Board index FlightGear Development Scenery

Rhode Island and southeastern Massachusetts

Questions and discussion about enhancing and populating the FlightGear world.

Rhode Island and southeastern Massachusetts

Postby montagdude » Thu Feb 27, 2020 1:11 am

Hello FlightGearheads (is that a term?)! I hope you enjoy this custom scenery I made covering Rhode Island and southeastern Massachusetts. It is a big improvement over the default scenery, if I do say so myself. Sources used include:
  • NLCD for landcover
  • Void-filled SRTM-3 elevation
  • OpenStreetMap for roads, buildings, pylons, and other objects, using osm2city (be sure to turn on OpenStreetMap buildings and detailed roads and pylons for the best experience, if your computer can handle it)
To download, click on the Releases page here and download RIandSEMass.zip or RIandSEMass.tar.xz. Then unpack it and point FlightGear to the directory in the usual way.
https://github.com/montagdude/FGScenery-RIandSEMass

The area covered is on the small side, but plenty big enough to spend a few hours cruising around in a Cessna or Piper Cub. The actual coordinates of the bounding box are: West 69.5 - 72, North 41 - 42.15. Here is an image showing the area:
Image

Despite the relatively small size, it is still 1.3 GB of data when unpacked. The OpenStreetMap coverage is really dense in this area.

Here are some screenshots:
Image
Cape Cod

Image
Martha's Vineyard

Image
New Bedford, MA

Image
Providence Airport (KPVD)

Image
Downtown Providence, RI

This is an initial release. There are some minor issues that I'm aware of:
  1. There are some line-data waterways from OSM that are on top of water landcover data. I think I could fix these in the future by manually trimming those lines.
  2. There are some buildings and roads in the water here and there.
  3. There is a small gap with the default terrain on the West edge of the scenery (expected), but:
  4. There is a really big gap (as in, probably a couple miles) with the default terrain on the North end. Maybe that's because I cut off the top boundary at 42.15 instead of an actual tile boundary? If anyone has insight on this, let me know.
Anyway, let me know what you think!
Last edited by montagdude on Thu Feb 27, 2020 1:29 am, edited 1 time in total.
montagdude
 
Posts: 154
Joined: Tue Dec 31, 2019 6:04 am

Re: Rhode Island and southeastern Massachusetts

Postby xen_steda » Thu Feb 27, 2020 1:18 am

The screenshots look absolutely stunning!
xen_steda
 
Posts: 234
Joined: Sun Jan 05, 2020 3:53 pm

Re: Rhode Island and southeastern Massachusetts

Postby xDraconian » Thu Feb 27, 2020 2:02 am

Agree that the north gap was caused by the non-standard 42.15 boundary. 42.125 or 42.25 would have been aligned to the tile boundaries.
xDraconian
 
Posts: 304
Joined: Sun Jan 21, 2018 5:53 am
Version: Git
OS: Linux Mint

Re: Rhode Island and southeastern Massachusetts

Postby xDraconian » Thu Feb 27, 2020 2:41 am

Very nice work @montagdude. I threw away the northern-most tiles and flew around the area. You've learned the terrain generation quite well. Looking forward to seeing more of your contributions.

Scott
xDraconian
 
Posts: 304
Joined: Sun Jan 21, 2018 5:53 am
Version: Git
OS: Linux Mint

Re: Rhode Island and southeastern Massachusetts

Postby montagdude » Thu Feb 27, 2020 2:58 am

xDraconian wrote in Thu Feb 27, 2020 2:02 am:Agree that the north gap was caused by the non-standard 42.15 boundary. 42.125 or 42.25 would have been aligned to the tile boundaries.

Okay, thanks. Maybe I will update this at some point and fix that. Thanks for the compliments too. Having good data makes all the difference. I can run codes but not do art. :D
montagdude
 
Posts: 154
Joined: Tue Dec 31, 2019 6:04 am

Re: Rhode Island and southeastern Massachusetts

Postby montagdude » Thu Feb 27, 2020 4:56 am

xDraconian wrote in Thu Feb 27, 2020 2:02 am:Agree that the north gap was caused by the non-standard 42.15 boundary. 42.125 or 42.25 would have been aligned to the tile boundaries.

Hey, so I just tested this and don't quite understand the results. I ran tg-construct again with max-lat=42.125, but the huge gap is still there. I flew close and discovered that my scenery actually ends at 42.2, and the default scenery picks up again at 42.25. This kind of makes sense, because the shapefiles I generated actually go to 42.2. I say "sort of" because I thought tg-construct would trim them at 42.125 as requested. But on the West border, tg-construct apparently is actually trimming the shapefiles, because they extend to 72.05. I set min-lon to -72, and the custom scenery actually does end at West 72 when I inspect it in the sim. So the behavior seems to be different on the West boundary and the North boundary, and now I'm not sure the proper way to fix the North boundary. I suppose I could totally regenerate the shapefiles to actually end at 42.125, but I'd rather not if tg-construct is supposed to do it for me.
montagdude
 
Posts: 154
Joined: Tue Dec 31, 2019 6:04 am

Re: Rhode Island and southeastern Massachusetts

Postby pb321 » Thu Feb 27, 2020 7:43 pm

How long did it take Terragear to process the scenery on your computer with a Core i3 processor?
pb321
 
Posts: 201
Joined: Sun Nov 27, 2016 4:08 pm
Version: 2019.2.0
OS: Windows 10 Pro

Re: Rhode Island and southeastern Massachusetts

Postby montagdude » Fri Feb 28, 2020 1:07 am

pb321 wrote in Thu Feb 27, 2020 7:43 pm:How long did it take Terragear to process the scenery on your computer with a Core i3 processor?

