What?

It looks like you were trying to point me to something, but it didn't make it into your post.

What unknown int?

Ah sorry, Unknown Short 1, in the nitrishapedate in a f3 nif exports as an 8. in skyrim nifs its always a zero. I have no idea if that affects anything if you miss it, but I mentioned it in the mini tut.

1. Found a few things while testing the settings in the shader property:

• In the unknown floats list (the one you need to expand) the first unknown float is the same as the 'alpha' setting in the material property of Oblivion nifs. If an alpha property is applied setting this to lower than 1 will add transparency to the mesh, so 0 is invisible while 1 is completely visible.
  
  
• Like already mentioned - in the same list the third unknown float is the same as the 'glossiness' setting in the material property of Oblivion nifs. Low values make the specular effect look 'dry', while high values make it look 'wet'. Unlike in Oblivion nifs the effective range seems to be 0-1000 and not 0-100.
  
  
• The seventh float in the list made my game lock up when I changed it to something else but 2 (only tried once setting it to 0.5). Not very helpful I guess.
  
  
• The unknown float set (which is grayed out by default) directly below the expandable list apparently controls the strength of the cubemap effect if a cubemap is added. I only found it in nifs with a cubemap and with a value of 15 the cubemap effect was very strong, while with a value of 1.5 it was subtle.

2. Again a new collision type...hrmph. Hope it won't take as long as the Mopp shape for Oblivion to get this working.

3. I generally have problems opening nifs with the old mopp shape in place since I installed the new xml. If I try it says 'failed to load block x, previous block was bhkCollisionObject' and 'failed to load nif'. And everything but the collision branch is gone. Doesn't happen with other collision types. But maybe it's just a messed up Nifskope install, I will try to reinstall. I have the latest version (1.0.22).

Oh, and thanks for your effort of course!

In the original Steel Plate F pair, that value is set to 8.

I duplicated as much as I could between the original and my modified version, and it still won't render.

@madcat and his not appearing sword

you need to replace the 1stperson model for the sword to appear in the players hands if the "normal" version of the sword has a WNAM entry, if you want it to appear in your inventory too you definately need that WNAM 1st person entry
WNAM is the 1stperson model of a given weapon and links to a static, hence create a STAT 1stperson item and MODL it to your new model, then change the non1stperson weapon's WNAM to direct to your 1stperson STAT formid... lol hope that didnt sound too confusing XD

So if I just delete the WNAM it'll work? Or will it asplode?

I just thought about what you are trying to do, and it might be due to the morph? have you replaced both nifs ?

Yes, I am replacing both nifs.

I've even tried replacing both morph halves with the same mesh and it's still not.

It seems I'm not the only one having problems with this exporter from the sound of things on the NifTools forum...

My objective is to do some rigging fixes to the female Steel Plate armor (meshes\armor\nordplate). I'm just doing a test run by just stripping out the unnecessary neck weighting in the collar piece. Some of the vertices at the very front tip of the collar have 100% weight to the neck bone.

@madcat and the sword
yup nothing will explode if you remove the WNAM but you wont have an inventory model anymore, rather make a new STAT entry and direct that WNAM to it

didnt try morphs and rigged stuff yet so cant really comment on that

This is now officially peeing me off.

You graphics modding thread on the nexus I am going to post in... until some sort of master thread happens.

And this thread on this forum. Seeing as Jaysus will probs update and link in th first post. SO it's mostly contained.

Seems I had an old version of your nif.xml, now opening nifs with a Mopp TriStripsShape works. But I still can't copy the old Mopp over to a Skyrim nif, it always says 'unkown block'. When I try to copy each part of the collision branch one by one it works, but the subshapes in the bhkPackedNiTriStripsShape get deleted and the field is grayed out so I can't add them back in. Nifskope also becomes incredibly slow when I copy/paste hkPackedNiTriStripsData into a Skyrim nif.

maybe import the mopps as obj and once you got em in a skyrim nif copy/paste over their shapedatas... not sure... didnt bother with collission so far, im too much in love how the game renders stuff now, the inventory item view is also great for screenshots in nearly render quality

Sounds like a plan.

Actually it looks like there is a problem somewhere, either in the nif.xml I'm using at the moment, or the build I'm using. Loading a .nif that has a CompressedMeshShapeData has that block floating off outside of the root node AND linked from the CompressedMeshShape. Looks like PackedNiTriStripsShape is set to end/die with Oblivion files (v20.0.0.5); Looking at the *Data block, changes to Fallout 3 needed that. It's likely that if this block exists for Skyrim it will be the same. You could try replacing that vercond with this one as a personal test: vercond="User Version != 12"

In my last play session, I was mucking about with console commands before I quit and found out that the game does geometry culling, this reduces the overall polycount nicely, especially in outdoor scenes.
I haven't started on that LightingShader node, but came up with a good testing plan yesterday. I'm too worn out from festivities to give that a go tonight however, hope everyone had a good thanksgiving!

I suppose this is the right place though OT from the current discussion. I've created my own fork of TESsnip to allow editing of ESP files and ESM files without FOMM. What I noticed is that the FULL substructure changes meaning depending on whether its an ESM or ESP file. When its an ESM, the strings are an integer used to lookup into the files in the Strings folder where there are English (at least for me) versions of the strings. When the file is an ESP, the FULL structure is just a normal string embedded in the file. TESsnip was never designed for this kind of behavior and was rather cumbersome trying to make work but I got enough of it going to make my own armor placeholder with custom description.

