Hard Light Productions Forums
Hosted Projects - Standalone => Fate of the Galaxy => Archived Star Wars Conversion Threads => Topic started by: Omniscaper on January 28, 2004, 12:38:32 pm
-
Looking sweet!!! I hope HomeOne is on the list. ::patiently waits::
I have a question. What are the current relations between the Freespace modding community and X-wing Alliance's? I recently wiped the dust off that old game and found my CD full of XWAUpgrade improvements. I know converting those would be a sinch, but I'm not so wild about their model accuracy. I saw some of the other screenshots, are they ports of their mods or cooked from scratch? If ported, did your team get the blessings of Darksaber and the rest of their gang? If so, I'm itching to work on the HomeOne model.
-
we have permission to use all the darksaber's models and the xwaup ISD.
We have already converted y-wing and are working on b-wing, r41, z95, isd, tb. But since I'm not looking at this personally I'm not sure what's their status. If you want to help us in this we can surely use an additional hand.
BTW the problem is that they doesn't work always in freespace2 (too intersections) and they require usually lots of work since they have dozens and dozens of texture with a max size of 256 pixels.
For example, their corvette will use my textures, and I hoped to use their model as high poly lod, but unfourtunately the intersection cause too many collision problems:(
-
Maybe those many intersections cause my R-41 model to crash while converting from TS->PCS.
Do I Object Union those intersections?
Stupid question, I frist try and come back with the results.
-
no way.
just be sure to have the lastest PCS (autofacet build)
it may take a while to convert but it will convert (99% of the times), wait for at least 5 minutes before closing it.
also I think sigma was giving a look at r-41 (if it isn't the same).
If you want to work on something new, give a try at z-95 since I suppose sigma has to already start it.
If we have to go with object union, I prefer to make a whole new model from scratch
-
I got around the intersection problem by seperating meshes from the original before exporting it to Truespace. The Galaxy Class was a nightmare to convert until I realized that seperating the multiple intersecting meshes that were considered "single". Once the seperations were made and turned into subojects in Truespace, PCS would convert without crashing in less than 8 seconds for a 15k mesh. I'm currently downloading Home One for a test conversion. I'll beef up its existing textures. Anyone working on Home One? I don't wanna undermine someone else's hard work and deserved glory!
In regards to texture sizes, is 256 the top size you want? I"m already using textures as high as 4096. Lightwave has this great tool to create a single texture merging all the textures through it's texture unwrap.
-
problem is that we can't use models w/o prior permission, and I don't know where this home one come from.
But If you can find the permission to use she and if the model is good enough (do you have some pic?) and not too high in pcount (I don't want a 20k mesh) we'll surely be happy to use she.
about intersections:
I had no problems converting intersecting meshes as one single subobject so far.
I never got PCS crashing doing this, and limiting the number of subobjects is the key to optimize for HTL.
PCS just takes an helluva time to convert a cob whenever a single object is very complex, but that's normal.
But I admit that I never converted a ship of the complexity of those from the bridge commander modding community, so it may be very possible that those kind of models make PCS to crash. Also remember that some gamecrashes with subobjects of more that 7-8k have been reported
About the texture size, go for 1024x1024.
-
I splitted the thread in two since the discussion derailed from the Assault Frigate
-
JMs Homeone by the sound of it. Only one for xwa anyways. 6kish iirc and 40 odd textures, only 2 are 256*256 though, the rest are smaller. Probably have some space left over on a single 1024*1024 texture.
-
It is JM's but I intend to make my own so no permissions are needed. I just needed a 3d reference to see how it was created. I'm upping the polycount to 10k since it is an uber capship and work on its textures since its pretty much homogenous tiling with some areas of color.
-
Omni, any chance you could post instructions on how to combine the textures? I have lw7 but the manual its very specific on how to do it, and i'm not very familiar with the prog.
-
http://rds.yahoo.com/S=2766679/K=lightwave+texture+baking/v=2/SID=e/l=WS1/R=1/H=0/*-http://www.newtek.com/products/lightwave/tutorials/surface/baking/
-
omni, if you want I can ask to give you private forum access, we have many reference links and it's faster this way rather than copying the links here:)
about pcount, 10k triangles is still ok, just don't go too over.
Since you'll make the model from scratch, unless you intend to make it as a single solid object, it should be easyer to avoid big, weird intersections (minors are usually ok).
