Hard Light Productions Forums
Modding, Mission Design, and Coding => FS2 Open Coding - The Source Code Project (SCP) => Topic started by: Petrarch of the VBB on March 19, 2006, 11:56:41 am
-
I'm not sure whether this should go in here, or in the Upgrade Project folder, (or even on Mantis), but here it is anyway.
The port-side turret on the Fenris and the Leviathan (Turret07), is all strange. That is to say, the targeting box for it is a hell of a lot bigger than it should be, and is centred in the wrong place. All the other turrets are fine, it's just this one.
(http://www.gbev.co.uk/gubbins/FenrisTurret.jpg)
I'm using the 2006-03-06 builds of Fred and FS2, and the, um, latest Media VPs, I think. mv_models is dated December 29th, anyway. Dunno if there's been a release since then. So yeah. It's a problem
And apologies if it's already been mentioned. Further apologies if it's already been fixed.
-
Should be in FSPort and maybe Mantis. I'd bet money that the problem lies in the centering of the turret. One corner is the geometry, the other corner is the physical center of the turret itself. If all it affects are bounding box issues then I could probably fix it manually, but unfortunately it probably also messes up missile targeting and the lead indicator. It would need to be recompiled to be fixed, and I'm not sure who to talk to for the original model.
-
Hmm. I'll put it on Mantis.
And while we're at it. Did anyone ever fix the turrets on the AWACS ship? They were completely messed up, iirc.
-
The FSPort doesn't change the Fenris/Leviathan model, only the textures. That looks like the hi-poly model anyway.
-
I never said anything about the Port.
Port side, yes. As in, not starboard.
-
I knew what you meant. I was responding to StratComm. :)
-
Good
-
I knew what you meant. I was responding to StratComm. :)
And I was being an idiot. I meant to say FSUpgrade ;)
VA was the last person to work on the LeviaFenris, wasn't he? I may bug him later for the source file so that I can correct that bug.
-
Nyargh! turret 07's attached light never got moved from the centre of the scene (PCS's pof>cob bug), thus the turrets bounding box is huge. :(
Well, while I fix it and make the thing much more efficient by merging maps as best I can, can anyone else point out other problems the leviafenris has?
-
I've had that happen more times than I can count.
I'm going to suggest that you perhaps try to merge the two trim maps into one. They are both tiled along one axis only, if memory serves, so you should be able to tweak their UV space so that the same map can be used for both. It's a fairly minor concern though.
-
I'm hoping to merge them onto the main map actually. Failing that, that's what I'll do though. :)
I'll also be fixing the lods in accordance with Taylors guide, so they use mipmaps of the main LOD.
-
Ok, this should be the final version of the fenris. It's considerably more efficient than any previous ones in that it only uses 1 texture for the main hull ( + nameplate) and all lods combined. It should also have all lower lod rotating subsystems working properly, and I've gone over and fine-tuned a few areas of the texture - lining up windows to decks and other stuff I should have done to begin with.
Oh, and the textures are now all DDS dxt3 with full mipmaps. :)
http://webzoom.freewebs.com/twisted-infinities-va/HTL%2DFenris/HTL%5FFenris%5FHigh%5FFinal.zip
Note: everything but the debris texture has been modified in some way in the above zip, so the old stuff will need to be pulled out of the mediaVPs or you'll get weirdness with texture alignments - especially around the rear windows as the current media VP version has (older texture with the newer model, so the glow spilled out in a stupid looking manner :\ )
Also, this version is not yet compatible with Lightys specific Levvy/Fenris maps for the same reason as above. I'll fix his maps up to match the current model asap.
If anyone finds any _more_ bugs with it now, please don't tell me, or I'll have to shoot myself. ;)
-
Oh, and the textures are now all DDS dxt3 with full mipmaps. :)
You aren't using alpha transparency in any of the maps so you don't need DXT3. Change them to DXT1 and they will use half the memory.
EDIT: And please, please, please don't shoot yourself, since this is a MASSIVE improvement over all previous versions of this model, but...
[attachment deleted by admin]
-
:blah: :sigh:
*bang*
I know what caused this - when I pulled the model out of the heirarchy to fix bits of it, and stuck it back in, the nameplate object suddenly became top of the pile, and so is apparently rendered before the main hull. So it'll take a 6th recompile just to fix about the only feature I didn't test with this version. :(
And ok - switched over to dxt1, but does env-mapping (when it works) default to the actual shinemap itself if it finds no alpha-map there? Or does it just turn off if it fails to find one? 'Cos I reckon the actual shinemap itself seems about right for an envmap, thus saving the need for a separate one created once env-mapping works properly. :)
-
Ok, everything seems to be working....except......what in the world is
this stuff? (http://webzoom.freewebs.com/twisted-infinities-va/HTL%2DFenris/GTC%5FTodlerPic.jpg)
This is what I get on a fenris with no nameplate - thus I assume it's using the 1x pixel big nameplate.dds in the media vps, so how in the world does it manage to get the above? :\
Edit: it definitely doesn't happen if a propper nameplate is defined, so I'm guessing it must be a bug to do with the nameplate.dds texture
-
Bleh, triple post, but meh. There's a reason for it. ;)
Assuming I haven't missed anything further, these ones should be the final versions. :nervous:
http://webzoom.freewebs.com/twisted-infinities-va/HTL%2DFenris/HTL%5FFenris%5FHigh%5FFinal.zip
http://webzoom.freewebs.com/twisted-infinities-va/HTL%2DFenris/HTL%5FFenANDLev%5FLs%5FFinal.zip
-
Edit: it definitely doesn't happen if a propper nameplate is defined, so I'm guessing it must be a bug to do with the nameplate.dds texture
There is a (closed) Mantis bug on this. It's basically the distortion from compression artifacts that are being enlarged to fit where the texture is being used. The blank nameplate.dds just needs to be a real size rather than trying to make it real small like that. WMC knows about it so I assume that it will be fixed with the next MediaVPs.
-
This is (obviously) a stop gap solution, but if you make a 0% alpha, 2X2 dds image and save it as nameplate.dds under data/maps, then the colors go away.