Dead or Alive series formats and tools

[Darko]

Not too sure, its similar to when I swapped kasumi dlc 4s ACSCLS block onto lisa dlc 4 (that's the one with the flying loincloth) soon as I swapped it long hair no longer goes through the body which was good. the clothe thingy however becomes static and no longer moves so I can only speculate that the info that determines how it moves was no longer there because I swapped it. Playing around with the values in that block also made other things static like bikini strings etc etc.

*****************************************************************************************************
>Dark
Here's the hitomi stuff you asked for, I put in the new swimsuit and hair as well. http://www.mediafire.com/download/p98cs ... istuff.zip

Thanks, really really thank you. If I have something you want to request me in a near future feel free to do it.

Edit:

Mario's fixed script seems to work perfectly with the files you provided me. DOA5 bony scripts don't work with those files at all, the texture extractor crashes when it's enabled in blender.

Code: Select all

x360 DoA5/DoA4/DoAX2/NG3 textures extractor

hi
i'm a texture extractor for xbox360 'Dead or alive 5/4/X2' and 'Ninja Gaiden 3'
supported filetypes: *.TMC;*.SPR;*.--P;*.--H;*.TNF;*.TTG;*.IBL;*.GMD;*.EFFMC;*.T
PR;*.XPR;*.C;*.CT;*.TP;*.TDP;*.TEX;*.CTB;
open this script in a text editor to see more settings that you can edit and sav
e

D:/ASSETS/DOA5/HITOMISTUFF/HAIRNEW.TMC files = 13
 13 DXT1  128 x  128 hit_bear_base.dds
Exception in Tkinter callback
Traceback (most recent call last):
  File "C:\Python32\lib\tkinter\__init__.py", line 1399, in __call__
    return self.func(*args)
  File "C:\Users\Darko\Desktop\tecmo360texturesExtractorV12.py", line 319, in <l
ambda>
    doitbtn = Button(text='extract now', width = 25, font='sans 13', command=lam
bda:convertnow(openentry.get(), saveentry.get(), samedir.get(), allfiles.get(),
subdirs.get()))
  File "C:\Users\Darko\Desktop\tecmo360texturesExtractorV12.py", line 41, in con
vertnow
    parseFile(infile, outdir, subdirs)
  File "C:\Users\Darko\Desktop\tecmo360texturesExtractorV12.py", line 153, in pa
rseFile
    Untile360(sf, l, inBuffOffset, TexelPitch, tiledWidth, tiledHeight)
  File "C:\Users\Darko\Desktop\tecmo360texturesExtractorV12.py", line 185, in Un
tile360
    untilechunk(outBuff, inBuff, sourceOff, rectWidth << v51, TexelPitch)
  File "C:\Users\Darko\Desktop\tecmo360texturesExtractorV12.py", line 209, in un
tilechunk
    data = bytes([data[(i-1) if i%2 else (i+1)] for i in range(size)])#swap ever
y 2 bytes
  File "C:\Users\Darko\Desktop\tecmo360texturesExtractorV12.py", line 209, in <l
istcomp>
    data = bytes([data[(i-1) if i%2 else (i+1)] for i in range(size)])#swap ever
y 2 bytes
IndexError: index out of range

[Protocol X27]

they got better compression so change in size largely due to that.

other than that the changes are very very small. sometimes only 10 bytes. it could be anything but 1 thing I notice is when you swap in costumes from doa5 the accessory physics no longer work like hair collision skirts become static etc etc sorta like when you ported ng ayane costume to doa5. It can be fixed by just adding the extra 10 bytes or whatever from the new tmc to existing doa5 tmc so I'd say that gives a good indication of what the changes are for. (new hair styles probably )

if you're talking about the omg boob setting that works even with regular doa5 tmc, i don't think ultimate added anything in the tmc to accommodate that.

Do you know which data is different between the new and old files now since you've seen the physics corrected w/ some changes? I find it odd that I've replaced one costume with the same one although modified and uncompressed of course and it still seems to mess up the physics. Hitomi's Tank top for example, suddently her hair has issues again. Do we know if that hair stuff is controlled stricly by the ACLS in the .tmc?

[Lilstormcloud]

can try a few things.

[Protocol X27]

you can try a few things.

Just swap the tmcl, although the ultimate tmcl might actually be a bit larger even for returning doa5 costumes so there's a high chance it'd just freeze.

Other sure fire way to fix it is just use a converted ultimate tmc. once you get a hold of one open it and look at the addresses around 0x140

it'll look something like this

00 00 01 90 00 00 41 30 00 00 52 E0 00 00 53 90 00 00 54 40 00 00 67 50 00 00 7F E0 00 00 DB 60 00 00 DC 00 00 01 43 60 00 01 88 60 00 01 CD 60 00 01 F7 70 00 03 18 C0 00 03 19 00

it says MdlGeo is at 190, TTX is at 41 30, so on, so on. compare it with the old doa5 tmc if you notice a change in address at any of these then you know the previous entry was larger which is where the difference is. (because if 1 entry is larger all the subsequent entries gets shifted down by that much).

so anyway if you paste in the difference and the new ACSCLS then all should work fine. (don't go changing anything other than cpf and ACSCLS). if you paste in extra stuff you need to then change the size and addresses at 0x140.

