Русский New site

Advanced search

[ New messages · Forum rules · Members ]
Page 20 of 70«1218192021226970»
Forum » SpaceEngine » Archive » Work progress - 0.9.7.2
Work progress - 0.9.7.2
AzirphaeliDate: Wednesday, 11.06.2014, 15:39 | Message # 286
Space Tourist
Group: Users
United States
Messages: 32
Status: Offline
Quote Enki ()
I'd also like to bring up bio-luminescence- you could have "night lights" from planets covered with forests of bio-luminescent plants, etc.


The problem I foresee with this is that as soon as you dip down to planet level you'd have floating lights attached to nothing.. or just pasted onto the terrain as a texture/shader. Until we have procedural structures (cities/forests/etc..) the lights will look great from afar, but likely very confusing and lackluster close up.





~
 
SanoconDate: Wednesday, 11.06.2014, 17:16 | Message # 287
Space Tourist
Group: Users
United States
Messages: 33
Status: Offline
Quote SpaceEngineer ()
Look at the system/shaders/ and in the cache/shaders/. You will be surprized how many shaders are used in SE. Basic shading? Atmosphere, eclipse shadows, complex materials, water, foggy objects, black hole warp effect - all this are impossible with "basic shading".


Huh, ok makes sense. It's just right now there are no shadow casts when it comes to landscape shading (i.e. a mountain's shadow covering a valley) or spaceships casting shadows on themselves and landscapes as far as i can tell. (the spaceship thing was what made me think this in the first place...)

Edit: Reinstalled space engine due to planets not loading at all. and apparently it's only the spaceships that have no secondary shadows. I Didn't notice the shadows on the terrain because it was subtle, in a good way.


Edited by Sanocon - Wednesday, 11.06.2014, 18:15
 
Destructor1701Date: Wednesday, 11.06.2014, 23:40 | Message # 288
Pioneer
Group: Users
Ireland
Messages: 531
Status: Offline
What we have here is a confusion between the optical term "shading", and the computer term relating to shader code for calculating effects on the graphics hardware.




 
anonymousgamerDate: Thursday, 12.06.2014, 05:02 | Message # 289
World Builder
Group: Global Moderators
United States
Messages: 1011
Status: Offline
Quote Sanocon ()
I Didn't notice the shadows on the terrain because it was subtle, in a good way.


Those shadows aren't really proper shadows... it's not the landscape of the planet casting them, it's the same sphere used to cast eclipse shadows.





Desktop: FX-8350 4.0 GHz, 8 GB DDR3 RAM, EVGA GeForce GTX 1080 FTW 8 GB, 2 TB HDD, 24 inch 1920x1080 screen
Laptop: Core i5 480M 2.66 GHz (turbo 2.93), 8 GB DDR3 RAM, AMD Radeon HD 6550m 1 GB, 640 GB HDD, 17.3 inch 1600x900 screen
 
SanoconDate: Thursday, 12.06.2014, 09:22 | Message # 290
Space Tourist
Group: Users
United States
Messages: 33
Status: Offline
Quote anonymousgamer ()
Those shadows aren't really proper shadows... it's not the landscape of the planet casting them, it's the same sphere used to cast eclipse shadows.


so that exsplains why there's no shadow casts on the spaceships. also

Quote Destructor1701 ()
What we have here is a confusion between the optical term "shading", and the computer term relating to shader code for calculating effects on the graphics hardware.


there wasn't much confusion actually. when I said basic shading, I meant I thought he was using just opengl. but the shaders files prove he's using GLSL as well, he's just using it more creatively than i could actually think up.
 
SpaceEngineerDate: Thursday, 12.06.2014, 10:16 | Message # 291
Author of Space Engine
Group: Administrators
Russian Federation
Messages: 4798
Status: Offline
Actually, OpenGL can't work without shaders. This was possible 15 years ago. In modern OpenGL you must use shader even to render a single triangle.




 
SalvoDate: Thursday, 12.06.2014, 13:25 | Message # 292
Star Engineer
Group: Local Moderators
Italy
Messages: 1400
Status: Offline
Actually Real Time Shadows are different from "shadows" you see on a normal 3D objects, and they usually has nothing to do with shaders:



Real time shadows are really heavy for GPUs, and on complex objects like spaceships and landscapes they can really kill your FPS...





The universe is not required to be in perfect harmony with human ambition.

CPU: Intel Core i7 4770 GPU: ASUS Radeon R9 270 RAM: 8 GBs

(still don't know why everyone is doing this...)


Edited by Salvo - Thursday, 12.06.2014, 13:27
 
VoekoevakaDate: Thursday, 12.06.2014, 13:30 | Message # 293
World Builder
Group: SE team
France
Messages: 1016
Status: Offline
One alternative is to generate a shadow map. It could be possible with a shader.




Want some music of mine ? Please go here !

 
AerospacefagDate: Thursday, 12.06.2014, 14:13 | Message # 294
Pioneer
Group: Users
Russian Federation
Messages: 401
Status: Offline
Salvo, I guess that's a real problem, working real time shadows is as resource-consuming as landscape generating.

Voekoevaka, shadow maps are only useful for static lighting, not the dynamic one like in SE. Well, maybe in interiors of spaceships...

SpaceEngineer, so, I guess, the current lighting system is already very complex and I even noticed some bugs in it. The situation is, some places like brown dwarf systems, may sometimes have as much as 5 light sources - which is an asteroid satellite, a planet, another planet or satellite nearby, a brown dwarf itself and a second component in binary star. With this, the situation is really complex and some bugs may happen, as my observation show me. No, actually, that is not an important issue, they are not really that easy to notice most of the time, unless the major source of light disappears.

