Board index FlightGear Development Scenery

osm2city.py development

Questions and discussion about enhancing and populating the FlightGear world.

Re: osm2city.py development

Postby merspieler » Mon Jan 04, 2021 10:19 pm

delay in loading is a "feature"....

The idea is, that objects that you don't see, don't need to be loaded... which is a good idea... (Until you do stuff like ray tracing)
It obviously comes with the issue, you observed... as it starts loading when you switch/move the view and things come into sight.
If everything is going against you, keep in mind that airplanes take off against the wind, not with it.
merspieler
 
Posts: 826
Joined: Thu Oct 26, 2017 10:43 am
Location: Wish to be in YBCS
Callsign: JST935. ORI1711
IRC name: merspieler
Version: next
OS: Debian Bullseye

Re: osm2city.py development

Postby awen » Tue Jan 05, 2021 12:29 am

merspieler wrote in Mon Jan 04, 2021 10:19 pm:delay in loading is a "feature"...

Thanks for the explanation! I never thought that FG can detect what the user sees instead of rendering everything.
awen
 
Posts: 78
Joined: Mon Oct 19, 2020 12:39 am
Location: Minnesota
Version: 2020.4.0
OS: Mac OS X

Re: osm2city.py development

Postby wkitty42 » Tue Jan 05, 2021 1:04 pm

awen wrote in Tue Jan 05, 2021 12:29 am:I never thought that FG can detect what the user sees instead of rendering everything.

if FG were to render everything, it could easily kill low spec systems from memory use... for example, the default pagedLOD of objects is set to 200... that's 200 objects... if you are in an area with thousands of objects, only 200 of them within the range will be displayed...

to explain this with a little more detail, i have another system with 32Gig RAM and a much much better GPU (RTX2060 [6Gig VRAM] vs GT730 [2Gig VRAM]) than my main system... on the new system, i have pagedLOD set to 4000(!)... the scenery looks much much better but FG starts off using a lot of system and GPU RAM, too... that system has 32Gig of system RAM... it starts up using maybe 1Gig before any of my apps are loaded... that puts FG using 6+Gig right out of the starting gate depending on where i spawn in... on a system with an OS that requires 4Gig of system RAM for the OS and only has 8Gig total system RAM, that leaves only 4Gig for FG to play in... i'll leave it to the reader to do the math to figure out what happens when you try to shove 6+Gig of stuff into a 4Gig hole ;)

we won't even mention that an increased number of objects to be rendered and/or moved around could very easily bring one's GPU to its knees, as well... FPS can take a huge hit in heavily detailed areas depending on how those details are implemented...
"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: 6962
Joined: Fri Feb 20, 2015 3:46 pm
Location: central NC, USA
Callsign: wk42
Version: git next
OS: Kubuntu 20.04

Re: osm2city.py development

Postby Paccidat » Wed Jan 06, 2021 8:41 am

Cool!

I wasn't aware of that setting.

My system is rather beefy, so by adding —prop:/sim/rendering/max-paged-lod=400 to the commandline I could double the amount of objects loaded?

And higher values if it can handle that ;-)
Paccidat
 
Posts: 21
Joined: Thu Dec 10, 2020 6:45 pm

Re: osm2city.py development

Postby wkitty42 » Wed Jan 06, 2021 5:05 pm

what version of FG are you running? newer versions have a slider in the in-sim F10->View->Rendering Options dialog... it is in the top left quadrant and titled "Maximum number of aircraft and scenery tiles"... you could set it there but yes, also using that prop line is fine...
"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: 6962
Joined: Fri Feb 20, 2015 3:46 pm
Location: central NC, USA
Callsign: wk42
Version: git next
OS: Kubuntu 20.04

Re: osm2city.py development

Postby vanosten » Sat Jan 09, 2021 7:07 am

A short reply on the last 10 or so messages. The LOD loading behavior can be tricky. In FG NEXT a new parameter (which osm2city uses) is introduced to make some calculations more fair. So in the next FG version with osm2city built with FLAG_AFTER_2020_3 some of this should improve - but not everything has been coded yet.

When discussing osm2city please always tell where you got the scenery from (e.g. worldbuild) and/or whether it is using building lists (aka. shader buildings - visually detectable by "rectangular houses seems to look more the same and are darker"). The loading behavior is different. Building lists are loaded per tile and in worldbuild larger and smaller houses (the ones not in building lists) are loaded in squares (ca. 3x3 km), with larger buildings having the medium LOD and therefore typically loading sooner.
If you are not within LOD distance, then FG typically does not load it. See also https://osm2city.readthedocs.io/en/late ... ry-objects (I use larger distances with my graphics card). Therefore, you also need to mention your FG LOD settings when discussing such matter.
In upcoming osm2city with FLAG_AFTER_2020_3 some of this will change as more buildings etc. will be rendered per tile.

Lastly: as osm2city is relatively heavy on disk usage, putting scenery on an SSD does not hurt.
Maintaining osm2city. Contributing with ground attack stuff to the OPRF FlightGear military-simulation community.
vanosten
 
Posts: 496
Joined: Sat Sep 25, 2010 5:38 pm
Location: Denmark - but I am Swiss
Callsign: HB-VANO
Version: latest
OS: Win 10 and Ubuntu

Re: osm2city.py development

Postby merspieler » Sat Jan 09, 2021 3:34 pm

So... I've managed to build with trees and it looks awesome...
But I've noticed two issues:
1st trees are placed in buildings:
Image

and 2nd: it doesn't seem to use the local tree definition making a hard cut between random and osm2city trees:
Image

I'll post the 2nd issue as well to the ml.
If everything is going against you, keep in mind that airplanes take off against the wind, not with it.
merspieler
 
Posts: 826
Joined: Thu Oct 26, 2017 10:43 am
Location: Wish to be in YBCS
Callsign: JST935. ORI1711
IRC name: merspieler
Version: next
OS: Debian Bullseye

Re: osm2city.py development

Postby vanosten » Sat Jan 09, 2021 4:00 pm

There is no point in posting stuff to the mailing list, when it is related to osm2city.

