Ogre Forums

Seems there's a lot of issues to look at right now O_o

Screenshot from CaelumTest1 with opengl, but it could be driver issue.

I just spotted this:
http://www.sundog-soft.com/

Not meant to discourage or anything!
But those gourgeous screenshots could serve as inspiration!

I just downloaded Caelum and will take it for spin!

Uh, strange hole, as it's in one of the latitudinal segments, not a cap (which excludes I forgot a segment when creating the sphere ) Try to update to the latest drivers if possible. Have you experienced this sometime else?

I tested little software lastnight, which modifies desktop version of drivers to work with laptops (pressed OK for reboot and covered my eyes and waited for blue smoke or black screen, but it worked ) so I'm running latest drivers now.

Those holes appears randomly anywhere when look around. But it works just fine with DX, and that's enought for me.

No-one else tested with ATI yet?

Edit: they appear just in skydome not nowhere else

jacmoe wrote:I just spotted this:
http://www.sundog-soft.com/

Not meant to discourage or anything!
But those gourgeous screenshots could serve as inspiration!

No worries, I was aware of it already, but don't have the time to work on it for now

Just letting you know I got caelum working with DirectX9 by setting *both* the PS and VS requirements to 2.0.

ie:

fp->setParameter ("profiles", "ps_2_0 arbfp1");

vp->setParameter ("profiles", "vs_2_0 arbvp1");

It looks great
Looking forward to seeing a new build with clouds

very cool!! great work!

Thanks a lot guys I'm dying to get some time to work on it again

Little addition to caelum.

klik

I'll send diffs to you Kencho.

Edit: forgot to apply rotations, now moon is facing in right direction

New version available!
http://es.geocities.com/kenchoweb/Caelum-2006-10-22.zip

I've done some changes blindly (can't test them in Linux, and can't use Windows right now ) so I would thank any spotted issue I've accidentally created today.
By the way, the clouds might work. You can play with them as much as you want. The aim of this release is to add some other features and bugfixes, so clouds aren't that important. Hope you like it (and everything runs fine)

PS: Don't hesitate to play with some of the new features as using the CaelumListener (you have an example in the clouds updater in the sample) or using a nice space skybox for the starfields throught Ogre's SkyBox. They have lots of chances in my honest opinion

LayeredCloudFP is trying to use invalid temporary register (r12), prints error for 24 times.

I'll try to find out why, but don't count on my shader skills. Time start learning something about them.

Looking forward to see it in action.

Edit: OpenGL works (with holes)

This is certainly an exciting project, Kencho. I'm downloading the new version now, so keep up the great work!

Hmm... the shader does 12 calls to Tex2D, so that might have some relation... In any case, I just found today that didn't lost the cloud shader I was working on, so I expected some kind of error. My suggestion is to simply disable the cloud layer meanwhile, and test the "final" version of the Caelum library itself...

Thanks for testing and supporting

I don't know how I missed this topic before. Maybe I was too busy pressing 'F5' in my own topic. =) Anyway, very nice work Kencho. The next project I'm involved with that has a visible sky will definitely include this.

eih, really cool

Ite remembers me the A-10 cuba sky, wich was simple, but very realistic at the same time.

I woudl like just to make a proposal to have the availability of placing a directional light where the sun is, and make its target the point you specify byt giving the coordinates (real world coordinates I mean) something like UTM coords.

I thinks that with max comes something similar, but haven´t tried yet.

In my opinion this would be nice for using it in architectural visualizations, so that the real impact of the shadows over the enviroment can be analized for the differente hours of the day or even the part of the year.

Dont know if this is extremely difficult or easy, just a proposal.

bye_ _ _

@johnpus: Thanks, glad you like it

@paulov: Yes, real sun position is on my to-do list already

Kencho:

In order for it to compile as it should in linux line 8 in CaelumPrerequisites.h needs to be changed to: Also, in order to get Dagon compatability line 13 of StoredImageSkycolourModel.cpp should read:

Have you had any problems with the source as is? About the first one, I applied a bugfix MWeb told me about, and it worked fine according to what he told.
About the second: 1<<16 = 0x100; 3<<8 = 0x30; so 1<<16|3<<8 = 0x130

I'm currently working on the sun movement so if you want to try mine Really hard to find right formulas and datas about that on the web, but my sun seems to be at the right position.


Hope I'll be able to understand and use your superb atmospheric effect...

Well, as I'm considering doing changes in the way Caelum works, I will investigate more about the sun position, so will probably interested on it

Here is my class. Still work in progress so the sources are maybe a little weird But sun position (azimuth and altitude) should be pretty ok in relation with date, hour, and latitude longitude.

http://davy.sg.free.fr/SunRise.zip

Result in Ogre http://davy.sg.free.fr/JourNuit.avi

If you find some infos on the web about how to calculate the sun position I'm interested as I'd like to verify thos calculus...

It seems to give the right results Will take a look to the source next week Thanks

Kencho wrote:Have you had any problems with the source as is? About the first one, I applied a bugfix MWeb told me about, and it worked fine according to what he told.
About the second: 1<<16 = 0x100; 3<<8 = 0x30; so 1<<16|3<<8 = 0x130

MWeb must then never have tested it; it doesn't compile as is on linux.
Check http://www.ogre3d.org/index.php?option= ... &Itemid=91 . OGRE_PLATFORM_WIN32 will always be defined, it's only OGRE_PLATFORM that will differ on different platforms.

About the second: 1<<16 = 0x10000; 3<<8 = 0x300 . So you would need to check for 0x10300.

It did compile on linux with my own change, since i just removed the definition for windows and I didn't try the new version.