Board index FlightGear Support osgEarth

3.2 osgearth or other integrated real time satellite data?  Topic is solved

osgEarth renders the terrain scene by building the textured geometry at runtime from raw source imagery and elevation data.

Re: 3.2 osgearth or other integrated real time satellite dat

Postby SmergFerguson » Fri Jan 02, 2015 1:58 pm

dont know why it made a link out of that bat file but that is NOT the right link to download. search this thread the right one is in here somewhere.
SmergFerguson
 
Posts: 15
Joined: Tue Dec 30, 2014 7:47 am

Re: 3.2 osgearth or other integrated real time satellite dat

Postby Thorsten » Fri Jan 02, 2015 3:53 pm

There are no interstates on the version 3.2 I downloaded. I watched terrasync download additional data. Still no accurate roads.

The default install is flawed on the GAME SIDE.


Who cares what people think when you KNOW what you are doing is COOL and they don't even know what your talking about, they are just generally negative towards anything different than what they already understand.


Indeed - so they are.
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: 3.2 osgearth or other integrated real time satellite dat

Postby Hooray » Fri Jan 02, 2015 4:45 pm

SmergFerguson wrote in Fri Jan 02, 2015 1:56 pm:Is there a way to change plane type in game and keep your altitude and position and direction and such?


Nope, there isn't.
For a full -technical- explanation, please refer to:
http://wiki.flightgear.org/Reset_%26_re-init
http://wiki.flightgear.org/FlightGear_Sessions#Problem

So people are working towards this, so that aircraft can be switched at run-time - without this necessarily retaining the current position/velocities though - that will be difficult for other reasons, e.g. imagine hovering in a helicopter at 50 ft AGL and switching to a 747-400. There are a number of workarounds to make this somewhat possible - but it will involve XML & scripting.
Please don't send support requests by PM, instead post your questions on the forum so that all users can contribute and benefit
Thanks & all the best,
Hooray
Help write next month's newsletter !
pui2canvas | MapStructure | Canvas Development | Programming resources
Hooray
 
Posts: 12707
Joined: Tue Mar 25, 2008 9:40 am
Pronouns: THOU

Re: 3.2 osgearth or other integrated real time satellite dat

Postby jarl.arntzen » Fri Jan 02, 2015 9:56 pm

poweroftwo wrote in Wed Dec 31, 2014 9:13 pm:With osgEarth scenery enabled, terrain height is derived from different height samples than those used in the native Flightgear scenery. Because of the difference, 3D elements such as buildings and airport lights must be surface clamped. The clamping occurs based on viewport proximity to retain decent performance, however there are occasional artifacts including some objects that don't clamp correctly.


OSGEarth is performing extremely well at my mid range PC, running at a consistent 60 fps compared to ALS at around 20 fps with every feature switched on. :mrgreen:
I've been flying around testing and enjoying the scenery for far too long now but it's just so fun to see OSGEarth in action. EHAM, KSFO and
LEBL has never looked better.

I'm just curious on what to expect wrt to object clamping:
I have recently populated ENCN with some hangars, aircraft, lights and a tower. Even if I fly right next to the using the UFO, they do not seem to readjust to the surface but remain a good 15-20 meters below it. The runway lights seem to cope very well, though and keep reclamping to the surface as each pass loads in.

The square boundary box of the the airports seem to have been flattened in the elevation data too with the centre of the flat square matching the real height which, since ENCN is slightly sloped, results in a ridge running along the south and west edge in the middle of the water and an equivalent cliff along the east and north edge. Trees are above the terrain and ships below the ocean..

