The FFED3DAJ Thread

This really looks great, the only reason I haven't tried this version is the download size...

Hi Zireael!

Well the original FFED3D beta 1.12++ download is only about 240MB from the elite-games Russian site. Ittiz's download is much larger as it contains videos and enhanced textures, but you don't *have to* use it if you don't want to. If you sign up with the Russian site (Google toolbar for translation in IE will help!) then once you've grabbed FFED3D, you can then add my patch which is about 2.6MB and the core-textures at 41MB. There's a 3rd optional set of example ship skins too at 15mb. I've deliberately split my patch files this way to avoid re-downloading all the files each update.
You can also find the patch files online at Space Sim Central as well as the beta in 4 7z files 1, 2, 3, 4 (be sure to have a current 7zip installed) and that's also free to join up to & download from.

Cheers!
AndyJ
 
Still making progress with the atmosphere shader. Had a bit of a busy week last week which stalled things, but today have sorted out a shader issue that was discolouring the internal sky when clouds were drawn.
Soft-coding is done but the next big problem is that the atmospheres in FFED3D were simple spheres, some of the planets that need them though are, well, rugby ball shaped!
Pretty sure that the shader only works for a sphere, so will have to figure out if I can somehow distort the end result to match the shape of the planet. :rolleyes:
 
Last edited:
AndyJ - thanks for the explanation, I guess I must have seen Ittiz's version and freaked out over the download sizes! 240 MB is manageable on my (admittedly poor) connection.
 
visual bug

Hi,
I want to report a visual bug in last FFED3DAJ_v1.09b. Below is a screenshot and a scenario. I congratulate AndyJ for the hard work in programming a better version and Zak Gordon for providing a nice texture pack for low-spec computers.

Keep going!

bug.png


Scenario: http://simutrans-germany.com/files/upload/bug.dat
 
And that's how planets are born.
LOL :D

When z-buffers go bad... planets get eaten!

unfortunately it's what can happen when two distant objects converge at a similar z-value. Where they intersect there'll be z-fighting and one will swallow the other. It's not something that's easily fixable, it's always been an issue in FFED3D and it's particularly a problem in the 3D planetary view.

I have an idea that I may be able to work-around it by disabling z-writes for far away objects now that they're depth-sorted in the next build and being drawn in back-to-near order. But it's not a solution that comes without introducing more issues!!! for example rings around planets would then draw over the top and not be clipped, as would orbital paths. And there's also a loss of sprite clipping on suns & lights.
But I'd already started to play with breaking up planetary rings into separate objects so that sections that should be behind the planet are drawn first ... if I can get that working - and if I can link orbital paths to their parent planets to get a depth - possibly this could be a way to (mostly) resolve it... if it doesn't break something else badly of course :rolleyes:
 
............
But I'd already started to play with breaking up planetary rings into separate objects so that sections that should be behind the planet are drawn first ... if I can get that working - and if I can link orbital paths to their parent planets to get a depth - possibly this could be a way to (mostly) resolve it... if it doesn't break something else badly of course :rolleyes:

That sounds like a really difficult thing to work out - good luck AndyJ! :D
 
That sounds like a really difficult thing to work out - good luck AndyJ! :D

well the rings are a combination of lots of lines to draw thin circles, and triangles that make up the thick rings. the theory is that knowing where the centre of the planet is, to split out anything with a 'z' value behind it (deeper into the screen) so that these can be drawn first, then draw the planet and afterwards draw the bits of the rings that are forward of centre. It does generate quite a few extra individual 'models' to draw but nothing that should impact on performance. needs a bit of testing to make sure I've not introduced gaps....

The planet-PacMan issue that polaris reported has now been resolved :D
(thanks for the test case!)



I'm sure that everyone will be thrilled to hear that DreamZzz had gotten in touch a little while back to offer any help I might need! Having dusted off his original FFED3D code, he's been kind enough to jump on this issue and msg me with a game variable which was unknown to me & that he thought could be tested, with the suggestion to clear the z-buffer each time that it changed. Fits perfectly with the depth-sorting render of the next build and the change seems to have resolved the issues with the system planetary view as well :D

The only issue that this alteration introduced was that the orbital paths in the system view were then drawn behind everything. This was because they are drawn separately from the other objects, and all at the same time.
But I've been able to tweak the assembly to expose which planet is the 'parent' of an orbital ring and then give these circles the same 'depth' value so that they are sorted and drawn with the parent. It's not entirely perfect - they can still draw in-front/behind the parent planets at times, but I think that this might well be a limitation of how they are drawn anyway. Closer planets/moons that are in front of inner rings will now mask them as you'd expect them to, and this is better than before and so I'm happy with that!

I think that the coding is pretty much done now for the next build, and quite glad as it has turned into a never-ending release compared to the usual once a month! (no more 'feature-creep' lol!)
nanite2000 is still working on getting the desired look in his texture pack for some new settings and which may yet also require a shader tweak or two. Also the eagle-eyed might notice that Saturn now has a glow around it. This is just a place-holder at the moment and it's a new shader that I've added around the irregular gas-giants/brown dwarfs. I'm needing now to add a few more gradient images and example shader setting values in the texture .cfg files to get the desired effects for final release... but... it's getting there!!

[edited 28/08/2018 to update image URLs to postimg.cc]
 
Last edited:
Agreed! Even if ED is on the horizon, i'm still getting time in with FFED3D+AndyJ, and will continue to do so. The improvements being made are great to see :)
 
I must admit I've gotten a little side-tracked this week checking out the Elite: Dangerous beta :p
Type-6 bought & 2mil credits, so probably just going to dip in and out now until the next beta build hits. Probably could do with earning some more credits and then settling into messing about with a Cobra, but don't really fancy the trading grind for the moment.... so... with an empty bank holiday ahead, it's back to fixing stuff!

have made few tweaks to the oxygen atmosphere shader and it's looking even better now, hugging the planet much closer. Within the atmosphere this means bluer skies and better sunsets!
Need to figure out what 'glow' colours to add to the gas giants/brown-dwarfs atmospheres now that I've created a new shader pass for these, the geometry it uses for their atmosphere is an inflated mesh of the parent planet. I realised I didn't actually need to recalculate anything and could just expand & reuse the mesh to match their deformation & ensure the atmosphere has the same 'rugby ball' shape.

Rings are now being split up as described last post, so the further side is drawn behind the gas giant and their new glow. Not quite finished this evening, but the idea appears to work as I'd hoped... just gotta dot the i's & cross the t's to finish it up. Once it's finished I'll try putting the orbital rings through the same routine to see if perhaps they'll then draw behind their parent planet when they should do.

And after playing E: D & seeing its streaks when travelling about at speed, I thought tonight that I'd take a quick look again at the issue of the missing space dust when far away from stations/planets. Success! Finally seem to have figured out what the issue was -their radii needed to be scaled up at higher speed- have apparently worked out the scale calculation and the dust seems to be appearing consistently once again :D

Next to do is to extend the external 'cloud' layer so that nanite2000 can shroud a planet with 'fog'. Everything goes through the same cloud shader pass at the moment and this makes a filled texture too opaque on the light side of the planet... well, bright white and not at all see-through :rolleyes:
I probably need to add another shader pass & extra settings in the textures file to choose which pass will apply to the cloud layer. Going to add a lower external cloud layer too, this could be used to add a cloud shadow texture.

Getting there... just a few last minute detours! :D
 
Pettytyrant will be as pleased as a group of pleased men graduating with honours at Pleased University after finding out some long lost great someone or other has left them a rather large amount of cash while winning a rollover lottery jackpot!*

*
Probably. Exact response may not be as described. You read this post at your own risk. No claim for anything can be made from reading this post.


Sounds like a good update!
 
thanks for the bug report & test case polaris,

It's actually the whole planet that's rotating rather than the texture. nanite2000 reported it back on page 3 ... way back in, errr... March... which shows how long this latest build has taken! oops! :eek:
It has been fixed however, turned out that the brown-dwarfs and large asteroids weren't initialising their rotation matrix from the Identity matrix before being drawn.

Cheers!
AndyJ
 
Sorry, I didn't know it was already reported. It's nice the next build fixes so many bugs. It would be very nice to have a chance to test it during the remaining summer holydays. :)
 
Sorry, I didn't know it was already reported. It's nice the next build fixes so many bugs. It would be very nice to have a chance to test it during the remaining summer holydays. :)
Hopefully you'll get that chance. We're so close to releasing it now. Just a few finishing touches here and there.

Here's a preview of what to expect in addition to all the fixes AndyJ has made:





Exciting or what!?
 
Top Bottom