New Orbiter Beta Released (r.13, Mar 7, 2015)

[fort]

What's the status of BaseTiles ? Is the Orbiter 2015 still supporting them ?

Here in a config where the base is registered in Earth cfg:

...named "a" ( Heathrow )

; === Surface Bases ===
; place additional bases or
; base directories in this list
BEGIN_SURFBASE
DIR Earth\Base
a: 0 51.6796875
END_SURFBASE

and in Orbiter/Config folder:

...a cfg, "a.cfg"

BASE-V2.0
Name = a
Size = 70

with...one tile
( 8192x8192px - first picture ):

BEGIN_SURFTILELIST
0 -1 73 3
END_SURFTILELIST

or... four tiles ( with transparency )
( 1092x1092px each - second picture )

BEGIN_SURFTILELIST
5 -20 2342 3
5 -21 2342 3
5 -22 2342 3
5 -23 2342 3
END_SURFTILELIST

...this with Orbiter rev 10 ( 15-01-2010 ) and on EarthHi_12_15 textures ( West London/Heathrow )

 

[martins]

What's the status of BaseTiles ? Is the Orbiter 2015 still supporting them ? If I recall, they are "baked" into the surface tiles but what about planets using the the old tile format ?

For planets using the old tile format, nothing should have changed, including base tiles.

For planets using the new format, base tiles are currently still rendered, but they are likely to cause visual artefacts (for example, they don't know anything about terrain elevation, and they skip various render parameters), so the effect of using them is "undefined", to put it mildly. I may disable them altogether in the future, to avoid confusion.

 

[Fabri91]

@Xyon: the title on your asset download pages for the Moon and Mars still reads "Orbiter: Earth Textures Download".

 

[martins]

@Xyon: the title on your asset download pages for the Moon and Mars still reads "Orbiter: Earth Textures Download".

My fault I'll fix it some time ...

 

[fort]

Sorry Jarmonik,

I forgot that the config and the pictures was made with Heathrow at sea level with a 000040.elv from a tile at sea level from another place in the world ( 0701 / niveau13 / 286 ).

It's different with the original Heathrow elv ( first image ) or Heathrow Himalaya-ized ( second image )

Edit: what do you mean by old tile format ? Without elevation ?

 

[4throck]

I may disable them altogether in the future, to avoid confusion.

Please do. Had to manually remove the tiles from all my world bases in order to test them on the new beta. The bases themselves work 100% OK out of the box, so the tiles just mess things up.

I guess that in the future high res terrain texture will always go as meshes. Makes sense, but for that we'd need some terrain export tool...

 

[Photon305]

Hi, just registered to say just how great the new beta is (the high-res textures make a massive difference, it's awesome to fly over land and see the individual fields below, makes it feel like I'm really up there!), and also to report some issues I'm having with it, hope this is the right place.

Firstly, when cloud shadows are enabled, they look fine above a certain altitude but close to the ground they switch from being black to white, looking identical to the cloud layer above them (I'd post a screenshot, but I can't figure out how to upload images here)

Also, when it crashes, the system font (all window titles, text in windows explorer, desktop etc) seems to get reset to an uglier one - this happens whilst the Orbiter Beta window is open normally, but reverts when I exit. If it crashes though it stays until I reboot the computer.

Speaking of clouds, are cloud textures at a higher resolution than the current level possible/planned? The ground now looks amazing from the altitude of the ISS, so the clouds though still impressive look a bit smudgey in comparison.

Also look forward to seeing your bases again in the new version, 4throck - I chose to launch from Gran Canaria and found myself parked on the sea Wasn't there talk of potentially integrating them with stock, or is that no longer the plan? It would be awesome if higher-res textures for all the base areas could be integrated with the new high-res Earth terrain rather than relying on base tiles, but I'd imagine that'd be quite a hefty undertaking.

 

[martins]

Firstly, when cloud shadows are enabled, they look fine above a certain altitude but close to the ground they switch from being black to white, looking identical to the cloud layer above them (I'd post a screenshot, but I can't figure out how to upload images here)

This is known. Cloud shadows are not completely implemented yet. For now, you can either disable them, or disable the rotation of the cloud layer as a workaround.

Also, when it crashes, the system font (all window titles, text in windows explorer, desktop etc) seems to get reset to an uglier one - this happens whilst the Orbiter Beta window is open normally, but reverts when I exit. If it crashes though it stays until I reboot the computer.

This is a feature, as they say. Orbiter by default disables font smoothing to improve performance. If it crashes, it has no opportunity to switch it back on. You can fix this by running Orbiter again and exiting normally, provided that "Extra | Debugging options | Performance options | Enforce font smoothing on exit" is turned on.

You can suppress Orbiter's disabling of font smoothing by turning off "Extra | Debugging options | Performance options | Disable font smoothing while Orbiter is running". But this may hit performance.

Speaking of clouds, are cloud textures at a higher resolution than the current level possible/planned? The ground now looks amazing from the altitude of the ISS, so the clouds though still impressive look a bit smudgey in comparison.

If you find a high resolution cloud map source, you can generate cloud layers up to resolution level 19 if you want (about 1 meter resolution). You'll just need a few petabytes of disk space to store them.

 

[4throck]

Also look forward to seeing your bases again in the new version, 4throck - I chose to launch from Gran Canaria and found myself parked on the sea

That's odd. I've tried them all and the placement matches well with the new global textures. That's expected, since I used Landsat textures in the first place.

As far as I'm concerned, by all means include the base cfgs on the default distribution. They are just text files after all

 

[jedidia]

I'd post a screenshot, but I can't figure out how to upload images here

In the reply view (not in the quick reply at the bottom of the page), click on "attachments" (the paper clip icon), and you'll get to a window where you can upload files.

 

[Photon305]

Thanks for the reply Martin! I had tried restarting and exiting Orbiter, but I hadn't got Enforce font smoothing on exit turned on.

Shame that clouds maps at higher resolutions than the current one don't seem to be easily available - I'd assumed the main limiting factor was the engine, but it looks like it doesn't really get any better than NASA's Blue Marble clouds...

That's odd. I've tried them all and the placement matches well with the new global textures. That's expected, since I used Landsat textures in the first place.
QUOTE]

I was referring to the stock (lack of) base, it was what reminded me of yours Now I've realised that fixing the configs is as simple as deleting a couple of lines at the end I've got them working fine

In the reply view (not in the quick reply at the bottom of the page), click on "attachments" (the paper clip icon), and you'll get to a window where you can upload files.

Thanks Jedidia, I was quick-replying so I didn't see it

 

[mariostivala]

Atlantis does not respond to script commands

Hello, I'm experiencing problems with launch.lua script for Atlantis in Orbiter beta. It does not respond to bank and pitch commands: the launch is vertical and the veicle does not enter in orbit. have not tried other scripts, but also UAP (that works with the DG and with Shuttle-A) can not drive well the Atlantis. Someone has already found a solution?

 

[martins]

You are right. In Orbiter 2010, the default Atlantis module uses a fake RCS system to control attitude during launch. In the new version, attitude is controlled by gimballing the SRB and Shuttle engines. The autopilot script hasn't been updated to reflect this, and is trying to address RCS engines which don't exist anymore. I'll see if I can fix this.

Or maybe you could have a look at the script yourself and find a solution. Let me know if there are any Lua interface functions missing (I can't remember if the thruster gimbal commands are exposed to Lua yet).

 

[jarmonik]

D3D9Client Beta 11 for Orbiter 2015 (rev 10) has been uploaded to CodePlex.

 

[Felix24]

(Continuing the discussion from the Screenshot Thread)

It doesn't show much in sunlit areas, but is distinctive in shaded areas, as one would expect.

Is your new VC rendering method designed with possible future features in mind, like cockpit self-shadowing and local cockpit lighting?

 

[Jumper_001]

Is there anything specifically in the Beta that needs testing attention right now?

 

[martins]

Is your new VC rendering method designed with possible future features in mind, like cockpit self-shadowing and local cockpit lighting?

Not really. DX7 is quite limited in that respect, so I think I leave this to external graphics client developers.

 

[Jumper_001]

You are right. In Orbiter 2010, the default Atlantis module uses a fake RCS system to control attitude during launch. In the new version, attitude is controlled by gimballing the SRB and Shuttle engines. The autopilot script hasn't been updated to reflect this, and is trying to address RCS engines which don't exist anymore. I'll see if I can fix this.