Unfortunately, I won't have access to the uncompressed files for a bit. XD

I understand the potential misalignment... so if a tmc contains all of the offsets to its own data and things still aren't lining up correctly, doesn't that mean a file external to the .tmc is still expecting specific data at a specific address? Considering this happened with Demo Ayane as well as NG3:RE Ayane. There was no ultimate .tmc so to speak to reference, yet the tmcs were still all inclusive but the physics was still borked. Offsets for ACLS would then need to remain unchanged (possibly?) to prevent the physics issue.

[Lilstormcloud]

I guess you can extend

[b0ny]

tried my best but could not find how the material colors from "*.cat" file, are linked to materials in "*.xpr" file, in dead or alive 3. i'm stuck with this right now, any halp?

thanks to mariokart64n i know how this works for dead or alive 2 ultimate, but i need also doa3 coz i'm working on a better converter for costumes from doa3 to doa online, everything pretty much is converted, and i spent a lot of time trying to figure out how the cat material colors are linked to the xpr materials, and i feel very dumb coz i could not find any thing.
please help me...

[mariokart64n]

in doa3? i only looked at doa2u

[b0ny]

in doa3? i only looked at doa2u

can you also give a try to find how materials are linked in doa3?
i converted everything but have no idea how cat colors are related to xpr materials. if i could link the colors to the right xpr materials that would be a complete conversion...

[edit]
i think that the fact that they have material colors in the xpr and also in the cat file for some materials, is caused by the tools they used to create these models - like for example, they created the model, then selected some parts and increased specularity, and these changes made for the entire group, resulted to a material in the cat file, that could be easily modifyed in the future for the entire group.

i think cat material colors are important. i loaded one of the leon's skins on a sunny stage, and, that could be just my imagination, but i think it's not specular enough. i could be wrong, but my first thought was, that the colors from the cat file could fix that.

i thik that finding the links can't be done just by looking at the hex. an xbox1 should be used to make tests on doa3, on obvious materials like on zac02(alien). by disabling/enabling some materials from the cat file and find how to disable them in xpr. i think that would be a much easier way of finding "the switch" that links the cat colors to xpr.

[Darko]

Bony, are gonna update your texture extractor for DOA5U?

[b0ny]

tecmo360texturesExtractorV15 - (click to download)

blender importers for team ninja games v25.zip - (click to download)

Darko - fixed
i think the problem was that your models are dumped from the memory, where the addresses for textures have been initialized

[Protocol X27]

tecmo360texturesExtractorV15 - (click to download)

blender importers for team ninja games v25.zip - (click to download)

Darko - fixed
i think the problem was that your models are dumped from the memory, where the addresses for textures have been initialized

Most Excellent!

I was having the same issue to since I was using dumped .tmcs. You bring up an interesting point. Could the fact that the addresses are different in dumped containers be a possibility for why using those dumped / uncompressed .tmcs still won't work in game?

Is there a way to get the texture addresses before initializing and fix the .tmc?

[Darko]

Thanks Bony .

[mariokart64n]

i'm soo lost on what happened, doa2u is completely reversed? did you make a blender plugin that can make new models for the game?

[b0ny]

Could the fact that the addresses are different in dumped containers be a possibility for why using those dumped / uncompressed .tmcs still won't work in game?

Is there a way to get the texture addresses before initializing and fix the .tmc?

if you'll give me a doa5 dlc skin dumped from doa5u's memory - i'll have a look for differences, and maybe it's possible to create a script that could undo the initialized parts of the model.

i'm soo lost on what happened, doa2u is completely reversed? did you make a blender plugin that can make new models for the game?

no, i reversed it enough for a doa3 to doa2u conversion. the joints, the physics, and some chunks from the cat info block are not reversed.

i put the most effort on reversing joints - trying to understand how to convert the doa joints to weights and backwards. i reversed some operations - but i'm stuck on a portion of code where the data is gathered and sent to an anonymous function(method?) of "d3d9.dll", which i can't identify

[Protocol X27]

if you'll give me a doa5 dlc skin dumped from doa5u's memory - i'll have a look for differences, and maybe it's possible to create a script that could undo the initialized parts of the model.

Here's Ayane's original Swimsuit DLC as pulled from memory.

http://www.mediafire.com/?b805niq5rqmhov0

It seems like penguin has figured out how to make the new .tmc's work. He posted a pic of some changes in the beginning of a file, but I haven't done a cross check yet myself.