depth map can't be exported currently. I need to think through some things:

1. Water depth is expressed in meters, how to convert it to 0...1 range? I can normalize it, but will it be useful after this?

2. Water depth is the only map that isn't raster map. It has infinite resolution and shows actual difference in meters between water surface and river bed. I can sample it to raster, but will it still be useful in this case?

Daniil

mhaze

Thanks for the maps it will be a really useful update. All I need is a map of the river positions and depths can be shown by grey scale so a simple raster map will be a very good start.

SILENCER

#114
I've never seen this before, and it shows up in the 3D preview as well as RTP.
At first I thought it was some bizzarro access to the hot pink default test color, but no.
Then I thought maybe some error if the flowmap masked surface is displaced, but Nyet.

Lowrez heightmap, maybe? It's only at 4096. Scene is reading a saved .erd file

I'll try this on my work machine which works at Formula 1 speeds, and see if the height map rez is the issue.


EDIT: I see above the Plug In was updated with a new build. We'll give that a stab and recalculate.

WAS

Quote from: SILENCER on July 04, 2018, 10:28:46 AM
I've never seen this before, and it shows up in the 3D preview as well as RTP.
At first I thought it was some bizzarro access to the hot pink default test color, but no.
Then I thought maybe some error if the flowmap masked surface is displaced, but Nyet.

Lowrez heightmap, maybe? It's only at 4096. Scene is reading a saved .erd file

I'll try this on my work machine which works at Formula 1 speeds, and see if the height map rez is the issue.


EDIT: I see above the Plug In was updated with a new build. We'll give that a stab and recalculate.

Looks like the maps are showing through your shaders from terrrain. Make sure you have at least a solid colour surface layer or PF before any actual colouring and masking.

SILENCER

Make sure you have at least a solid colour surface layer or PF before any actual colouring and masking.


They way I solved this was by piping the flowmap scalar data mask into the PF color function of the surface, instead of the actual surface.  I also re calculated the hieghtmap at 8K resolution.

Daniil

Quote from: SILENCER on July 04, 2018, 10:28:46 AM
I've never seen this before, and it shows up in the 3D preview as well as RTP.
At first I thought it was some bizzarro access to the hot pink default test color, but no.
Hi Kyle,
Do you mean these strange color artifacts? I think they are because of flow map - this map can have values higher than 1.0. Clamp flow map (you can insert Color adjust shader with default settings) and it should solve this problem.

Daniil

Daniil

#118
Classic Erosion v1.1

- Fixed behaviour that caused strange peaks and overshoots om the area of sharp terrain edges;
- New thermal erosion.

[attachimg=1 align=center width=800]
(There is only thermal erosion is used)

New thermal erosion is presented as two new thermal erosion models. There is the parameter "Thermal erosion model" where you can select desired model.

Model A: this is old thermal erosion, left for compatibility with existing projects.

Model B: new erosion model. This thermal erosion model applies thermal stress uniformly to the rock. Resulting rock has similar look to uneroded, keeping it's shape as long as possible. Of course, if high thermal erosion intensity or long simulation time is used, resulting rock can lost it's with original. This model can be helpful for imitating erosion that often occurs in a dry climate.

Model C: new erosion model. This model emulates non-uniform thermal stress taking into account rock shape. For example, if you erode perfectly rounded truncated cone, you would get perectly round cone with nice talus "skirt" beneath it. But if original cone is distorted, it's irregularity will drive and control the erosion process, resulting in a complex shape. This gives rock it's distinctive "mountain" look, so often just thermal erosion is enough to get good mountain-like landscape.

Example (top-down view):
[attachimg=2 align=center]

There is also new "Smoothness" parameter available for "C" model. It controls the thermal stress of convex parts of rock: the rock pieces that are raised. The higher is smoothness, the higher is probability of destruction and falling down of these rock pieces. Combining with mild fluvial erosion this can be helpful for getting mountain shapes that often exist in cold climate.

New thermal models also allow resolution independance similar to fluvial erosion.

Thermal erosion strength in new models can be set higher than 1 up to quite high values (10 or may be more), but usually this isn't needed. Moreover, in many cases (for example, if combined with fluvial erosion) it is better to set Thermal erosion strength to quite low values, sometimes even less than 0.1.

Daniil

D.A. Bentley (SuddenPlanet)