New Release D3D9Client Development

[Donamy]

I'm having the same RMS attachment issues, with the BETA that were fixed in 2010 P1.

I'm using latest BETA with D3D9beta4.

 

[jarmonik]

D3D9ClientRC34b for Orbiter Beta

Here is RC34b build for the Orbiter Beta [110824].

There are two new options in D3D9Client.cfg

"ManagedTiles" setting this to "1" will load surface tiles in managed pool instead of default pool. (May help with video memory issues).
"LoadTexturesInSystemMem" when this is set to "1". All textures will be loaded in system memory from where they are moved into a video memory when required.

 

[Donamy]

That fixed it !!

 

[neutronium76]

Hello,
I have been following this thread from the beginning - I am reading it for 2-3 days now.
I want to try this client
1. On standard 2010P1 orbiter installation
2. On the latest 082411 beta
So I am going to create 2 folders: One for the standard 2010P1 orbiter and one for the latest beta 240811
My questions are:
1. Which client goes with which version?
2. Regarding the beta (240811) Orbiter (found in sourcforge) do I need to install 1st the official 2010P1 release and then extract on top of it the beta OR extract the beta on an empty folder directly (and then the appropriate D3D9client)
3. can I try hires textures for the earth, moon, mars and outerplanets or shall I expect trouble? At what stage shall I extract the hires textures during the installation? Orbiter distribution --> D3DClient --> hires textures or Orbiter --> Hires textures --> D3DClient?

Sorry for my bad english. A big thank you to the developers and creators of this project.

 

[Hayab]

Hello,
I have been following this thread from the beginning - I am reading it for 2-3 days now.
I want to try this client
1. On standard 2010P1 orbiter installation
2. On the latest 082411 beta
So I am going to create 2 folders: One for the standard 2010P1 orbiter and one for the latest beta 240811
My questions are:
1. Which client goes with which version?
2. Regarding the beta (240811) Orbiter (found in sourcforge) do I need to install 1st the official 2010P1 release and then extract on top of it the beta OR extract the beta on an empty folder directly (and then the appropriate D3D9client)
3. can I try hires textures for the earth, moon, mars and outerplanets or shall I expect trouble? At what stage shall I extract the hires textures during the installation? Orbiter distribution --> D3DClient --> hires textures or Orbiter --> Hires textures --> D3DClient?

Sorry for my bad english. A big thank you to the developers and creators of this project.

As first install the Orbiter 2010P1
- download the orbiter110822-100830diff.7z from SF and extract it in your Orbiter 2010P1 Folder (override all Folder and Files)
- get the orbiter110824-110822diff.7z and extract it in Orbiter2010P1 Folder
- now you have the current Beta 110824
- as the last Step download D3D9ClientRC34b_forBeta.zip, extract it in Orbiter2010P1 Beta Folder.
- run the orbiter_ng.exe (not the orbiter.exe) and enable in orbiter Launchpad window, Modules, D3D9Client.
- optionally you can install 2010P1 addons, but not all does work with D3D9 client
p.s.
- you can run the default beta client with orbiter.exe, (all 2010P1 Addons should work with it)

 

[Astronut25]

RC34 Bug List

Everything looks wonderful, great job.


Here is a list of the bugs I've noticed

The red circle marks the Guiana Space Center, as seen from the Kennedy Space Center.


There is an odd line perpendicular to the ecliptic when it is shown.


The base objects don't transition between night/day at the right time.


My most recent one, the textures with transparencies make the mesh groups behind them dissapear as if it was higher on the .msh's list (but they are on the bottom)(also, this only effects untextured groups)

left: DX7
right: DX9​

note: in RC33+ surface tiles it looks like the alpha channel of the diffuse texture is inverted for effect of the normal map (transparent for diffuse is opaque for normal & vice versa)

 

[jarmonik]

The red circle marks the Guiana Space Center, as seen from the Kennedy Space Center.

I have never seen anything like that and there is no line of sight. Must be caused by some addon I don't have.

There is an odd line perpendicular to the ecliptic when it is shown.

I thought it was there on purpose marking the vernal equinox. Should we remove it or keep it ?

The base objects don't transition between night/day at the right time.

I haven't detected any significant difference there. The day/night transition is currently controlled by the Orbiter. I suppose the darkness may come in the D3D9Client a little earlier or later than in the inline engine. I haven't done any exact measurements about the matter. But anyway there is a bug in the rendering of the night textures. Currently the day texture is replaced by the night texture when it should be rendered on a top of the day texture.

My most recent one, the textures with transparencies make the mesh groups behind them dissapear as if it was higher on the .msh's list (but they are on the bottom)(also, this only effects untextured groups)

Currently the rendering order is:
- Normal mapped groups
- Textured groups
- Non-textured groups
So, If there is a transparency used with normal mapped or textured groups then a the problem you described will appear. I suppose I need to add one more rendering pass for a transparent surfaces after all other groups are rendered for all vessels.

for every vessel render {
- Normal mapped groups
- Textured groups
- Non-textured groups
}
for every vessel render {
- Transparent groups (with distance sorting)
}

But, there is still a problem of how to detect a use of transparent textures when the material is fully opaque.

in RC33+ surface tiles it looks like the alpha channel of the diffuse texture is inverted for effect of the normal map (transparent for diffuse is opaque for normal & vice versa)

I took a look into the shaders and couldn't find anything. Could this be related into the rendering order problem described above. But I'll run a few test and see if I can reproduce this.

