Board index FlightGear Development Scenery

v0 landclasses

Questions and discussion about enhancing and populating the FlightGear world.

Re: v0 landclasses

Postby jaxsin » Thu May 05, 2016 10:53 am

bugman wrote in Thu May 05, 2016 9:26 am:Wayne, maybe you should get yourself into the WS 3.0 team ;)

Regards,
Edward


Yes, the guy has made some awesome scenery.

@wlbragg Did you guys not have a conversation that at some point that talked about using some kind of a different class to surround the water to fix a lot of the issues? IE. where roads etc would just end up in the water?
jaxsin
 
Posts: 395
Joined: Mon Dec 28, 2015 3:54 pm

Re: v0 landclasses

Postby statto » Thu May 05, 2016 2:33 pm

Rounding is common, and as I've stated before the resolution was reduced in the process from approx. 30m to about 100m in order to save on vertices and processing time, and to ensure the processor could finish properly.

Straight raster-to-vector grids aren't used very often.
Custom Scenery available from http://www.stattosoftware.com/flightgear
statto
 
Posts: 2106
Joined: Fri Jan 25, 2008 9:57 pm

Re: v0 landclasses

Postby wlbragg » Thu May 05, 2016 4:26 pm

Did you guys not have a conversation that at some point that talked about using some kind of a different class to surround the water to fix a lot of the issues?

