Musje Posted August 2, 2019 Posted August 2, 2019 I'm trying to convert the nice ZAP HDT Ankle chain to SMP. The chain consists of a left part attached to the left ankle, a right part attached to the right ankle, and a floating middle node. The chain link nodes in the left and right parts are all in a parent-child relation. So I set up an SMP XML file to link each chain end to the ankle and the middle node. So far so good. The only problem is how SMP positions each of these parts when the item is equiped. It appears that SMP positions each part as it appears on the character rather than as it appears in the NIF. If I equip the item on a character in T-pose (similar to the pose in the NIF), the parts align correctly and appear as a chain. This also happens when I load the game with the item already equiped. However if I put the item on a character in the default stance, with one leg slightly turned and forward, the parts will appear misaligned, even though the physics still work. Is there a way to make SMP use the node positions in the NIF rather than the positions at time of equiping the item? Spoiler Spoiler Item equiped after loading game: Meh. Spoiler Loaded game with item already equiped. Yay!
t.ara Posted August 2, 2019 Posted August 2, 2019 This is not anymore relying on the nif-file and it´s set-up, it´s relying on the HDT behavior-path. So the NIF-FILE IS OKAY. The smp/HDT-path tells each boned chain-piece, where to be or where to connect next, or how the "piece" should behave. The piece in the middle is the target of both chain-parts...and target itself is a "free" part, which is depending on both last bones, the R/L 8 bone. Skyrim SE seems to have a different SMP engine-i guess it has. So from the pics you are maybe close to a result, but that is maybe depending on a different SETTING inside of the SMP-PATH....what I know about is, that SMP-paths are always edited "manually"-there´s no export-process from a 3D-engine possible compared with HDT-PE. In SKYRIM, I have never seen somebody connecting a two-side-chain with a middle piece. Two anchor points is with that older SMP (OLDRIM) not working. HYDROGENSAYSHDT could maybe illustrate some help here. Where is that processing SMP path ??? You could show me once... (I´m editing often HDT-PE paths manually-maybe I can see something there...ggg)
Musje Posted August 2, 2019 Author Posted August 2, 2019 This is the XML I use. The problem seems to be with the SMP "solver" which doesn't seem to solve for all constraints at once, but instead follows the parent-child chain and leaves it at that. ZaZAnkleChainsRagDollsSMP.xml
t.ara Posted August 2, 2019 Posted August 2, 2019 2 hours ago, Musje said: This is the XML I use. The problem seems to be with the SMP "solver" which doesn't seem to solve for all constraints at once, but instead follows the parent-child chain and leaves it at that. ZaZAnkleChainsRagDollsSMP.xml 3.04 kB · 0 downloads :-))) maybe you could give me your item idk if I have that for SE here-maybe it´s not in my zap-I do not know if my zap for se is working properly-but I want to see the effect-maybe you could send me the suiting SMP-mod with a small guide for installation? This is something like a "key"-item-work: if this works, we could also have the whole party-that´s what you mean to do here-yes?
Musje Posted August 2, 2019 Author Posted August 2, 2019 14 minutes ago, t.ara said: :-))) maybe you could give me your item idk if I have that for SE here-maybe it´s not in my zap-I do not know if my zap for se is working properly-but I want to see the effect-maybe you could send me the suiting SMP-mod with a small guide for installation? Here's the item for SE, and the SMP mod. Install the latest version of that when prompted. And don;t install the "performance" profile, it is really crap. By the way, ZAP8/9 work fine on SE after doing a standard conversion. Except the HDT doesn't work 14 minutes ago, t.ara said: This is something like a "key"-item-work: if this works, we could also have the whole party-that´s what you mean to do here-yes? Hell yes ZaZAnkleChainsRagDolls_1.nif HDT_SSE_Combined_Compatibility_Set_1.0.0.7z
t.ara Posted August 2, 2019 Posted August 2, 2019 HuHu, MUSJE:-))) ..puh, I hope I can do that manually....version is not interesting...i found a version folder....I look what I have installed in the past i use 1.5.53...for a special reason I checked all the stuff and I must say I have it so far installed-but your collection is amazing stuff:-)) Oh, did you add the ground plate to the file-I´m behind the moon when it comes to SE..believe me:-))
t.ara Posted August 2, 2019 Posted August 2, 2019 Where do I find the suiting PATH of that chain, please? I still have that old HDT path there inside. OMG forget it-anyway the SMP path is in same folder as usually-i think so-yes I have it-now I want to see it ingame so far...did you create that xml? Here i do not need to change the . with ,:-))))
Musje Posted August 2, 2019 Author Posted August 2, 2019 Yeah I made the XML. Drop it in the same folder as the NIF (Meshes / Zap folder / HDT folder) and it should work.