But with that in mind, any system of real dynamic mapping is out of question.

On the other hand, let's return to the asteroids. They are really small, and their surface is considerably curved, so there's always a part of the landscape on terminator, that are really , in the shadow of the mountain, and sun shouldn't be able to reach there, but they are brightly lit and very confusing to observe.

Finally, if there's a sunset on the planet, then we have similar problems - even if the sun is lower than local horizon, some surfaces remain quite bright. It looks a bit like the sky itself is shining, but I know that this system is not currently supported. In reality, the landscape at the horizon should obscure most of the light, casting long shadows, but that is not possible to imagine without dynamic lighting.

Actually, I want to clarify - it is only a suggestion, it's not like I think that you don't know about this issue, I just want to know your opinion or if you're going to resolve that somehow in the future.


Edited by Aerospacefag - Thursday, 12.06.2014, 14:16
 
SpaceEngineerDate: Thursday, 12.06.2014, 14:54 | Message # 295
Author of Space Engine
Group: Administrators
Russian Federation
Messages: 4798
Status: Offline
You all wrong smile
Shadow mapping is a technique what uses render the scene depth to a special texture (called shadow map) form the point where light source is located. After this, this texture is used in the scene render pass to determine whether pixel is in the shadow. Using of shadow map texture is only possible in the shader. So:
- shadows of course uses shaders (as I said, all render in modern OpenGL is done throug shaders);
- shadow mapping don't depend on geometry complexity, it drops fps of course but not so significantly;
- shadow mapping is useful for dynamic lights.
Aerospacefag, I guess you mix shadow mapping with lightmaps what is indeed only for static lights.

Currently in SE any dynamic object (planet or spaceship) can use up to 4 light sources. As long as SE engine uses forward rendering pipeline, its performance linear depends on number of light sources, so adding new light is a bad idea. When I rewite engine to use deferred rendering, it would be possible to use hundreds of local lights. But number of global lights (sun/moon) will still be limited, because they must use atmospheric scattering effect, what is quite complex and slow (you may notice that airless planets shows better performance).

Shadows are still don't implemented in SE. I delay it for deferred rendering too. But SE uses "fake" shadows on sunset - it is a simple calculation of occluding of the sunlight by the spherical body of a planet. It makes illusion of "incorrect shadows" at sunset, but it is still better than nothing - without this system you might see lighted mountain peaks even at night side of the planet.

SE also have its own "precise" eclipse shadow technique. Eclipse shadows are calculated directly in the planet's (and ship's!) pixel shaders ad give much more accurate and artefactless shadows (umbra and penumbra!) what shadow maps can give. It will be used together with shadow maps in the future. Shadow maps are still needed for shadows of irregular bodeis like asteroid-like moons.

Rings shadows are implemented in similar way - they computed directly in the rendering shader. Future versions will take into account rings from any planet, not only from ring host. Ie Saturn's moons will receive shadows from Saturn rings.





 
FastFourierTransformDate: Friday, 13.06.2014, 19:16 | Message # 296
Pioneer
Group: Local Moderators
Spain
Messages: 542
Status: Offline
About the lava issue, I have found a procedurally generated lava flow that looks simple and also quite realistic.

https://www.youtube.com/watch?v=YrOLPKnNe_M
 
WatsisnameDate: Friday, 13.06.2014, 23:27 | Message # 297
Galaxy Architect
Group: Global Moderators
United States
Messages: 2611
Status: Offline
I don't think that qualifies as simple -- that's a five hour rendition made using Blender. Not exactly practical in the real-time rendering system of Space Engine.

Looks nice though.





 
DoctorOfSpaceDate: Saturday, 14.06.2014, 05:06 | Message # 298
Galaxy Architect
Group: Global Moderators
Pirate
Messages: 3600
Status: Offline
Quote Watsisname ()
Looks nice though.


If you wanted any sort of fluid dynamics in SE you would need to do a system similar to From Dust and that would be difficult to code.



Maybe in the far future of SE we will see something close to this.





Intel Core i7-5820K 4.2GHz 6-Core Processor
G.Skill Ripjaws V Series 32GB (4 x 8GB) DDR4-2400 Memory
EVGA GTX 980 Ti SC 6GB
 
FastFourierTransformDate: Saturday, 14.06.2014, 13:15 | Message # 299
Pioneer
Group: Local Moderators
Spain
Messages: 542
Status: Offline
Quote Watsisname ()
I don't think that qualifies as simple -- that's a five hour rendition made using Blender.


You are totally right
 
JoeCapricornDate: Saturday, 14.06.2014, 21:29 | Message # 300
Observer
Group: Users
United States
Messages: 14
Status: Offline
Those nightside images of Io look amazing!

Could this mean we might have rocky planets that orbit so close to their parent star that the nightside is glowing from seas of molten lava? Or otherwise highly volcanic worlds?

Might that even be a new planet class? Io is designated as a selena, but it is neither a selena, icy world or desert. There could be other Ios out there, both from tidal heating and from intense stellar heat.

Are we going to see any new planet classes in 0.9.7.2? I remember reading an article about a new class of planets being discovered called Gas Dwarfs, as well as a rocky planet that is more or less a Rock Giant (or Superterra).
 
Forum » SpaceEngine » Archive » Work progress - 0.9.7.2
Page 20 of 70«1218192021226970»
Search: