TG Earth test

[bigben]

As you may have read in the forum I'm working on a data framework for building a high res model of the earth for TG.  I've played around with a few ideas and data sources already and it's looking fairly promising.  Taking a step backwards for now and focusing on getting the base data right.  Here's a quick resolution test of my first composite DEM quadrant.

GEBCO-30 (for the bathymetry) overlayed with GMTED-2010 (above 0.1m) overlayed with SRTM and output at 0.005°/pixel (18 arc sec) ... although some of Antarctica is only available at 30 arc sec.   I still have coastal bathymetry for the US to clean up and add in, but from an earlier test it looks pretty sweet. 
Elevation is exaggerated by 1.7, camera altitude ~250km

[zaxxon]

This is a very ambitious and exiting project, certainly one that I am looking forward to seeing (and possessing  )!

[bigben]

Yes, it doesn't take too much for it to become challenging as you can see in a recent test of a river/lake model. https://flic.kr/p/ra5iq5  There will be limitations, some of which I have some ideas for fixing... others I (and others) have no practical solution for.  But it's certainly an interesting challenge   

[TheBadger]

 

[bigben]

50km is too close but for static renders you can throw in an SRTM tile. (Haven't masked out the sea in the SRTM for this quick demo)  I'll be making a tutorial on preparing geotiffs to go with this model.

[TheBadger]

So after you get the planet done, then you add images over the displacement, and get a finished planet, or?
What does the end game look like, and how do you see people using your work? I am asking because I imagine that with the correct displacement in the right places, then you just need a NASA image map to finish it off, ideally... Correct?

[Ariel DK]

this is without doubt very interesting for me
I saw your profile on Flickr, and I've seen some lights nights there?
the night side is another story that still I can not finish 
the reason? It's was impossible for me, creating a day-night dynamic Earth
I decided create a most realistic night side in other project...
good job in these terrains... I love to see them with clouds and atmosphere

[Oshyan]

The goal is a complete data set necessary for reproducing a realistic Earth in TG. It's a project Ben and I have had in mind for quite a while (he's doing most of the heavy lifting ). I've seen many people showing Earth images here from time to time, of varying quality and realism, and most are based on the same or similar data. But due to lack of knowledge, software tools, or other reasons, the results are often less good (or at least less consistent) than they could/should be. So the aim here is to provide a known "best" quality base Earth model that people can just download and use. Clouds will be the main thing that will be done simplistically as creating good, realistic procedural clouds across a globe is a whole challenge in itself, but the Earth model will be somewhat modular (TGCs), so hopefully people will build good procedural cloud data to go with it.

- Oshyan

[bigben]

Data I have in hand at the moment includes elevation, the blue marble textures, lakes and rivers (tricky compromises there), city lights and a couple of nerdy experimental ideas.  I'm working on the high end data at the moment (e.g. 10gb elevation data to start with) to see how far I can push TG before my (work) system falls over.  Even at this early stage it's pretty clear that something like this has been long overdue. The elevation composite I have now is much nicer than my previous files.

...  I imagine that with the correct displacement in the right places, then you just need a NASA image map to finish it off, ideally... Correct?

We'll get to a basic model first which will include rivers and Then there's a couple of ideas that may or may not work...  Basically, if you're using imagery, you're stuck at a fixed resolution and if you're trying to go from space down to ground level, you'll have that awkward middle ground of switching from image to TG surfaces... but the ideas are largely theoretical at this stage