t.ara Posted August 2, 2019 Posted August 2, 2019 The problem seems to me quite easy: The original gamebryo-file has information inside, which is WRONG for SMP and can not being shown. The most hard problem is, that the angles(how they told it inside the 3dsmax-I have not in mind....)do not SUIT. There´s nothing really together because the SMP-file can not handle the anglulations of each part of the bones. If you begin in NIFSCOPE to change something (angles of bones or bone-length), the mesh will be tordiated to death at least. The only chance is to create a rope instead of a chain, because ropes are round and their bone-angles are not interesting (as long the item is not under TENSION !!!). SMP is describing a bone, then the next bone in attachement -then the next one....the angles of each bone are ignored: hanging down is everything for SMP to have realized. For clothing is it perfect, because meshes, which form planes will be working without angles like a flag in the wind, the mesh is holding together and turns automatically the angles of the bones correctly: hanging:-) The SMP xml file only provides the bones to be working behind each other, but the stretching is very high-so the meshes become like fluid, if the chain is put away from each other. And the weight-painting doesnt work either, because chris weigthpainted every single chain-part as an static single ring-part. If the bones get this factor of stretch, the chain is looking like cheap-plastic-stuff. The first, which would be have to be added to SMP are working bone-angular-informations. And then you need to have a connection chance-I can not see, that this item is connection really with the middle part (bone 001). That´s why I thought it´s not possible to convert HDT-PE boned-physics to a suiting SMP - item. If someone has in between this technology then sorry. I can not help as long I have no parameters inside of the XML file here. The file is build up likewise a "chain" of perl without commands for the behavior structure...that´s what HDT PE instead is offering so much perfect. (Inside of a HDT-PE - path is one bone´s info about 4-10 pages DINA 4 or longer!!) Sorry-really pity....but maybe there´s suddenly another opinion or maybe an update for SMP in future....the best would be the transformation of HDT PE into that 64 bit engine-i have NO idea !
t.ara Posted August 2, 2019 Posted August 2, 2019 28 minutes ago, Musje said: Yeah I made the XML. Drop it in the same folder as the NIF (Meshes / Zap folder / HDT folder) and it should work. WOW, you made it!!!! wow;-)))
t.ara Posted August 2, 2019 Posted August 2, 2019 Again, it´s errors are showing up, because the smp-path has not all suiting parameters inside of it...it has the connection -set-up of mostly the whole chain, but we need to get parameters for bone-angles-with such an add-on, the parts could be turned to be more suiting-and: the direction of the bones could be made also...the parameter, in which angle the following bone is sticked to the before going one....
t.ara Posted August 2, 2019 Posted August 2, 2019 Anyway I have a very fine idea (just when I was in the kitchen cooking-lol) the bone-angle-trouble, we can surround!!! Instead the complicated information, which is handled thru all the HDT havok-path, we can instead kill the whole angle-stuff away: You can keep the existing SMP path by adding a parallel bone strip: those bones will stand inside of the 3-d-room beside the other bones and will offer a BAND like a TAPE, which will be used to weightpaint a similar chain. The chain´s parts will be fixed to the bones because they now have orientation inside of the 3d-room, likewise their angles are then FIXED (like glewed to the tape). In that case would the overall behavior of that chain be similar to a "mesh-tape", which is between the legs, and that tape would also get it´s double connection to the ankles as well. If the tape has enough polygons (this amount polys of the mesh do allow the movement-the more polys, the more smooth) and if their distance is suiting to the chain-parts, the chain will behave in that case quite WELL, then. I would please you to check out with a simpler path, if you can create an end-bone to be really connected to the other ankle. If that works, I´ll create for you a double-boned and weightpainted chain, so that you can create it´s new SMP path for it. That result should be looking FAR more pretty !!!
