TG Earth model progress.

[archonforest]

Doing a fantastic work here!! Thx for all your efforts ben
And it will be free to play with it? Wow that is very cool.

[lat 64]

Well I sure hope Ben gets rewarded for the efforts.
I am working on animating a sounding rocket lift off, and I want to have the camera follow it up to very high altitude.
Would be so great to have the land mass recognizable. As it is, the Earth model looks already useful.

You go!
r.

[Oshyan]

Ah right, that's the reason to use a sphere instead of planet, of course. Glad you figured that out. The sub-surface is looking great now. What's the source of the black shore artifacts hough?

And yes, the intention is that the data be hosted by Planetside and available for free. Earth as a subject is a common interest and need for many, and the base data is already freely available. The problem has been that the data needs some work in various ways, not just for TG but for almost any renderer. To get a really good result it takes some work, and frankly I don't think anyone has achieved the best possible quality yet. So this project is intended to tackle two main issues. We want to eliminate the duplication of effort (reinventing the wheel, so to speak) that people have had to do in setting up their Earth models, which I've seen people do at least 10 times just here in the forums alone. And then we also want to increase the actual overall resolution and quality of the data that's available and easy to use. The end goal is an easy-to-use Earth model with the best possible quality.

- Oshyan

[bigben]

I honestly have no idea on the artefacts, or indeed the subsurface detail.  There is a substantial difference between using a planet and a sphere in this case, especially regarding decay depth.  I'll run a render tonight with both using the same shaders to demonstrate but the two sets of images in this post have very similar shaders.

[Ariel DK]

The end goal is an easy-to-use Earth model with the best possible quality.
- Oshyan

for me, that would be a utopia made reality
If this is successful, I would like share my effort with this project
I'm work in a "semi-procedural" clouds layers
that would be the 1st part of my personal project
the 2nd part would be the terrain, and share here when it finish
but with the effort of users like Big Ben, this is not necessary now
I am anxious to see the result and help in whatever with this project
I wish you the best of luck Mr Ben

[Oshyan]

Ben, did you disable shadow casting in the sphere?

Ariel, yes I think your clouds could become a useful part of the project. There will be one or more base cloud masks (image-based), but we will want procedural detail to fill it in. If you have useful cloud masking techniques (e.g. continent-based, as you have shown), then you could potentially provide your own additional setups for that.

- Oshyan

[bigben]

Yes, shadows were disabled on the sphere.  I must have also changed the decay depth. Rendered both ocean models with the same settings: planet and sphere with cast shadows on look the same.

[bigben]

Got to spend a bit more time on this.  Basic city lights this time. I've set up a mask for the shaded side of the earth based on the sun's position, providing a gradated border to fade them in.  The image below was based on a preview without the atmosphere so it looks like the lights come on a little early.  The mask is adjusted via a colour adjust shader specifying the ON and OFF limits in degrees from a vector opposite the sun's position.

[Dune]

This looks really sophisticated, Ben. Awesome work!

[bigben]

Thanks.

Most of the nodes can probably be buried into internal networks once it's done. I'm trying to keep it functional for animations and also relatively straightforward for people to add their own tweaks.

Here is a rivers and lakes test.  Based on an image map derived from vector data for lakes and rivers (and islands). The offset from the image may take a bit of editing of the texture map, although I'm hoping most of it is due to resizing in Photoshop rather than exporting from GlobalMapper at lower resolutions.  I've checked without any other surfacing and the water IS sitting on the surface of the planet.  The width of the rivers is a bit unrealistic, but I've narrowed them as much as possible using a colour adjust shader on the mask.  The water is just a masked surface with a water shader. To keep the lakes flat the DEM is loaded into 2 heightfield nodes, one with fractal detail and one without, using the water mask to apply one or the other.

[Dune]

How will you deal with areas that are below sea level, but still dry? Mask? The western half of my country is well below sea level 

[bigben]

Separate coastal mask for the sphere that forms the ocean surface. Didn't want to flood Lake Eyre either. I have 3 water-related masks in a single RGB image, pulled out using convert to R,G,B nodes although I think one of them is unused at the moment

Must check what happens at the Netherlands coast. Hopefully the ocean isn't floating above

[bobbystahr]

Getting more excited about this with every up date. Awesome work Ben!

[Oshyan]

Great to see further progress on this, especially the city lights setup. The dark side of the Earth is rather bright in general though, isn't it?

Also, still worried about the appearance of shadow from the lakes which makes it look like they're hovering above the land. Why is that? Almost like it's "embossed" or something (simple Photoshop effect). But the amount of detail is great!

- Oshyan

[bigben]

There is a fill light on the dark side in that render... wanted to see where stuff was still.

As for the lake edges, I've georeferenced the blue marble image (quite easy with a global image) and am exporting the lower resolution versions from that. Checked the new texture maps with the masks and there are some 1px differences. This would also be exacerbated by trying to sharpen the edge of the lake/river mask with a colour adjust shader. I might have to try a content-aware deletion of the lake areas from the blue marble image to try and "shrink" the lakes.