Only about 30 minutes. The osm2city stuff takes a lot longer. That was more like 24 hours on 3 processes.
montagdude
 
Posts: 154
Joined: Tue Dec 31, 2019 6:04 am

Re: Rhode Island and southeastern Massachusetts

Postby montagdude » Fri Feb 28, 2020 1:38 am

Yeah, cutting it off at 42.125 N just isn't working. Even though I give tg-construct a max-lat of 42.125, it goes up to 42.2 anyway and leaves a big gap with the default scenery, which picks up again at 42.25. I guess I'll try redoing the landclass shapefiles and have them go all the way up to 42.25 and see if that fixes it. I did get rid of the waterways on top of water landclasses last night.
montagdude
 
Posts: 154
Joined: Tue Dec 31, 2019 6:04 am

Re: Rhode Island and southeastern Massachusetts

Postby Figaro » Fri Feb 28, 2020 2:14 am

montagdude wrote in Fri Feb 28, 2020 1:07 am:
pb321 wrote in Thu Feb 27, 2020 7:43 pm:How long did it take Terragear to process the scenery on your computer with a Core i3 processor?

Only about 30 minutes. The osm2city stuff takes a lot longer. That was more like 24 hours on 3 processes.


So what are the specs of the computer you use? I'm soon going to be building some scenery. I'm just curious about processing times. Obviously changes with system specs, but still just curious about rough numbers.

-S
User avatar
Figaro
 
Posts: 1323
Joined: Fri Feb 25, 2011 9:23 pm
Callsign: 4L-FIG
OS: ElementaryOS/Win10

Re: Rhode Island and southeastern Massachusetts

Postby montagdude » Fri Feb 28, 2020 2:32 am

Thinkpad T-430
4 x Intel Core i5-3210M @ 2.5 GHz
8 GB RAM

So not high-end specs by any means. If you are going to be doing osm2city, I think 8 GB RAM is probably the minimum you would need, and make sure you have some reserve swap space. It's also a good idea to break up the osm2city work into different sections. I used three for this one. The .pbf files ranged from 19 to 39 MB, just as a point of reference. The 24 hours estimate was to complete all three sections. It might have been a little less, actually, but it was on that order.
montagdude
 
Posts: 154
Joined: Tue Dec 31, 2019 6:04 am

Re: Rhode Island and southeastern Massachusetts

Postby xen_steda » Fri Feb 28, 2020 2:46 am

I just had to say once more , this scenery is really awesome! You did a really good job on the coastline for sure.

Image
Super accurate coastlines near 3MA5 Airfield.
xen_steda
 
Posts: 234
Joined: Sun Jan 05, 2020 3:53 pm

Re: Rhode Island and southeastern Massachusetts

Postby Figaro » Fri Feb 28, 2020 3:10 am

montagdude wrote in Fri Feb 28, 2020 2:32 am:Thinkpad T-430
4 x Intel Core i5-3210M @ 2.5 GHz
8 GB RAM


Thanks for letting me know.
I'll be running on a 6C/12T AMD Ryzen R5 2600X with 16GB of 3200MHz RAM. So hopefully will complete quite a bit faster. Thanks for sharing - timeframe was something I'm curious about!

Anyway, enough off topic chatter - this scenery does look really excellently detailed - lots of OSM and land use definition - looking forward to exploring it! Thanks for your work!

-S
User avatar
Figaro
 
Posts: 1323
Joined: Fri Feb 25, 2011 9:23 pm
Callsign: 4L-FIG
OS: ElementaryOS/Win10

Re: Rhode Island and southeastern Massachusetts

Postby montagdude » Fri Feb 28, 2020 3:33 am

Thanks guys. I haven't actually gotten to fly it too much myself yet, except for inspecting it with the UFO. I would tonight, but I'm too tired.
montagdude
 
Posts: 154
Joined: Tue Dec 31, 2019 6:04 am

Re: Rhode Island and southeastern Massachusetts

Postby xDraconian » Fri Feb 28, 2020 4:11 am

re: the northern gap between tiles... reduce your bounding box size a little -- e.g. 42.100 instead of 42.125.
Rationale: The bounding box defines which tiles you wish to produce. Since 42.125 is treated as the southern edge of the tile just north of your desired area it is being included. You need to be a bit shy of 42.125 to prevent that tile from being processed.
FYI... I dislike this behavior and I'm pretty sure that I improved this on TerraGear next. You may be running a different branch or version.
xDraconian
 
Posts: 304
Joined: Sun Jan 21, 2018 5:53 am
Version: Git
OS: Linux Mint

Next

Return to Scenery

Who is online

Users browsing this forum: No registered users and 1 guest