[XBox360] Dynasty Warriors 8 (Early Access)

Release and comment on game extractors and other tools here.

Moderators: Dazzy, semory

tianmuxia
Posts: 2
Joined: Thu Mar 27, 2014 6:37 am

Re: [XBox360] Dynasty Warriors 8 (Early Access)

Post by tianmuxia »

TRDaz wrote:We cant get physic parts of the model either unfortunately. We have helpful obj's that help create them however.
What 'helpful objs' ?
Vertices parsed from NUNO part of G1 Model?
User avatar
semory
Site Admin
Posts: 7755
Joined: Sat Aug 04, 2012 7:38 pm
Custom Rank: Kitty pu tu tu lay!
Location: Torrance, CA

Re: [XBox360] Dynasty Warriors 8 (Early Access)

Post by semory »

yeah, somehow the real cloth is computed from the weights in the vertex buffer and the control points in the NUNO section. i still cannot figure them out. i've tried a shitload of different equations and formulas but nothing has ever worked.

someone told me on OPKM2 Monkey has a version of his main player model with the cloth intact, which can be used as reference. i just don't have time anymore to figure this out so if you want to... do give it a go, you'll make a lot of people happy as nobody's been able to figure out the cloth now for years.
SaintLouisX
Posts: 16
Joined: Wed Jun 04, 2014 10:04 pm
Custom Rank: Aaaa

Re: [XBox360] Dynasty Warriors 8 (Early Access)

Post by SaintLouisX »

I suppose I'll leave this here before I forget. I don't know if it'll help you figure out the cloth, but here's the format for reading the NUNO table: http://pastebin.com/Vakaxg9f
User avatar
semory
Site Admin
Posts: 7755
Joined: Sat Aug 04, 2012 7:38 pm
Custom Rank: Kitty pu tu tu lay!
Location: Torrance, CA

Re: [XBox360] Dynasty Warriors 8 (Early Access)

Post by semory »

thx, the output of my ripper includes nuno and nunv obj files that contain the geometry represented by the data for which your python code stores in your "table" variables. one table type stores connectivity information for cloth control points and another stores the actual control points. the other tables never really figured out what they are used for. the hard part is determining what mathematical formulas are used to generate the cloth from the nuno, nunv, and vertex buffer data.

hahaha i was even thinking about trying to find one of the developers and ask for a hint lol. typically they are bound by non-disclosure agreements though :( .
SaintLouisX
Posts: 16
Joined: Wed Jun 04, 2014 10:04 pm
Custom Rank: Aaaa

Re: [XBox360] Dynasty Warriors 8 (Early Access)

Post by SaintLouisX »

Ah, well if it helps, I just found out today that they use other files for rigging the models, the RIGB files. I'm not sure if your ripper can read/process those ones. 0x4-0x8 are bytes, everything else is a dword. All the character rigging files are in file 8096 for PS3 XL, but I'm not sure about Xbox sorry.

The 0x70 vertexes control cloth though right? I know chrrox said they did, and I before I reversed those properly I would get errors in the cloth. Fortunately it's easy for me to whip up an example by just breaking my script again:

http://i.imgur.com/cRATzso.jpg - A rip in his cloak, maybe a bit hard to see but it's easy to see in-game.

http://i.imgur.com/rMG70Wn.jpg - The 2 front thingies that hang down are just mashed together in his middle.

This is from only breaking a couple values. Reversing 12 bytes at 0x40 - 0x4c of the vertex as dwords instead of keeping them as bytes. So the cloth isn't entirely corrupted, most of the vertex is fine, but yeah, those 12 bytes make that happen if reversed that way.

I'm really bad at modelling and such so I can't really debug *how* it's computed, because I have no idea what I'm looking for really, but the 0x70 vertexes are what you want I guess.