(Dont mind the colors. I've just been messing with the contrast.)
Image
Image

16 GB RAM, 3.6 GHz CPU 4 cores
gl-vendor:NVIDIA Corporation
gl-version:4.2.0
gl-renderer:Quadro 600/PCIe/SSE2
gl-shading-language-version:4.20 NVIDIA via Cg compiler

Now, I just need to see if this will also run at my Linux box too :)
Frequent Flyer. Occasional Lander.
jarl.arntzen
 
Posts: 106
Joined: Thu Jan 03, 2013 10:43 pm
IRC name: Jarl Arntzen
Version: 2017.1.1
OS: Ubuntu 14.04

Re: 3.2 osgearth or other integrated real time satellite dat

Postby poweroftwo » Fri Jan 02, 2015 11:55 pm

jarl.arntzen wrote:
OSGEarth is performing extremely well at my mid range PC, running at a consistent 60 fps compared to ALS at around 20 fps with every feature switched on. :mrgreen:
I've been flying around testing and enjoying the scenery for far too long now but it's just so fun to see OSGEarth in action. EHAM, KSFO and
LEBL has never looked better.

I'm just curious on what to expect wrt to object clamping:
I have recently populated ENCN with some hangars, aircraft, lights and a tower. Even if I fly right next to the using the UFO, they do not seem to readjust to the surface but remain a good 15-20 meters below it. The runway lights seem to cope very well, though and keep reclamping to the surface as each pass loads in.

The square boundary box of the the airports seem to have been flattened in the elevation data too with the centre of the flat square matching the real height which, since ENCN is slightly sloped, results in a ridge running along the south and west edge in the middle of the water and an equivalent cliff along the east and north edge. Trees are above the terrain and ships below the ocean..


Sorry to hear that the buildings are not clamping properly. I just took a quick look at the buildings at ENCN and I see the behavior. As I don't have time to debug in detail now, my guess is that the structures are not broken out into individual models and may be getting clamped as a single model which would cause them to be placed at seemingly random heights.

The square box is an attempt to roughly flatten out the airport surfaces as the DEM height data is generally too coarse and statistically noisy. Unfortunately, the box transitions sometimes appear quite abrupt. osgEarth is quite flexible, so if we find better sampled data for the airports, we could reference that in the future.

jeff
poweroftwo
 
Posts: 100
Joined: Tue Mar 05, 2013 4:35 am
Location: USA - Alabama

Re: 3.2 osgearth or other integrated real time satellite dat

Postby jarl.arntzen » Sat Jan 03, 2015 1:36 am

poweroftwo wrote in Fri Jan 02, 2015 11:55 pm:Sorry to hear that the buildings are not clamping properly. I just took a quick look at the buildings at ENCN and I see the behavior. As I don't have time to debug in detail now, my guess is that the structures are not broken out into individual models and may be getting clamped as a single model which would cause them to be placed at seemingly random heights.

Don't worry. I just wanted to let you know and find out whether this is a known problem which could possibly be fixed on my side. BTW: Those buildings, poles, aircraft and apron workers are all individual objects, placed using the UFO and then mass imported as STG to scenemodels.flightgear.org. A few misplaced objects isn't going to ruin any of the fun. :)
Frequent Flyer. Occasional Lander.
jarl.arntzen
 
Posts: 106
Joined: Thu Jan 03, 2013 10:43 pm
IRC name: Jarl Arntzen
Version: 2017.1.1
OS: Ubuntu 14.04

Re: 3.2 osgearth or other integrated real time satellite dat

Postby Thorsten » Sat Jan 03, 2015 8:09 am

OSGEarth is performing extremely well at my mid range PC, running at a consistent 60 fps compared to ALS at around 20 fps with every feature switched on


Because it doesn't have to do the (rather complicated) light and how fog is illuminated by light computations - that's a rigged comparison. If you'd /not/ have ALS do atmospheric light and assume light and fog are always white or have a simple color, it'd be way faster as well.
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: 3.2 osgearth or other integrated real time satellite dat

Postby Johan G » Sat Jan 03, 2015 8:33 pm

Stepping in as a moderator.

Enough is enough. I removed a number of posts to get back to the point before everything erupted. Most if not all of the posts was off topic and some of them being way more about personalities than about the subject of the topic.