t.ara Posted August 2, 2019 Posted August 2, 2019 I-ll prepare that file right now-remember to CONVERT that result to SE....!!! I´ll add a second bone-chain and you only have to double their names in the path....the way how much their OVERALL torsion shall be, can be GLOBALLY set up inside the SMP path-don´t ask me, which parameter that has been::---))
Musje Posted August 2, 2019 Author Posted August 2, 2019 It takes the initial angles from the NIF, which is fine. But in-game they still twist a bit. What I have now is an unhappy compromise between links that twist too much and a too rigid chain. To do a really proper job, you'd have to constrain rotation over 2 axis, and make sure the frame of reference for rotation is that of the parent node. Not sure if SMP even does that, I think rotation contraints are relative to the scene root. At least that is what it look like if I play with the rotation constraints. Another irritating limitation: the stretching of the chain is as limited as it gets with generic constraints in a constraint group. SMP supports rigid spring constraints too which will make a true non stretchy chain, however.., it appears you can only create a rigid constraint between an HDT enabled node and a fixed node, not between 2 HDT nodes. It does look like SMP is a somewhat limited system geared towards clothing animations... or I missed something in the "documentation"
Musje Posted August 2, 2019 Author Posted August 2, 2019 7 minutes ago, t.ara said: You can keep the existing SMP path by adding a parallel bone strip: those bones will stand inside of the 3-d-room beside the other bones and will offer a BAND like a TAPE, which will be used to weightpaint a similar chain. The chain´s parts will be fixed to the bones because they now have orientation inside of the 3d-room, likewise their angles are then FIXED (like glewed to the tape). In that case would the overall behavior of that chain be similar to a "mesh-tape", which is between the legs, and that tape would also get it´s double connection to the ankles as well. If the tape has enough polygons (this amount polys of the mesh do allow the movement-the more polys, the more smooth) and if their distance is suiting to the chain-parts, the chain will behave in that case quite WELL, then. I would please you to check out with a simpler path, if you can create an end-bone to be really connected to the other ankle. If that works, I´ll create for you a double-boned and weightpainted chain, so that you can create it´s new SMP path for it. That result should be looking FAR more pretty !!! That sounds like an interesting workaround, but I am not sure how to create the NIF for this. And yes we still have the endpoint problem. But keep in mind: the endpoints are connected, but there is an offset that shouldn't be there. To verify the XML, equip the chain, save, then load the save. You will see the chain working correctly. I still have no idea how to correct that offset, except perhaps by a manual parameter (which I will work on now)
27X Posted August 2, 2019 Posted August 2, 2019 <constraint-group> [node grouping] </constraint-group> 13 minutes ago, Musje said: It does look like SMP is a somewhat limited system geared towards clothing animations Not even kind of. Just because Patreon users just cannibalize xmls copied from other patreon abusers doesn't mean that's all it's good for.
t.ara Posted August 2, 2019 Posted August 2, 2019 Boning is ready....now I need to weightpaint that... I try use BodySlide..ggg all bones have now their partners, same line, added "b" at their name´s ending.
t.ara Posted August 2, 2019 Posted August 2, 2019 10 minutes ago, 27X said: <constraint-group> [node grouping] </constraint-group> Not even kind of. Just because Patreon users just cannibalize xmls copied from other patreon abusers doesn't mean that's all it's good for.