I really like the idea of the strings being externalize but would be nice for ESP as well without having to convert everything to ESM files.

Here is that build if anyone is interested in it. http://www.mediafire.com/?c1h4hb6lqx73j09 (git link is available in the readme). Anyway i think it might be useful with the existing tutorial though i was only using it for ARMO records not WEAP as per the tutorial.

there also seems to be something wrong with the formids, had a piece of the riverwood inn dissappear because one of the swords in my mod seemingly had the wrong formid, i can only speculate that it overwrote the other one tho tessnip didnt tell me its a duplicate and with the mod index numbers in front that shouldnt happen anyway... guess a CS would truly be nice afterall lol

il post that skyrimsnipper on the mainthread holo

So is there any way to get armor into blender yet?

I can not seem to find a answer or explanation on that, just lots of people arguing over blender being good or not lol. o_O

If by "the viewer" you mean Skyrim Viewer, that's because the current release was going off of old information about the FULL and other related fields. See my response below:


Actually, that's not entirely true. We've decoded this a bit more:
bit 7 (hex 0x80) in the TES4 record header is the 'isLocalized' flag. If this flag is True (which it is in Skyrim.esm and Update.esm), then this tells the engine that any field of type 'lstring' (aka, localized string) should be a 4 byte number referencing the string entry in the .DLSTRING, .ILSTRING, and .STRINGS files for that mod. If the flag is False, it means that any field of type 'lstring' should be a zero-terminated string. Also, those string files can be packed into BSA's as well, so it's a [censored] to look them up if you're doing it with an external program.

The 'lstring' type fields I've found so far are FULL, DESC, RNAM, CNAM, BPTN, DNAM, ALFD, ANAM, TNAM

I've not made any progress on the persistence issue. Kind of getting annoying. Decided to work on other projects . The update on the localized string is good but I swear I tried that and it didn't try loading my strings. But I was trying a lot of things and could have missed it. Thanks for the heads up on the other ones.

I have to admit I don't really know what I'm doing with the nif.xml, I couldn't find a vercond in the section for PackedNiTriStripsShape so I simply replaced every vercond in the xml with the line you suggested. That may be completely wrong, but it didn't cause more bugs, neither did it help.

I've got one more question not directly related to the new nif structure. I replaced the mesh of the 'ambient eagle' (the birds you see circling in the sky sometimes) with an animated cliffracer from my Sheogorad mod. That works fine (the cliffracer shows up with textures and moves along the same path as the eagle would), however the wing flap/head/tail movement animations of my Cliffracer don't work and I can't find the error. The cliffracer has a skin instance and is rigged to a few bones. The bones do animate in game (tested that by applying a mesh to a bone directly), but the cliffracer does not. So it seems to be a problem with the rigging/skin instance. In Nifskope everything looks fine. If anyone would be kind enough to look at it, http://www.mediafire.com/file/ia7yjlu228s67er/Cliffracer.nif is the nif.

EDIT: Oh, and of course the cliffracer with all its animations did work in Oblivion, so it's not messed up by default already.

arrrrrr... could the game be broken in regards to mod support?
just realized that all the items i added replaced references in the game... like whole buildings and stuff lol, seems like it doesnt add a mod index to the formids at all -.-
any suggestions which formids are save to use?

edit: oh lol, forgot XD
form ids starting with 01 should be random... worked in fallout at least... hence just add an 01 at the start where there is probably 00....

edit2:
mmh... well its ok now but the objects cant link to each other anymore since i added the 01 -.-

edit3:
all is good lol, just dont name your formids like 01000001 ...rather do 010a0001 (the first 01 will be replaced with the mod index... if you start with 00 however it will be static and most likely conflict with somin)

Any chance you would share the source, the version you have posted seem to have the "copy / paste" issue I run into when not running TESSnip through FOMM, or just the changes you did to TESSnip.

Yours truly
Zilvereyes

I was using 0A000001 and was not having success so switched to 00A00001. The later does not persist in the save file. The former does not get instantiated using player.additem. Is there something else that has to be changed for the upper byte to be used? Or should I just be using 01000001 and the game will sort out the conflicts for me somehow?


Thought I posted the link in the readme but maybe not. You can download a zip or get it from https://github.com/figment/tesvsnip. Original source code is available on sourceforge somewhere.

-snip- Guess it's all good

Ok now I understand how that upper byte works a little better. Just set that upper byte to 1 and its changed automatically (in my case it changed to 02xxxxxx) and the added benefit was that the item was actually persisted across saves where it was not when the id was 00xxxxxx. This ensures some level of scope to avoid conflicts. I assume that they will update mods that refer to other mods so that IDs are remapped for scripts.

@formids

-ALWAYS start em with 01 (01 tells the game to replace the 01 with your mod index hence preventing any possible conflicts, 00 makes it fixed and unchangable basicly always overwriting vanilla refs or somin, 02 etc dont do nothing but also can overwrite vanilla stuff

-01000001 does not work, try 010a001 instead and go from there (010a002, 010a003,etc)