 

[neutronium76]

Hi again. I am wondering if this http://en.spaceengine.org/ would be of any help to the developers? Just discovered it from youtube. If it could be somehow merged into orbiter it would be a :10sign::woohoo::11sign: spacesim!

 

[Astronut25]

I have never seen anything like that and there is no line of sight. Must be caused by some addon I don't have.

It happens when there are surface tiles on the bases. I've added them to all major space centers, so they all do that. (GSC has the largest surface tiles, so it is the most noticable)

I thought it was there on purpose marking the vernal equinox. Should we remove it or keep it ?

That seems like a good idea, I didn't see that on the inline version and didn't know that it was marking that.

Oh, and one more bug I forgot, I'm having trouble with the virtual cockpit lighting for SC3 vessels. I'll try to describe it, it is like the range of the light is too short, and the contrast is screwed up.

DX7 DX9
angle 1


angle 2

 

[FordPrefect]

Hi again. I am wondering if this http://en.spaceengine.org/ would be of any help to the developers? Just discovered it from youtube. If it could be somehow merged into orbiter it would be a :10sign::woohoo::11sign: spacesim!

I discovered Space Engine about a month ago, and fell in love with its graphics. Especially, the atmospheric rendering in- and outside the atmosphere (view from space) is absolutely mindblowing beautiful, because it looks very real (as far as one can judge from looking at photos taken from Earth orbit). It is my understanding, that Valdimir Romanyuk, the creator of SpaceEngine, used Eric Brunetons spectacular real-time atmospheric rendering source code in his engine. However, SpaceEngine is based on OpenGL, so I am not sure if there's any chance of combining the two (besides SpaceEngine is not open source).

And as far as I know, jarmonik once had a D3D9 client version, which actually used part of that atmospheric rendering code as well, but I seem to remember he was running into some issues with the transition of the rendering from inside to outside the atmosphere. At that point, jarmonik dropped the new atmospheric rendering method (for the time being) to focus on the solution of the major bugs. But what was there, looked great, much better than the default inline rendering method.
I very much hope, that jarmonik, as soon as the major bugs are ironed out, can include again the new atmospheric rendering technique into his D3D9client, maybe even with the help of the good doctor Martin Schweiger, in case some changes were needed in the core of the Orbiter engine.
The day jarmonik will manage to do that, and perhaps along with planetary normal/bump maps for better visuals/shadows (especially on moons - see this SpaceEngine screenshot) near the terminator area (...please, please...), it would be a dream come true.

:tiphat:

Now back to the D3D9client bug search and destroy subject...

 

[Cras]

I am trying the D3D9 client for the first time.

I started with a fresh Orbiter 2010p1, installed UMMU and Orbiter Sound and the XR-2 1.4

I am getting some strangeness in the 2d panel.

 

[Fabri91]

What version of the XR vessels were you using? Recently a new version was released to make them work better with this client.

 

[Cras]

XR2 version 1.4, which I believe is still the newest one.

In the manual, it says it is recommended to use the D3D9 client so I assume it is the same version you are referring to.

I have no experience with the D3D9 client. I tried the Open GL client once with little successs. I have since been running Orbiter in XP Pro x64 since I have been using Shuttle Fleet almost exclusively.

Now that I am starting some XR2 missions, I figured I would give the D3D9 client a try since I have heard great things both in the visuals and the performance under Windows 7. The default Atlantis launch scenario seemed to work just fine in D3D9, but the XR2 is not working right. All the 2d panels are corrupt in similar fashion, with certain information areas not rendered correctly. So far, the only XR vessel I have installed is the XR2.

 

[jarmonik]

I am trying the D3D9 client for the first time.
I started with a fresh Orbiter 2010p1, installed UMMU and Orbiter Sound and the XR-2 1.4
I am getting some strangeness in the 2d panel.

It's a graphics driver bug. Installing the latest drivers should fix the problem. The same problem is reported here. http://www.orbiter-forum.com/showthread.php?p=284233&postcount=32

---------- Post added at 00:57 ---------- Previous post was at 00:51 ----------

It happens when there are surface tiles on the bases. I've added them to all major space centers, so they all do that. (GSC has the largest surface tiles, so it is the most noticable)

Oh, and one more bug I forgot, I'm having trouble with the virtual cockpit lighting for SC3 vessels. I'll try to describe it, it is like the range of the light is too short, and the contrast is screwed up.

Could you post a screen shot from both of these bugs.

 

[Astronut25]

...
Could you post a screen shot from both of these bugs.

They are in the previous posts (#826 & #829)

...
The red circle marks the Guiana Space Center, as seen from the Kennedy Space Center.

...
Oh, and one more bug I forgot, I'm having trouble with the virtual cockpit lighting for SC3 vessels. I'll try to describe it, it is like the range of the light is too short, and the contrast is screwed up.

DX7 DX9
angle 1


angle 2

...
"

Was that what you were asking about?

 

[jarmonik]

They are in the previous posts (#826 & #829)

Sorry, I can't see any screen shots in your posts. Can anyone else ? There's only a few lines of text. Is the forum playing tricks ?

DX7 DX9
angle 1

angle 2

 

[Cras]

I dont see any screenshots either.

 

[Turbo]

Same, just text

 

[kuddel]

Same here, just text. The only post containing a picture is Cras' #831

 

[Wishbone]

There are hyperlinks to album pics in the text, try hovering mouse over the dashes.