Or maybe you could have a look at the script yourself and find a solution. Let me know if there are any Lua interface functions missing (I can't remember if the thruster gimbal commands are exposed to Lua yet).

I'm going to take a crack at this just because I am bored and I have some time.

As far as I can tell the strategy used by the "manual" controls on the new Atlantis is that prior to SRB sep, pitch is controlled by the SSME thrust vectors deflecting and roll/yaw are controlled by the SRBs, and then after SSME sep all three are controlled by the SSMEs?

It looks like the v:set_thrusterdirection is exposed in the Lua interface. Instead of controlling RCS thruster groups, the code in the SetRCS function could scale and combine the required rotations for the commanded pitch, roll, and yaw rates to create a final set of gimbal settings and update the thrusterdirection for the SSMEs + SRBs (before t=126) or SSMEs (t>=126) accordingly.

Edit: I've been thinking about this a a bit more and looked into the per-timestep code in the Atlantis.cpp file. It seems to me that if I start modifying engine gimbal settings directly via the Lua interface (my first thought was to replicate the AutoGimbal C++ code in Lua), the .dll for Atlantis will see that it's manual controls are set to zero when it runs its timestep code, and will reset the gimbals to null the rates every step, and the two systems will basically fight each other.

I also see new code for a C++ Atlantis autopilot in there that wasn't in the old 2010 release, so maybe the LUA Atlantis AP example is going to be obsolete anyway?

 

[Gingin]

Question about D3D9Client Beta 11 for Orbiter 2015 (rev 10)

Good Morning everyone.
Thank you Martin for this incredible new beta version , it's amazing the work done since My begginings with Orbiter 2006 :thumbup: :thumbup:

I have been trying the new beta for 2 weeks now.
I have a question for Jarmonik (very nice work with D3D9 ,amazing )
I tried your last plugin for direct x 9 : D3D9Client Beta 11 for Orbiter 2015 (rev 10)


New render for water is amazing but I have the feeling that something changed in environmental shaders .
We can't select them anymore (default ,advanced , experiment )


Here is a picture with D3D9Client Beta 11 for Orbiter 2015 (rev 10) :




Here a picture with D3D9Client Beta 9 for Orbiter 2015 (rev 9) and advanced environmental shaders



And here a picture with D3D9Client Beta 9 for Orbiter 2015 (rev 9) and experiment environmental shaders.



The first one picture with last plugin looks like to be in shadows compare to the 2 others.
Like if an Umbrella was between the sun and the Earth.
No sun reflection anymore on the water.
Is it normal to have that with the last plugin D3D9Client Beta 11 for Orbiter 2015 (rev 10) ??

 

[jarmonik]

I have a question for Jarmonik (very nice work with D3D9 ,amazing )
I tried your last plugin for direct x 9 : D3D9Client Beta 11 for Orbiter 2015 (rev 10) ....

Right now, atmospheric shaders are experimental (all of them). Some features may be broken and not working at a moment. I'll take a better look into them when migrating a cloud layer shadows in the shaders. There are a few bugs and problems with the current shaders and I am trying to fix them from the future releases.

---------- Post added at 15:05 ---------- Previous post was at 14:48 ----------

Local light sources are a little problematic. Having a multiple light sources active in a scene with a high vertex count is causing a performance impact. There could be a way to reduce the impact.

First of all, it should be pretty easy to compute a bounding cone for a spotlight. Also, cone-sphere intersections are easy and fast to compute. Which could be used to narrow-down which meshes and mesh groups need to be rendered with local lights sources active.

Also, for every lit mesh group, it would be possible to select maximum of 2-4 light sources those are contributing most of the light to the group (taking some kind of directional weight factor in count). For an example in the D3D9Client it would be possible to render 1-2 most effective local lights in a per pixel basis and the rest in a per vertex basis. (This could, of course, produce some lighting anomalies in a mesh group boundries)

So, a number of active lights in a scene could be increased where as a number of active lights during rendering would be decreased.

Also, it would be a good idea to explore a possibilities of pre-computed (baked) lights via light-maps.

---------- Post added at 15:09 ---------- Previous post was at 15:05 ----------

The ideas a self-shadowed virtual cockpit has crossed my mind. But I have considered it from a sun light perspective. How would the sun light lit the cockpit when shining through cockpit windows.