EDIT: Actually maybe not, at least not entirely. Just to see what would happen I didn't reverse the 0x70 vertex at all, and the output is exactly the same as the screenshots up there.
User avatar
semory
Site Admin
Posts: 7755
Joined: Sat Aug 04, 2012 7:38 pm
Custom Rank: Kitty pu tu tu lay!
Location: Torrance, CA

Re: [XBox360] Dynasty Warriors 8 (Early Access)

Post by semory »

actually, it is not the vertex size that determines if something is cloth or not. chris' (chrrox's) script is incomplete. he told me he quit because of the cloth. there is a section that determines vertex layout: if you look at my debug output search for word "semantic" in the file. any vertex buffer that references an input layout that has an 0x700 (if i remember correctly) semantic, is for cloth. this is why his script only works with dw7 (because he doesn't process the input layout section).

i can also post my code if you want to see it, and my code works with a bunch of dw games from 2010 to 2014 (it is amazing how tecmo koei milks this engine, it really hasn't changed at all in five years).

btw, is that pic from PC dw8? the models look horrible if it is lol.

i will try per your suggestion messing with the bytes from 0x40 to 0x4c, as i have only tested the ones from 0x00 to 0x0c. i will let ya know the results soon (i have some test data from nene's vtx buffer and nuno that i use for quick testing). and thx for the info on those rig files, i will be sure to check them out!
SaintLouisX
Posts: 16
Joined: Wed Jun 04, 2014 10:04 pm
Custom Rank: Aaaa

Re: [XBox360] Dynasty Warriors 8 (Early Access)

Post by SaintLouisX »

Yeah it's PC DW8, but I don't have the graphics on max, actually have them on medium or low. Actually seems that DW8 is a lot better optimised than DW7 is, especially in regards to numbers of enemies on screen, so I could crank things up, but yeah it's not representative of the best possible.

Here's all the vertexes if you want: http://pastebin.com/N2yY2w50

0x70, 0x20 and 0x30 were the ones that had bytes in. 0x70 is actually 16 bytes not 12.

How do I get your debug output? I'm not sure if the archive in the OP of this thread is the same as the xentax.exe in your other tool. The one you posted in this thread and the WO3 thread are corrupt for me, can't open them in WinRAR or 7Zip. Didn't see any debug output for xentax though, is it a commandline option?
User avatar
semory
Site Admin
Posts: 7755
Joined: Sat Aug 04, 2012 7:38 pm
Custom Rank: Kitty pu tu tu lay!
Location: Torrance, CA

Re: [XBox360] Dynasty Warriors 8 (Early Access)

Post by semory »

im not on pc atm, but the 7z file in the OP of the gas machine thread has latest ripping program. i can also post a sample file and output later today. i will test your code today on nene from sengoku musou 4. i will post her files and data here after i wake up today (going to sleep now). if this cloth is ever figured out lol, a few peeps around here will love ya as it is a pain to replicate using the UVs as guides in blender.
SaintLouisX
Posts: 16
Joined: Wed Jun 04, 2014 10:04 pm
Custom Rank: Aaaa

Re: [XBox360] Dynasty Warriors 8 (Early Access)

Post by SaintLouisX »

If you can give me the whole file (with the textures) I can probably import it into DW8, which would be awesum.

But like I said, I really know 0 about modelling and things related to it at all, so I really can't help you figure out what parts of the file actually do, all I did was get the data format. I remember chrrox posted on Xentax asking "how" to read the 0x70 vertexes, so thought it may help if you can at least read the file correctly, and then guess what the sections do going from there.
PoophairJones
Posts: 1068
Joined: Wed Oct 17, 2012 5:09 am
Custom Rank: Baebsae
Location: Idk where am i?
Contact:

Re: [XBox360] Dynasty Warriors 8 (Early Access)

Post by PoophairJones »

Just a few thoughts about these NUNO meshes... is it possible that they are just some meshes that tell the physics where to act? Somehow like a starting point or so... cause in most cases (hair mainly) it's not possible to recreate the actual meshes from those because the shape is entirely different and way more simple^^
Post Reply