Flowing robes?

[RepJunkieJr]

So, think about this for a moment. Would it be possible to make a new skeleton, exactly like the original, but add robe bones, then make new animations for said bones, and apply these animations and skeleton to a model that uses a robe? Take the picture for example. The robe is stiff ingame and it gets kind of annoying after a while. I remember looking at a Galen Marek skin that changed the animations for that specific skin, so essentially, this would be the same thing except you're adding a skeleton as well. So, thoughts? Comments? Snide remarks?

[Noodle]

It's totally doable but it'd take an enormous amount of work so nobody has done it yet. 

[General Howard]

It’s such a shame that the Quake3 engine can’t handle cloth physics. It would totally finish and polish off so many characters that have caps/cloaks or any type of material that’s free-flowing (or should be free-flowing).

[RepJunkieJr]

I was actually thinking of putting a team together to do this, but I don't know who would actually want to do this, seeing as how it would take so much time and effort.

[Ramikad]

Actually Psyk0Sith and DT85 did something about this, but I think the idea has been abandoned.

 

 

Basically DT created cloth animations for a modified _humanoid skeleton with cloth bones which was imported with the old _humanoid animations.

[RepJunkieJr]

Yes! This is what I'm looking for! If we can just pick this back up and work on it, that'd be awesome!

[Lancelot]

@

You should contact Psyk0sith or DT85 (or both) if you really want to do it. If you're lucky, maybe they can hand you the necessary files.

And look out for experienced modders/coders that can help you with that. A project like this one is pretty ambitious and requires a lot of teamwork.

[katanamaru]

Base JA has about 24,000 frames of animations. You're going to have to animate and test each of those sequences out.

 

Maybe 26,000 frames. I forget.

[Psyk0Sith]

My approach was to get the GLA sequence into 3Ds Max and create a cloth sim, i hit a roadblock with the first test, it borked the torso's animations. Possibly because i didn't use the source files and didn't shift the root pose at frame 0. Another reason is that autodesk are a bunch of retards and broke the cloth sim in max 2010, my original sim was done in max 9 but i didn't bother reinstalling when i moved to win7. I didn't even get to adding cloth bones and do a test, so chances of that workflow working are slim. I would expect the skeleton to get rejected without code changes.

[AshuraDX]

I would expect the skeleton to get rejected without code changes.

It will be rejected in multiplayer, the devs added a .gla lock for characters for some reason.

Will be fine for singleplayer though

[RepJunkieJr]

It will be rejected in multiplayer, the devs added a .gla lock for characters for some reason.

Will be fine for singleplayer though

There's no way to bypass this?

 

EDIT: My whole intention is to add flare to the MP version of the game. Sure, it'd be nice if we could do this for SP, but I'd rather do this with MP.

[AshuraDX]

There's no way to bypass this?

 

EDIT: My whole intention is to add flare to the MP version of the game. Sure, it'd be nice if we could do this for SP, but I'd rather do this with MP.

Not without code changes.

And that undertaking would require everyone you play with to use that code mod.

[RepJunkieJr]

That undertaking would require everyone you play with to use that code mod.

That shouldn't be much of a problem.

[Tempust85]

Shouldn't require any code change, unless you want to add bone remapping (like jk2 models use in jka) for base jka models to work with the new gla (just wouldn't be able to use the cloth bones without a model re-rig).

 

I've done tests in SP and the new gla works fine - no reason why it wouldn't.

 

But yeah - I get shit results with clothsim (I use max 2014) and was hand-animating. There are just over 1,000 animations that need to be imported onto a skeleton with cloth bones, then the cloth bones need to be animated. In short, it's a massive task.

 

On the subject of physics engine - there's a WIP implementation of bullet in jaownt. Those guys abandoned it but I'd love to see someone pick this up and finish it. Bullet can do both rigid and soft body physics.

[AshuraDX]

Shouldn't require any code change, unless you want to add bone remapping (like jk2 models use in jka) for base jka models to work with the new gla (just wouldn't be able to use the cloth bones without a model re-rig).

 

I've done tests in SP and the new gla works fine - no reason why it wouldn't.

 

But yeah - I get shit results with clothsim (I use max 2014) and was hand-animating. There are just over 1,000 animations that need to be imported onto a skeleton with cloth bones, then the cloth bones need to be animated. In short, it's a massive task.

 

On the subject of physics engine - there's a WIP implementation of bullet in jaownt. Those guys abandoned it but I'd love to see someone pick this up and finish it. Bullet can do both rigid and soft body physics.

sure, no code changes required for SP

 

But in Base JAMP models that do not use the _humanoid.gla can not be chosen as your character model. It throws some warning/error message and defaults you to kyle instead iirc

[Tempust85]

I know but if you replace the base _humanoid.gla and re-rig models to update their bone order (to reflect the bone order of the new GLA), then it will work without a code change.

[AshuraDX]

I know but if you replace the base _humanoid.gla and re-rig models to update their bone order (to reflect the bone order of the new GLA), then it will work without a code change.

While that works it is a ridiculous amount of effort

[Archangel35757]

I'm wondering if it would be viable to have Bolt-On GLM/GLA models for things like armor, capes, robes, etc.?

 

It would require a coding change to allow bolting on of robes, cloaks, capes, additional armor, etc. and would also require having the base humanoid.gla to tell the bolt-on models what anim to play. So they would have to be animated 1-to-1 to match the base anims.

 

What do you think? Seems like then you would not have the humanoid gla burdened with ancilliary bones...

 

@@Raz0r, @@redsaurus

@@eezstreet, @@DT85

[Tempust85]

While that works it is a ridiculous amount of effort

 

Well duh

 

 

@@Archangel35757

 

You'd still be exporting out 1,000+ anims for this, but anyway.

 

Afaik, we would need to make sure the bolt-on models have all the animation lengths matching the base GLA. We could also get any bolt-ons to use the base GLA's animation.cfg seeing as they would have to match anyway to avoid visual glitches, etc. I *think* SoF2 does something like this for their viewmodels so they could have different arm meshes.

[McGroose]

 

Does nobody care about how this is creature is an abomination

[JediBantha]

 

"Skywalker, what are you doing?!"

[Tompa9]

Any progress on flowing robes?

[ChalklYne]

About the animating 12000+ frames.... it's not as bad as it seems. A huge majority of those anims aren't different, just tweaks of animations.cfg. I got into a nice work flow to streamline them also. I've done every saber anim for 1 style in 1 day. What I do... and try to follow me here... I start with Tr_BL. I animate the character going from that particular saber stance to the top right... then animate the swing to the bottom left... then animate a return to the stance. Then go back to about halfway through the swing and animate a lock and a bounce back... then a win and a couple loses for the lock... then I save that as TR_BL_Set. This new set now contains a ton of anims. I'm recalling from memory... but I believe that one set will take care of s1_tr... s1_tr_bl... r1_tr_bl... lk_tr_bl... the lock... the wins and loses of the saber lock... and whatever the bounce was I forget... now... you make a "set" just like you did with Tr _Bl for each of the swing angles... i.e. t_b... tl_br... r_l... l_r... br _ tl and bl_tr... then you have a ton done. Now go through and link your tr _ bl to tl_br... then link tr _ bl to l_r and etc. Then repeat for each swing. Each swing must be animated to swap to a different angle in the beginning. These will be your t_tr _ tl or "transitions". Biggest headache is just the anims used probably once in the game. I.e. cinematic stuff. The force powers are simple.. the navigational anims are self explanitory... so before we go shooting down everyone who wishes to try to reanimate... let's be realistic. You could get a perfectly fine model even without the cinematic stuff in a few days with some knowledge of your programs animation layers or whatever. You would be hand animating like 40-50 or so anims... and breaking them down into a shit load of anims with animations.cfg... not exactly rocket science and it has been working flawlessly for me

[Noodle]

About the animating 12000+ frames.... it's not as bad as it seems. A huge majority of those anims aren't different, just tweaks of animations.cfg. I got into a nice work flow to streamline them also. I've done every saber anim for 1 style in 1 day. What I do... and try to follow me here... I start with Tr_BL. I animate the character going from that particular saber stance to the top right... then animate the swing to the bottom left... then animate a return to the stance. Then go back to about halfway through the swing and animate a lock and a bounce back... then a win and a couple loses for the lock... then I save that as TR_BL_Set. This new set now contains a ton of anims. I'm recalling from memory... but I believe that one set will take care of s1_tr... s1_tr_bl... r1_tr_bl... lk_tr_bl... the lock... the wins and loses of the saber lock... and whatever the bounce was I forget... now... you make a "set" just like you did with Tr _Bl for each of the swing angles... i.e. t_b... tl_br... r_l... l_r... br _ tl and bl_tr... then you have a ton done. Now go through and link your tr _ bl to tl_br... then link tr _ bl to l_r and etc. Then repeat for each swing. Each swing must be animated to swap to a different angle in the beginning. These will be your t_tr _ tl or "transitions". Biggest headache is just the anims used probably once in the game. I.e. cinematic stuff. The force powers are simple.. the navigational anims are self explanitory... so before we go shooting down everyone who wishes to try to reanimate... let's be realistic. You could get a perfectly fine model even without the cinematic stuff in a few days with some knowledge of your programs animation layers or whatever. You would be hand animating like 40-50 or so anims... and breaking them down into a shit load of anims with animations.cfg... not exactly rocket science and it has been working flawlessly for me

 

I agree. I've been doing some hand-crafted animation recently and the only thing that's really giving me issues are the things that require IK constraints like walking and such. It takes time, but not an eternity like some would believe. 

[Tempust85]

It's not that bad, just time consuming. Process being: import the existing 1,000+ anims one-by-one onto a new skeleton, animate then export.

 

If you're going to do capes though may as well do hair, tabard "flaps", lekku, tits, add back all 5 fingers, add back toes and proper facial bones while you're at it.