One thing I often notice about the time a topic more or less suddenly is in flames is that people step aside from two equally simple and powerful principles* (in that order):

  1. "Assume good faith"
  2. The Golden Rule

Those two are the small and (too?) simple things that can keep things at a civilized (and productive) level, no matter if it is among siblings, in a marriage, between nations or global financial interests... or in this forum. :roll: :wink:

* If anyone of you are a non-native English user or does not recognize them, consider doing a Google search and you will quickly find that you are already familiar to them.
Low-level flying — It's all fun and games till someone looses an engine. (Paraphrased from a YouTube video)
Improving the Dassault Mirage F1 (Wiki, Forum, GitLab. Work in slow progress)
Some YouTube videos
Johan G
Moderator
 
Posts: 6625
Joined: Fri Aug 06, 2010 6:33 pm
Location: Sweden
Callsign: SE-JG
IRC name: Johan_G
Version: 2020.3.4
OS: Windows 10, 64 bit

Re: 3.2 osgearth or other integrated real time satellite dat

Postby poweroftwo » Sat Jan 03, 2015 8:47 pm

jarl.arntzen wrote in Sat Jan 03, 2015 1:36 am:
poweroftwo wrote in Fri Jan 02, 2015 11:55 pm:Sorry to hear that the buildings are not clamping properly. I just took a quick look at the buildings at ENCN and I see the behavior. As I don't have time to debug in detail now, my guess is that the structures are not broken out into individual models and may be getting clamped as a single model which would cause them to be placed at seemingly random heights.

Don't worry. I just wanted to let you know and find out whether this is a known problem which could possibly be fixed on my side. BTW: Those buildings, poles, aircraft and apron workers are all individual objects, placed using the UFO and then mass imported as STG to scenemodels.flightgear.org. A few misplaced objects isn't going to ruin any of the fun. :)


@jarl.arnstzen, I pushed a fix to the osgEarthNext branch. I plan to update the binary package later today. In the meantime, you can try to fly near the buildings at ENCN (where the clamping is invalid), from menu -> Rendering Dialog->Reload Scenery. Most likely the clamping will be improved.
poweroftwo
 
Posts: 100
Joined: Tue Mar 05, 2013 4:35 am
Location: USA - Alabama

Re: 3.2 osgearth or other integrated real time satellite dat

Postby poweroftwo » Mon Jan 05, 2015 6:18 am

@jarl.arnstzen, uploaded a new binary build with the minor building surface clamp fix along with a fix that properly sets the far clipping ranges when in detail (local) weather modes so that the globe can be seen. Windows binary download is again available here: http://forum.flightgear.org/viewtopic.php?f=22&t=25000&start=15#p228522

Here are some screenshots of KLSV, Nellis Air force base from various heights showing osgEarth ability to display various source imagery resolutions:

KLSV, Nellis AFB
Image

KLSV, viewed from a bit higher
Image

KLSV, higher still
Image

KLSV, can you find it? (near Lake Mead)
Image

Western US
Image

Continental US, Canada and Mexico
Image
poweroftwo
 
Posts: 100
Joined: Tue Mar 05, 2013 4:35 am
Location: USA - Alabama

Re: 3.2 osgearth or other integrated real time satellite dat

Postby jarl.arntzen » Mon Jan 05, 2015 6:57 pm

poweroftwo wrote in Mon Jan 05, 2015 6:18 am:@jarl.arnstzen, uploaded a new binary build with the minor building surface clamp fix along with a fix that properly sets the far clipping ranges when in detail (local) weather modes so that the globe can be seen.


Thanks! The clamping works far better now but seems to be dependent on the terrain height as measured directly below the UFO at the time of loading the final terrain tile. As such it does work pretty well on perfectly flat airports but on sloped airports like ENCN, the start position at one end of the runway is some 15 higher than the other and since I luckily have lots of RAM, I'm guessing that all LOD levels for all tiles in 1 km radius around the UFO are actually finished loading by the time I even start flying towards the buildings.
This means that the objects are still clamped some meters above the terrain and even if I fly towards them, no further LODs are left to load which means that they remain clamped to the height of the terrain where the UFO was when the last LOD tile loaded...I think.

