Project D3D11Client Development

[Frogisis]

This new version fixed the hard edges of the atmosphere, but now unfortunately the upper reaches of it no longer cover the edge of the clouds on Earth, leaving a pixellated line of cloud texture above the limb/horizon, and that one glitch is back where the day/night sides of the clouds are often misaligned with the day/night sides of the planet.

To be more specific in regards to the dots/range thing, what I was trying to describe was how currently, as you're approaching a planet from deep space, it first appears as a flat gray dot, but as you get closer, at some point it will instantanously pop into being as a sphere lit in the proper planetary color.
That "jump" seems to be a function of the way the core handles things, but I wonder if it wouldn't be possible to do something like extend a patch of the celestial body HDR to interplanetary distances so it would be superimposed over the dot, giving it an appropriate hue and making that transition less abrupt.

 

[asmi]

Hy Asmi,

I've noticed that a fresh checkout of the current terrain branch is not building due to 2 missing files:
* pnglibconf.h referenced by png.h since the start of the terrain branch and
* pngstruct.h referenced by resources.h.

Would it be possible to add those files to the repository?

cheers,
Face

I've uploaded these and some other required files. Please re-download and let me know if there is anything still missing...

 

[luki1997a]

I'm really impressed, it looks wonderful :thumbup:

But:

Ground looks like water and sun is not AA'ed.

 

[BruceJohnJennerLawso]

Apologies if Im asking a really obvious question, but do I need a copy of the D3D11 terrain package intalled over the regular D3D11 package or can the terrain package work on its own? The bitbucket site appears to have changed a bit since I last checked in, and I cant seem to find any reference to the terrain branch.

 

[Face]

Apologies if Im asking a really obvious question, but do I need a copy of the D3D11 terrain package intalled over the regular D3D11 package or can the terrain package work on its own? The bitbucket site appears to have changed a bit since I last checked in, and I cant seem to find any reference to the terrain branch.

With BitBucket's (and Mercurial's) URL-driven archive-engine, it is quite trivial to get a particular version as full ZIP: http://bitbucket.org/asmi/ovp/get/D3D11Client - Terrain.zip

The part "http://bitbucket.org/asmi/ovp" is clear. It is "scheme://server/user/repo".
The part "/get/D3D11Client%20-%20Terrain.zip" is almost self-explanatory, too. It is "/command/argument" with the command being "get" and the argument "D3D11Client%20-%20Terrain.zip". The strange %20 thingies there are just URL escapes for the space character.

With this knowledge, you can get a link to every version of his repo, even using tags or bookmarks, or simple hash-codes:

• http://bitbucket.org/asmi/ovp/get/tip.zip - gets you the tip-most (what Asmi pushed last) archive.
• http://bitbucket.org/asmi/ovp/get/5583c99.zip - gets you the archive of version 5583c99 (it is the one the D3D11Client branch is at the moment of this writing).
• http://bitbucket.org/asmi/ovp/get/0.4.a.zip - gets you the archive of the tag "0.4.a".

Besides that, the new GUI lists the branches under "Downloads", but you have to select the proper tab-page "Branches" there.


regards,
Face

 

[Ripley]

But do I need to download both D3D11 and D3D11Terrain if I want the terrain??

Is the D3D11Terrain to be considered a "patch" to the plain D3D11, or are they 2 standalone versions?

Thanks.

 

[SolarLiner]

But do I need to download both D3D11 and D3D11Terrain if I want the terrain??

Is the D3D11Terrain to be considered a "patch" to the plain D3D11, or are they 2 standalone versions?

Thanks.

Both are standalone versions, but AFAIK D3D11Terrain is the version that asmi keeps updated.

 

[Face]

But do I need to download both D3D11 and D3D11Terrain if I want the terrain??

Is the D3D11Terrain to be considered a "patch" to the plain D3D11, or are they 2 standalone versions?

Thanks.

Normally, such repositories always contain everything you need to run besides prerequisites like DX itself or C runtimes. It makes no real sense to create a branch that only contains patches to another branch. To me it looks like Asmi is using it normally.

 

[asmi]

I'm really impressed, it looks wonderful :thumbup:

But:

Ground looks like water and sun is not AA'ed.

It's not and it won't be - that's how atmosphere renderer works. Make Sun a normal size and re-check.

---------- Post added at 09:22 ---------- Previous post was at 09:21 ----------

Apologies if Im asking a really obvious question, but do I need a copy of the D3D11 terrain package intalled over the regular D3D11 package or can the terrain package work on its own? The bitbucket site appears to have changed a bit since I last checked in, and I cant seem to find any reference to the terrain branch.

No you don't, either branch is all-inclusive.

 

[Zachstar]

Has there been any more discussion on the idea of a standardized Orbiter terrain and collision system by chance?

 

[BruceJohnJennerLawso]

Has there been any more discussion on the idea of a standardized Orbiter terrain and collision system by chance?

Anything down that alley so far is either Artlavs work or Dumbo2007s thing. Artlav has created Orulex to create procedural terrain, Meshland to support mesh collisions (and possibly vessel mesh-Orulex terrain collisions), Meshland apparently not working in Orbiter 2010 (I think). Dumbo2007 is working on integrating an external physics engine, Bullet, to provide proper orulex collisions. He appears to have made some terrific progress but hes only one man, so if you really want 3d terrain, go over to his site at http://code.google.com/p/surface-physics/, or PM him offering to help test.

 

[Zachstar]

I am referring to the discussion between asmi and Martins in this post some time ago.

http://orbiter-forum.com/showthread.php?p=358056&postcount=53

 

[BruceJohnJennerLawso]

I am referring to the discussion between asmi and Martins in this post some time ago.

http://orbiter-forum.com/showthread.php?p=358056&postcount=53

I suspected as much, but I wanted to fill in what I knew. Suffice to say that it doesnt sound like complete collision detection is likely to happen through the Orbiter core for quite a while. :shrug:

 

[Deep Blue]

There seems to be a bug on drawing the planet surface tiles. It looks like it's on all the planets, but its most noticeable on the moon. This only happens when I use a higher level texture, and I am currently using
http://www.orbiter-forum.com/showthread.php?t=26097

https://dl.dropbox.com/u/6919533/pics/tile problem.jpg

Frame 1: Everything looking fine
Frame 2: Can see individual tiles, with a weird shading error
Frame 3: Can see it displaying what looks like two different textures
Frame 4: Can see a seam, with what looks like a flipped normal
Frame 5: Life returns to normal.

Wince poking around a bit more, I discovered it only happens when there is the level 8 normal.

 

[Frogisis]

^ I'm getting the exact same thing, but since I can't seem to find the normals online anymore when checking for a new version, I'm guessing they were taken down because they were obsolete in some way.

 

[asmi]

Glider is currently working on terrain engine, so expect changes soon

 

[SolarLiner]

I re-post the cloud shaders problems, because I (may be) found the problem:

(On the first image we can see the clouds dark on the lighted surface. The Venus one is visually better: the clouds are totally opaque, so we see the bug clearly)

I think the light source for the planets is the Sun itself (well, a VECTOR3 point as the center of the Sun) but the light source of the clouds are not parented to the Sun at all, and it is a non-moving point.

I hope you understood me :lol:
Anyway, your D3D11 is great anyway !

 

[Zachstar]

Glider is currently working on terrain engine, so expect changes soon

Looking forward to this! And I now am fully DX11 as well!

 

[Poscik]

Hi, have some problem when using "Compute shaders" mode.


Also when using Pixel Shader mode, atmosphere looks little bit scary...



Anyone got this?

I have GeForce GTS450, and if It is important Pentium D 925 (yeah, still ;d). Forceware is 301.42, Win7 x64.

[EdiT]
Checked also on i5, GTX660. The same happens.

 

[Face]

I've uploaded these and some other required files. Please re-download and let me know if there is anything still missing...

Just wanted to confirm that a32935c8d6b3a is building here. It took me so long because it refuses to build with VS2010 (although I've tried changing toolset and whatnot), so I had to install VS2012.