And please as always: tile index, lon/lat or nearest airport, FG version, parameters used. Also I can see that the background is not standard BTG.

Also: did you just generate trees - or did you run everything from scratch for the specific tile?
Maintaining osm2city. Contributing with ground attack stuff to the OPRF FlightGear military-simulation community.
vanosten
 
Posts: 496
Joined: Sat Sep 25, 2010 5:38 pm
Location: Denmark - but I am Swiss
Callsign: HB-VANO
Version: latest
OS: Win 10 and Ubuntu

Re: osm2city.py development

Postby merspieler » Sat Jan 09, 2021 4:13 pm

It's related to how flightgear interprets the tree list... so.. ml is probably the better place.

Tile: 5329469
Coords: 145.743556 -17.013742 facing south
Airport YBCS
FG: next, build at 2021-01-03
btg: same effect with default scenery
australia materials applied
If everything is going against you, keep in mind that airplanes take off against the wind, not with it.
merspieler
 
Posts: 826
Joined: Thu Oct 26, 2017 10:43 am
Location: Wish to be in YBCS
Callsign: JST935. ORI1711
IRC name: merspieler
Version: next
OS: Debian Bullseye

Re: osm2city.py development

Postby vanosten » Sun Jan 10, 2021 3:19 pm

merspieler wrote in Sat Jan 09, 2021 4:13 pm:It's related to how flightgear interprets the tree list... so.. ml is probably the better place.

Tile: 5329469
Coords: 145.743556 -17.013742 facing south
Airport YBCS
FG: next, build at 2021-01-03
btg: same effect with default scenery
australia materials applied


I have replied on the mailinglist. What is missing are materials definitions in FGData with a fair set of regionalisation. The beauty of that is that no coding is needed, only experimentation and maybe a talk to me. See also https://gitlab.com/osm2city/osm2city/-/issues/176.

For the second issue: I would be interested whether you ran osm2city from scratch or only parts of it -> which -e option?
Maintaining osm2city. Contributing with ground attack stuff to the OPRF FlightGear military-simulation community.
vanosten
 
Posts: 496
Joined: Sat Sep 25, 2010 5:38 pm
Location: Denmark - but I am Swiss
Callsign: HB-VANO
Version: latest
OS: Win 10 and Ubuntu

Re: osm2city.py development

Postby merspieler » Sun Jan 10, 2021 4:45 pm

Complete build.
Tho there was a previous build in that directory.
If everything is going against you, keep in mind that airplanes take off against the wind, not with it.
merspieler
 
Posts: 826
Joined: Thu Oct 26, 2017 10:43 am
Location: Wish to be in YBCS
Callsign: JST935. ORI1711
IRC name: merspieler
Version: next
OS: Debian Bullseye

Re: osm2city.py development

Postby vanosten » Sun Jan 17, 2021 10:51 am

My guess is, that the materials press stuff like trees through the photoscenery and you have not done: https://osm2city.readthedocs.io/en/late ... n-textures
Maintaining osm2city. Contributing with ground attack stuff to the OPRF FlightGear military-simulation community.
vanosten
 
Posts: 496
Joined: Sat Sep 25, 2010 5:38 pm
Location: Denmark - but I am Swiss
Callsign: HB-VANO
Version: latest
OS: Win 10 and Ubuntu

Re: osm2city.py development

Postby merspieler » Sun Jan 17, 2021 11:32 am

no, I didn't do that... but the trees have very distinct colors so I can easily tell osm trees apart from the random ones.

Also I've done a clean build with the same result.
If everything is going against you, keep in mind that airplanes take off against the wind, not with it.
merspieler
 
Posts: 826
Joined: Thu Oct 26, 2017 10:43 am
Location: Wish to be in YBCS
Callsign: JST935. ORI1711
IRC name: merspieler
Version: next
OS: Debian Bullseye

Re: osm2city.py development

Postby wkitty42 » Sun Jan 17, 2021 3:21 pm

would someone please kindly add some instructions to the docs on installing the prerequisites? i have no idea how to do that if i want to try to use OSM2City...
"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: 6962
Joined: Fri Feb 20, 2015 3:46 pm
Location: central NC, USA
Callsign: wk42
Version: git next
OS: Kubuntu 20.04

Re: osm2city.py development

Postby merspieler » Sun Jan 17, 2021 4:09 pm

In my install scrip I install the following packages:
git osmosis postgis pgadmin3 postgresql-contrib postgresql-client postgresql-postgis postgresql python-pip python3-venv python3-dev
Then I create a python venv and install the from the requirements.txt like so:
pip install -r osm2city/requirements.txt

You might as well want to take a look at easy-osm2city.
If everything is going against you, keep in mind that airplanes take off against the wind, not with it.
merspieler
 
Posts: 826
Joined: Thu Oct 26, 2017 10:43 am
Location: Wish to be in YBCS
Callsign: JST935. ORI1711
IRC name: merspieler
Version: next
OS: Debian Bullseye

PreviousNext

Return to Scenery

Who is online

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