I suppose clamping maybe even the first(?) object to the correct height at its position and then perhaps offset all the other objects according to the offset on the first one will also yield good results.

Scenery loaded while UFO at aircraft position. UFO was then moved back 150 m, further downward the sloped apron, prior to taking the screenshot.
Image

Scenery loaded while UFO at same position as above image for easier comparison
Image

Gotta get this working on my Linux box soon :)
Frequent Flyer. Occasional Lander.
jarl.arntzen
 
Posts: 106
Joined: Thu Jan 03, 2013 10:43 pm
IRC name: Jarl Arntzen
Version: 2017.1.1
OS: Ubuntu 14.04

Re: 3.2 osgearth or other integrated real time satellite dat

Postby poweroftwo » Mon Jan 05, 2015 7:27 pm

jarl.arntzen wrote:
Thanks! The clamping works far better now but seems to be dependent on the terrain height as measured directly below the UFO at the time of loading the final terrain tile. As such it does work pretty well on perfectly flat airports but on sloped airports like ENCN, the start position at one end of the runway is some 15 higher than the other and since I luckily have lots of RAM, I'm guessing that all LOD levels for all tiles in 1 km radius around the UFO are actually finished loading by the time I even start flying towards the buildings.
This means that the objects are still clamped some meters above the terrain and even if I fly towards them, no further LODs are left to load which means that they remain clamped to the height of the terrain where the UFO was when the last LOD tile loaded...I think.

I suppose clamping maybe even the first(?) object to the correct height at its position and then perhaps offset all the other objects according to the offset on the first one will also yield good results.


Yes, still work yet to be done here. The struggle is quality vs. performance. The elevation data can be queried very accurately but causes a noticeable loss of performance whereas the polygon intersection queries suffer as much faster but is a bit difficult to ascertain the LOD for a given location. Hope to focus on it in the coming weeks.

jeff
poweroftwo
 
Posts: 100
Joined: Tue Mar 05, 2013 4:35 am
Location: USA - Alabama

Re: 3.2 osgearth or other integrated real time satellite dat

Postby wlbragg » Mon Jan 05, 2015 11:01 pm

jeff,

Do you have any instruction or setup for building in Visual Studios?
Kansas and Ohio/Midwest scenery development.
KEQA, 3AU, KRCP Airport Layout
Intel i7/GeForce RTX 2070/Max-Q
User avatar
wlbragg
 
Posts: 7574
Joined: Sun Aug 26, 2012 12:31 am
Location: Kansas (Tornado Alley), USA
Callsign: WC2020
Version: next
OS: Win10/Linux/RTX 2070

Re: 3.2 osgearth or other integrated real time satellite dat

Postby poweroftwo » Tue Jan 06, 2015 12:53 am

wlbragg wrote in Mon Jan 05, 2015 11:01 pm:jeff,

Do you have any instruction or setup for building in Visual Studios?


I scripted a custom auto builder that works well for me. Let me know if you have any problems.
viewtopic.php?f=6&t=23404#p213447
poweroftwo
 
Posts: 100
Joined: Tue Mar 05, 2013 4:35 am
Location: USA - Alabama

Re: 3.2 osgearth or other integrated real time satellite dat

Postby Hooray » Tue Jan 06, 2015 3:09 am

for the sake of getting this accepted upstream, it would definitely make sense to revisit adapting the cmake superbuild, so that things work across all FG platforms
Please don't send support requests by PM, instead post your questions on the forum so that all users can contribute and benefit
Thanks & all the best,
Hooray
Help write next month's newsletter !
pui2canvas | MapStructure | Canvas Development | Programming resources
Hooray
 
Posts: 12707
Joined: Tue Mar 25, 2008 9:40 am
Pronouns: THOU

PreviousNext

Return to osgEarth

Who is online

Users browsing this forum: No registered users and 0 guests