Don't use too many submodels per subobjects (I got crashes when going over 130....). This way you should be able to have a model not crashing pcs/game
-
btw...
http://www.theforce.net/swtc/links.html#models
http://www.theforce.net/swtc/mcc.html#home1
-
Let me try a test conversion using JM's Homeone first. Of course it won't be released or even used, so permissions are not needed. But so far as I've seen, Its gonna be a toughy making the HomeOne NOT go over 10k. Theres a heck of a number of rounded edges. JM's looks a little sloppy with its polygon count so low.
Private forum access?? Um, sure if its easier to talk that way. Keep things on the downlow.
-
ohh omni, c'mon:)
we are making game models, not CG ones. We have to be efficent, avoiding useless details and oversmoothing. 10k is a lot, and is much more that the average of even actual commercial games...
-
::conceeds::
Ok then, a 10k cap it is. Homeone will definitely not be a single mesh then. I am already forseeing at least 15 subobjects. JM's Homeone was quite helpfull as well as your links. It really is a simple vessel to build without the 10k cap =). I just gotta find a balance between aesthetic and practical.
-
I'll not mind if it will be (little) more or less polys, I'll complain if the model will be too polywasting. Just use 10k as reference.
As I said I got in game models with more than 100 submodels, some intersecting, converted as one subobject, so 15 submodels shouldn't cause too problems.
just limit the intersections, trying to have only very simple ones (like a bulge with the base inside the hull) and not too complex ones (like multiple intersections in the same area).
I suppose that the problems you got converting multiple models as one subobject were caused by too complex intersections. Maybe this time you'll be more lucky.
Also remember that intersection MAY cause the model to be nonsolid (you can fly through the ship).
Another good reason to limit them only when they are really necessary= when they let you save lots of polys
-
My first successfull conversion of JM's Homeone. It consists of 31 subobjects. Non of the objects were joined and kept seperate. All objects are also solid. I'm gonna start retexturing and remodeling. Should I merge alot of these "bulges". That would increase polygon counts. JM's HomeOne has 5400 polys, so I'll try to border on 10k since it is huge.
(http://www.penguinbomb.com/omniscaper/Screenshots/Jm_Homeone.JPG)
-
that's my 2 cents (you are doing the model so you decide how to make it):
keep em separated till the end to reduce pcount.
If the model will work converting all of them as a single object, then better, elseway you can try making different subobjects if you will perform the same crashes you got with some ST ships.
If you'll still have problems (flythroughs), you may start to join some of them, converting after each join to test the mesh, and stop when the problem will be solved.
-
Here is my first draft of Homeone using generic textures. Its Polygon count is still pretty high, 20k. Reducing as I type this.
I was curious if it I could make the most detailed LOD at 15k (like the Galaxy Class) at an extreme close range, lets say 20 meters. It would only make flybys much more pretty.
(http://www.penguinbomb.com/omniscaper/WIPS/My_Homeone.JPG)
-
That's rather impressive, especially for an early version. I wish I had the tools to do something like that, considering how long that would take to do in Blender....
-
I basically made a number of half spheres, stretched and tapered where necessary, and arranged using JM's as a reference.
It was actually quite easy to make the buldges since most of them have the same shape, just different orientation. Thank god for cut and paste.
-
sorry but no:)
mostly because with 2 high poly lods out of 4 lods, it'd mean to have an unbalanced LODs succession.
BTW, it's really a nice start, but I give you a suggestion for when you work on blubby models:
Consider the main body of the ship. Look at the ventral area: it has a very wide angle, it's almost flat. You don't need more than..say 4 edges on the trasversal section for this area.
The borders instead have a narrow angle, they are more curved, and need (proportionally) more edges. 6 Edges should be enough.
That's a total of 20 edges for the trasversal section. Want to oversmooth? go for 24 if your pcount isn't already too high.
This mean to not only deform a sphere or a rounded cylinder, but also to control the edge density, smoothing only where needed.
I don't know how many edges you used (but to have 20 k you probably used a lot of them)
It also seem from your pic that the longitude of the various elements is more rounded (proportionally) than the latitude (the number of trasversal elements).
But I may be wrong because it's hard to judge a rounded model with smooth shadings from a perspective view.
-
Ok Homeone is gonna take MUCH more work than I expected. I'm still focused on my Trek conversions. Would you folks like my current conversion of JM's Homeone? Maybe it will be of some use to you folks. Or does someone have it already?
I'm gonna keep working on my version of Homeone but it will be second on my priority list. It is a favorite of mine so its definitly gonna be a WIP.
-
ok, no hurry for homeone.
meanwhile we may use that JM model, if it's good (never really looked at it) and if JM will give us permission.
Don't post the link, give me it with a p.m
thanks:)
-
Making some good progress there, Omni. Keep it up. :):yes: