New Release D3D9Client Development

[jarmonik]

Although I think the micro-textures on the moon do not show up here.
All needed .dds files are in Textures folder[*], so that should be O.K. Have I missed a switch in the options dialog?

Did you forget MicroTex.cfg file ?

 

[n72.75]

Pardon my ignorance, but what was the impetus behind the atmosphere remake initiative?

 

[jarmonik]

There is one small glitch I noticed: The D3D9 Atmospheric Controls dialog doesn't allow me to switch between Surface / Low orbit / High orbit configurations on the Moon (scenario Delta-glider | Brighton beach). It always shows the Surface configuration. It works fine for Earth and Mars.
I noticed this because I wanted to edit my configurations for the Moon. Similar to what I reported earlier for Earth, there seems to be a very pronounced and very rapid change in albedo for the Moon, from very brightly lit with the camera close to the surface, to very dark at about 10km altitude and above.

The "surface / low orbit / high orbit" configuration is available only if there's atmosphere. I thought that non atmospheric planets/moons wouldn't benefit from it.
I haven't seen that kind of albedo change on any non-atmospheric planet or moon. It remains very much constant with and without micro textures. If micro textures are not installed and the shader uses them anyway then it might create near white surface at low proximity. That could explain it, I'll have to check that....

 

[martins]

In case it helps, this is what I am seeing (left at 3.5km altitude, right at 10km):

 

[jarmonik]

In case it helps, this is what I am seeing (left at 3.5km altitude, right at 10km):

This issue should be fixed now. It was a micro texture problem.

 

[martins]

Thanks for the quick fix! I can confirm that with the latest sources the albedo is now consistent across all altitudes.

 

[kuddel]

Did you forget MicroTex.cfg file ?

No, it's present in Config folder.
It is deployed during install step!
Shouldn't the Textures itself also be in the repository then? Mine were from ZIP and those might not be compatible anymore?

MD5 sums of the files I have:

a997077cfb13c9c1554fadeac23ee7ad *D3D9Mars_A.dds
c3e408df5385f97e9c3daaaa976c12fb *D3D9Mars_B.dds
7eda9dc168dbe2fa41917019742e4ae8 *D3D9Moon_A.dds
6f2c6d4fd4e4cab44b012b9b44c816f7 *D3D9Moon_B.dds
a75c60afa7d4a374cf4dbad6f67667cd *D3D9Moon_C.dds
49a465848e213cb668a3fc1785cb9557 *D3D9Moon_C2.dds

 

[kuddel]

I've checked agains the "main" branch:
The microtextures seem to be rendered, but not so 'sharp' as I would like them to be (or as I'm used to see them)

1. current "main" branch (see the "edgier" rocks in foreground):


2. current "d3d9_atmosphere_remake" branch ("soft"rendered rocks in foreground):


I've also tried to tune all the "Surface micro textures" settings, but none changed much.


Maybe this is intentional, but I would love to see the rubble again

 

[kuddel]

Sorry, but I found another issue (in d3d9_atmosphere_remake branch):

The shadows seem to behave very odd. When you run the attached scenario and rotate a bit around you will see what I mean.
Those shadows behave as if the casting body is 2D and rotation wierd...

 

[jarmonik]

Thanks for the reports, I will look into the shadow and rock edge issues. There's been no change in micro texture formats so the files you got from the zip are good. I was hoping to upgrade the textures to support color and normals, this would work better on Mars but do nothing on the Moon. But I haven't had time to prepare new textures. Including the textures in the repository is fine by me but it will increase the size by 28MB

 

[jarmonik]

I looked at the shadow problem and it also occurs in D3D7 client but not in the inline-engine. It looks like the problem could be in the Orbiter. Surface normals don't match, for some reason the normals for payload canisters are misplaced. Problem seems to be in HorizonInvRot(). So, it might be better to hand this issue to @martins

VECTOR3 hn, hnp = vessel->GetSurfaceNormal();
vessel->HorizonInvRot(hnp, hn);
D3D9DebugLogVec(name, FVECTOR4(hn, 0));

 

[jarmonik]

The problem mentioned above is tied to vessel status. As long as it remains "landed" HorizonInvRot() is incorrect for attachments but becomes correct when vessel status changes to "active".

 

[jarmonik]

The surface micro texture problem should be fixed now. I noticed that due to ambient light, terrain slopes facing to the sun or away from it has very little difference in lighting. I have to do something about that.

 

[kuddel]

The surface micro texture problem should be fixed now. I noticed that due to ambient light, terrain slopes facing to the sun or away from it has very little difference in lighting.

Thank you,
that's the way I used to see my pebbles

 

[kuddel]

[...]So, it might be better to hand this issue to @martins

Done: https://github.com/orbitersim/orbiter/issues/328

 

[jarmonik]

Atmosphere remake is now merged to main.

 

[n72.75]

Atmosphere remake is now merged to main.

If you wanted to dazzle us with pretty screenshots we'd be totally cool with that

 

[n72.75]

Minor bug? The sun texture appears to be visible through Jupiter at certain distances from the vessel.
Also, the sun texture gets hidden by the DG hud glass, but seems to be okay with other transparent things.

Looks beautiful in earth orbit, by the way

 

[jarmonik]

Minor bug? The sun texture appears to be visible through Jupiter at certain distances from the vessel.
Also, the sun texture gets hidden by the DG hud glass, but seems to be okay with other transparent things.

DG HUD glass is intentional, I will look into it later on when working on virtual cockpit rendering, it's a little more problematic since the glass should change the color of the glare. Thanks for reporting Jupiter issue.

 

[DaveS]

I've done some early checking of the new atmosphere rendering and while it is miles better than what came before it, it's still not quite right. For one thing, Earth is extremely hazy and dark, even at high orbital noon. Also, it has this weird green tinge to it, and playing with the "Green" slider does to eliminate it. Right now the atmosphere appears to very sick, not letting much sun light through, very much what it would be a few minutes after the Chicxulub impact happened.

Good videos for calibration: