
But illumination maps would definitely allow special one-off glowing ( highly volcanic? recently bombarded? ) planets/moons!

Moderators: winston, another_commander
Illumination ideas...Kaks wrote:But illumination maps would definitely allow special one-off glowing ( highly volcanic? recently bombarded? ) planets/moons!
Wow! Even at this early stage, these are looking great.submersible wrote:A couple of early screenshots
http://swarm.perlide.org/static/oolite-124.png
http://swarm.perlide.org/static/oolite-125.png
Happy to try - what OS are you running ? I'm building on ubuntu 686 and have never attempted cross compiling.Cmd. Cheyd wrote:Submers-
Could you compile it for me and post somewhere? I am on the road 90% of the time now, and can't install the stuff I'd need to build from source onto my laptop. Or would someone else be kind enough?
Eric Walch wrote:Just one thing: its is not PlanetEntity that handles the new_planet stuff, but OOPlanetEntity. I am completely unaware how all the classes are linked, but when the new-planets were active in the trunk builds, two years ago, I noticed they rotated around the wrong axis. Rotation axis was not through the poles. It must always have been wrong, but without white poles, you just never noticed it. I fixed that but was badly surprised that in the following Oolite release version, the rotation was wrong again. For the release, the new_planet stuff was disabled again and planets used PlanetEntity instead of OOPlanetEntity were I made my fixes. In current oolite those rotational fixes are copied to PlanetEntity.submersible wrote:One thing is fairly apparent - PlanetEntity/Drawable look like prime candidates for refactoring into several implementations of the same interface. Not my goal right now
Wish me luck
Oddly - regarding mountains , with debug build - I do see that a normal map is generated. Why those are not displayed I am unsure and still picking through the code.Kaks wrote:Well, I couldn't resist the challenge, and I did get the code to compile with NEW_PLANETS set to 1.
Still, though the seas are better than I remembered (Ahruman did some extra work on the seas which I didn't notice at the time), all the NEW_PLANETS land portion is very much featureless now, so we can't see the 3D mountain ranges anymore...
I'll look some more into the matter, hopefully I'll figure out what happened between then (2 years - & change - ago) and now.
PS: Yes, bug #5254 seems to be a typo, can't quite figure out what bug Ahruman was referring to.
PPS: The NEW_PLANETS stuff now compiles in rev4805, so if you really, really want to you can start playing with that code now!
A bunch of material defaults and shaders were removed.submersible wrote:Oddly - regarding mountains , with debug build - I do see that a normal map is generated. Why those are not displayed I am unsure and still picking through the code.Kaks wrote:Well, I couldn't resist the challenge, and I did get the code to compile with NEW_PLANETS set to 1.
Still, though the seas are better than I remembered (Ahruman did some extra work on the seas which I didn't notice at the time), all the NEW_PLANETS land portion is very much featureless now, so we can't see the 3D mountain ranges anymore...
I'll look some more into the matter, hopefully I'll figure out what happened between then (2 years - & change - ago) and now.
PS: Yes, bug #5254 seems to be a typo, can't quite figure out what bug Ahruman was referring to.
PPS: The NEW_PLANETS stuff now compiles in rev4805, so if you really, really want to you can start playing with that code now!
Code: Select all
svn merge -c -4363 svn://svn.berlios.de/oolite-linux/trunk .
Something is also happening with the current planets. At a ceirtain distance to the planet, you see a small color change. Then, when flying back it jumps back at the same distance. I never investigated deeper as its hardly noticeable, but it might be more pronounced with shaders.submersible wrote:when using the jump drive there is a 'clunk' and the lighting of the planet shifts - weird.
Probably because it is just unfinished work. Ahruman and Kaks put a lot work in it, but than stopped because we needed a 1.74 release and it was apparent that this works would never be ready in time. And than with a feature freeze for 1.75, work was also postponed.submersible wrote:Oddly - regarding mountains , with debug build - I do see that a normal map is generated. Why those are not displayed I am unsure and still picking through the code.
Could be a LOD thing? I wonder if LOD is dropped when travelling under jump drive.Eric Walch wrote:Something is also happening with the current planets. At a ceirtain distance to the planet, you see a small color change. Then, when flying back it jumps back at the same distance. I never investigated deeper as its hardly noticeable, but it might be more pronounced with shaders.submersible wrote:when using the jump drive there is a 'clunk' and the lighting of the planet shifts - weird.
However, in this particular case, restoring the vertex & fragment shaders that were removed in rev4363 does display the generated normal maps & I reckon it's got us back to at least 90% of what we had before 1.74! Those files are in trunk - once again - and the NEW_PLANETS are more or less just as I remembered them!Eric Walch wrote:Probably because it is just unfinished work.submersible wrote:Oddly - regarding mountains , with debug build - I do see that a normal map is generated. Why those are not displayed I am unsure and still picking through the code.
I have tested using materials and shaders keys in planetinfo.plist with my patches and can confirm it is still rather broken, also probably very inefficient as the dynamic generation still occurs in spite of the successful creation of an OOMaterial OOShaderMaterial via this mechanism.another_commander wrote:Not sure if this has been done already, has submersible's code on the previous page been tested by anyone yet? If it is working and does not cause problems, maybe we can include it in the mainline and add one more contributor to the list? Kaks, you seem to be the one overlooking the planets code right now, shall we go ahead?
Most welcome. I think part of the missing 10% would be to sort out which 'surface' is bumpy and which is shiny. To my eye - sometimes the ocean coloured parts are bumpy - with shiny land , other times it is the opposite.Kaks wrote:However, in this particular case, restoring the vertex & fragment shaders that were removed in rev4363 does display the generated normal maps & I reckon it's got us back to at least 90% of what we had before 1.74! Those files are in trunk - once again - and the NEW_PLANETS are more or less just as I remembered them!Eric Walch wrote:Probably because it is just unfinished work.submersible wrote:Oddly - regarding mountains , with debug build - I do see that a normal map is generated. Why those are not displayed I am unsure and still picking through the code.
Thank you submersible, very nice find!