t.ara Posted August 2, 2019 Posted August 2, 2019 Maybe you simply concentrate on the connect-point-here I have my problems with that....-anyway can you later change the ending or start of bones-points.....even I have some neckchains in mind, which have no real end-connection bones. I also experimented with looping SMP....anyway seems your smp path to react somehow to that endings....whatever... OK this is a VERY quick weightpaint...ONLY FOR TESTING-! the overall behavior. Convert it before a try-if it succeeds, I can redo that more better then. Double your SMP, add the bone-amarda with always "b" ending names and parallel connective-behavior... ZaZAnkleChainsRagDolls_1b3.nif
Musje Posted August 2, 2019 Author Posted August 2, 2019 15 minutes ago, 27X said: <constraint-group> [node grouping] </constraint-group> Not even kind of. Just because Patreon users just cannibalize xmls copied from other patreon abusers doesn't mean that's all it's good for. Cannibalizing XMLs is pretty much required with the limited documentation available (and yes I have seen and read the tutorials, they cover basic stuff only). But if you have more detailed resources, I am all ears
t.ara Posted August 2, 2019 Posted August 2, 2019 34 minutes ago, Musje said: Cannibalizing XMLs is pretty much required with the limited documentation available (and yes I have seen and read the tutorials, they cover basic stuff only). But if you have more detailed resources, I am all ears Forget that...there´s enough info in those files-they do not have interest in sharing (guides). HAVOK ´s engine in itself is a much more complicated physical stuff, than this SMP ever can handle. Outdated but all fine for US here. You can rely on the HCT guide and read guides from first hand. Gamebryo has got lot of changes inside of skyrim-in the past it was a different thing. It´s everything "specialized". Our tools do not read out everything of the files....there´s some BLACK-MATTER around in the space, which we can not see:-))) "patreon" users- THERE`S sadly NOTHING MORE TO OFFER THAN WHAT I DESCRIBED more higher: HANGING DOWN
t.ara Posted August 2, 2019 Posted August 2, 2019 I started a real nonsence..every 2nd-ring got extra changes...result seems to me better than the original-chain-but anyway I have a problem with that middle part yet... ZaZAnkleChainsRagDolls_1everysecondringgota90-degree-angular-change.nif
Musje Posted August 2, 2019 Author Posted August 2, 2019 24 minutes ago, t.ara said: I started a real nonsence..every 2nd-ring got extra changes...result seems to me better than the original-chain-but anyway I have a problem with that middle part yet... ZaZAnkleChainsRagDolls_1everysecondringgota90-degree-angular-change.nif 383.83 kB · 0 downloads Still can't get that node to compute properly (it does attach, but in the wrong position as described above: the position depends on the stance of the body) I have a visual fix at least: the chain appears properly attached if your character puts it on while standing (or after zoning). But for example if you put it on while sitting down, the attachment is incorrect again Maybe this is as good as it gets. I'll give your NIF a go tomorrow
t.ara Posted August 2, 2019 Posted August 2, 2019 22 minutes ago, Musje said: Still can't get that node to compute properly (it does attach, but in the wrong position as described above: the position depends on the stance of the body) I have a visual fix at least: the chain appears properly attached if your character puts it on while standing (or after zoning). But for example if you put it on while sitting down, the attachment is incorrect again Maybe this is as good as it gets. I'll give your NIF a go tomorrow Depending on the fact, that this technique has never been made therefor, you did again a very charming work inside of that small XML file. Anyway does SMP not like the attachment to the root-node of the skeleton, because it´s a LOOP. And that loop causes this nervous and deangled zipper-movements-the zippering will shut down, if all angles are suiting...I can´t edit on a single bone two incoming /outgoing ankle-parameters-there´s only one set...maybe you could also try to connect only between the bones without the 001...they are not physically (skeleton-) connected, but you can loop them inside the XML file-please also check that without the 001 tomorrow-maybe that´s working?!...
shaidien Posted August 18, 2019 Posted August 18, 2019 That nif is set up wrong. Bone001 get rid of that one place the chains in right spot like foot maybe and turn the top off by changing 524294 under flags on top starting bone node with 524302 . That usually is for your perpus changing both boneR.01 flag into 524302 and let it be stationary bone the rest with 524294 as movable bones but if it were me I just add one in the top called boneR and boneL or even boneR00 and boneL00 then ths rest of your bones.
Recommended Posts
Archived
This topic is now archived and is closed to further replies.