Rounding is common,
~~~
Straight raster-to-vector grids aren't used very often.

I propose in the next gen to go ahead and do the rounding, we probably have to, to fit "most users", ie; "save on vertices and processing time, and to ensure the processor could finish properly". But to do it smart, by that I mean make all the water class sand or grass during that rounding process and then when that step is finished, cut a proper water layer in. What you should end up with at that point is a correct water layer and anywhere where the rounding grew the water past its original size ether sand or grass. No harm to roads. Not too difficult and then we would have the best of both worlds. I would even volunteer to collect and cut in the water layer. I think it is very important.
Also, as I said before, if we don't do something about the smaller chunks of deciduous forest, "add them back in", then there will be miles and miles of very boring scenery that might otherwise be spectacular.
Maybe some regional shader magic along river/stream class would do the trick? :idea:

Wayne, maybe you should get yourself into the WS 3.0 team

Thanks Edward, all I can do is suggest, as I am not on the 3.0 team. Although I do have my own team of one. :lol:
Kansas(2-27-15)/Ohio/Midwest scenery development.
KEQA (2-27-15), 3AU, KRCP Airport Layout
Intel i5 3570K AMDRX480
User avatar
wlbragg
 
Posts: 5874
Joined: Sat Aug 25, 2012 11:31 pm
Location: Kansas (Tornado Alley), USA
Callsign: WC2020
Version: next
OS: Win10/Linux/AMDRX480

Re: v0 landclasses

Postby psadro_gm » Thu May 05, 2016 4:39 pm

heh - there is no WS3.0 team. just a bunch of people working on tools, scenery, and infrastructure. mostly independently :)
8.50 airport parser, textured roads and streams...
psadro_gm
 
Posts: 751
Joined: Thu Aug 25, 2011 2:23 am
Location: Atlanta, GA USA
IRC name: psadro_*
Version: git
OS: Fedora 21

Re: v0 landclasses

Postby wlbragg » Thu May 05, 2016 4:48 pm

just a bunch of people working on tools, scenery, and infrastructure. mostly independently

And make no mistake, doing a very good job with the tools at hand.
Kansas(2-27-15)/Ohio/Midwest scenery development.
KEQA (2-27-15), 3AU, KRCP Airport Layout
Intel i5 3570K AMDRX480
User avatar
wlbragg
 
Posts: 5874
Joined: Sat Aug 25, 2012 11:31 pm
Location: Kansas (Tornado Alley), USA
Callsign: WC2020
Version: next
OS: Win10/Linux/AMDRX480

Re: v0 landclasses

Postby bugman » Thu May 05, 2016 4:51 pm

Well, maybe then a metaphorical team ;) One whose combined work will result in the next world scenery release.

Regards,
Edward
bugman
Moderator
 
Posts: 1799
Joined: Thu Mar 19, 2015 9:01 am
Version: next

Re: v0 landclasses

Postby wkitty42 » Thu May 05, 2016 7:18 pm

perhaps there should be a WS3.0 team working together so that others don't reinvent the wheel or get stuck on something?? :)
"You get more air close to the ground," said Angalo. "I read that in a book. You get lots of air low down, and not much when you go up."
"Why not?" said Gurder.
"Dunno. It's frightened of heights, I guess."
User avatar
wkitty42
 
Posts: 6833
Joined: Fri Feb 20, 2015 3:46 pm
Location: central NC, USA
Callsign: wk42
Version: git next
OS: Kubuntu 20.04

Re: v0 landclasses

Postby psadro_gm » Thu May 05, 2016 7:53 pm

a bit out of date, but I'm working on it at least a few hours every week :)

http://wiki.flightgear.org/TerraGear_roadmap

I was hoping to have a mesh at the end of April, but got delayed on this bug: https://github.com/CGAL/cgal/issues/781

It looks like they are getting close to a fix. Once I have a mesh, I plan on releasing a large area to dropbox, and get input from shader guys, like Thorsten and I4DNF, to see if they have anything additional they would like to see.
8.50 airport parser, textured roads and streams...
psadro_gm
 
Posts: 751
Joined: Thu Aug 25, 2011 2:23 am
Location: Atlanta, GA USA
IRC name: psadro_*
Version: git
OS: Fedora 21

Re: v0 landclasses

Postby legoboyvdlp » Thu May 05, 2016 8:52 pm

Pasdro, is it too late for me to make new shapefiles? I don't know what Florida is like, but I was quite disappointed when I flew into KMIA. There were entire islands missing in the Miami Beach area -- could I make those in QGIS? The area is just not the best, and I'd like to improve it.

I am sure that SHM would also like India to get in.
User avatar
legoboyvdlp
 
Posts: 7816
Joined: Sat Jul 26, 2014 1:28 am
Location: Northern Ireland
Callsign: G-LEGO
Version: next
OS: Windows 10 HP

Re: v0 landclasses

Postby psadro_gm » Fri May 06, 2016 12:30 am

It's never too late to make shapefiles. We are no where close to starting actual scenery generation. The last I heard, there may be a new world scenery built sometime to check new infrastructure - but this would just be a bugfix release + whatever has already been added to the cs_layer.

The new toolchain I linked above would be for a completely new scenery - departing with what was done in the past.
- draped airports, and roads, and other linear data, which could be turned off completely. This make the mesh much smaller for lower end machines
- true LOD - hopefully making earthview obsolete - we'll see.
- mesh curvature in addition to normals - a feature Thorsten requested, to differentiate between peaks and valleys. normals can point in the same direction on top of a hill, or bottom of a value, but the curvature will be opposite.
- raster image of landclass per tile - a feature requested by i4dnf - to aid in neighboring landclass lookup from shaders.

It also writes all intermediate files during processing to shapefiles. So debugging scenery issues / terragear bugs should be easier. You can load the intermediate data files into qgis :)

All I am close to having is a really good base mesh right now. World Scenery 2.0 had many mesh irregularities - breaking the urban shader, and making mesh simplification required by LOD impossible. I will generate a test area in Northern Italy with a corresponding simgear patch in the near future ( maybe another month ) to start testing it.
8.50 airport parser, textured roads and streams...
psadro_gm
 
Posts: 751
Joined: Thu Aug 25, 2011 2:23 am
Location: Atlanta, GA USA
IRC name: psadro_*
Version: git
OS: Fedora 21

Re: v0 landclasses

Postby Thorsten » Fri May 06, 2016 5:06 am

- true LOD - hopefully making earthview obsolete - we'll see.


I guess you can't beat a hires photo-textured sphere for orbital visuals... so ultimately we want to fall back to that I guess. Combine this with the requirement that GLSL doesn't like large numbers, and you end up using ray optics of some sort.

Earthview makes a lot of sense conceptually, though not necessarily in its actual implementation.
Thorsten
 
Posts: 11765
Joined: Mon Nov 02, 2009 8:33 am

Re: v0 landclasses

Postby erik » Fri May 06, 2016 8:36 am

psadro_gm wrote in Fri May 06, 2016 12:30 am:- draped airports, and roads, and other linear data, which could be turned off completely. This make the mesh much smaller for lower end machines

Draping is probably not as much as a problem as one first might think.
I've redone my local airport after it is reopened again (it used to be a military airbase). And it extends the previous mesh in some places which already get draped on top of the scenery (with a bit of z-fighting). So it's probably merely making sure the underlying terrain elevation and the airport elevation match closely. And then add an offset to the actual airport layout.

Erik
erik
 
Posts: 1705
Joined: Thu Nov 01, 2007 1:41 pm

Re: v0 landclasses

Postby psadro_gm » Fri May 06, 2016 11:38 am

The big issues with draping in fg are:

1) exclusion zones - we don't want the underlying terrain (i.e. forest) applying trees when there's a runway on top :)
2) terrain smoothing. we can't just drape an airport over unsmoothed terrain, and I'd still like mountain roads to be a cut in - hopefully to a lesser extent than what we do now. as we subdivide triangles with every road vertex. The new cut in will subdivide mesh triangles only at the road intersection points with the mesh triangles.
8.50 airport parser, textured roads and streams...
psadro_gm
 
Posts: 751
Joined: Thu Aug 25, 2011 2:23 am
Location: Atlanta, GA USA
IRC name: psadro_*
Version: git
OS: Fedora 21

Re: v0 landclasses

Postby erik » Fri May 06, 2016 1:41 pm

WED allows you to define two sort of exclusion zones. One real which is a big square, which should include all airport sections, and an airport boundary, which basically defines where to turn the terrain into airport grassland.

A lot of airport developers also define their own set of roads, parking lots and shrub areas. So maybe it's safe to define that if a road crosses an exclusion area (or airport boundary) it will most likely enter a tunnel. This is the case for Schiphol at least.

Erik
erik
 
Posts: 1705
Joined: Thu Nov 01, 2007 1:41 pm

Re: v0 landclasses

Postby bugman » Fri May 06, 2016 1:51 pm

Thorsten wrote in Fri May 06, 2016 5:06 am:Earthview makes a lot of sense conceptually, though not necessarily in its actual implementation.


It sounds like the current Earthview visuals should be quite easy to combine with and merge into the lowest level mesh LOD :) If a moon landing is considered a serious goal for the future, then maybe an even simpler mesh for large distances from the Earth would be a good idea? This would allow WS 3.0 to seamlessly integrate with deep space operations.

Regards,
Edward
bugman
Moderator
 
Posts: 1799
Joined: Thu Mar 19, 2015 9:01 am
Version: next

PreviousNext

Return to Scenery

Who is online

Users browsing this forum: YandexBot [Bot] and 2 guests