Dead or Alive series formats and tools

[Rosalin]

it's looking that you've done rightly.

and yes. that's normal.
It's not going to be exactly same usually. that's the way float works.

[Protocol X27]

it's looking that you've done rightly.

and yes. that's normal.
It's not going to be exactly same usually. that's the way float works.

Awesome, thanks!! Okay, so it sounds like it's just a rounding issue then. That's a huge help to one of my last major projects.

I finally had a chance to look at some of your utilities last night. I had no idea how far along you were with those. The UI is really nice. Regarding one tool, the Vertex Matcher, I think... So does that affect materials & normals too? Besides the neck issue, I still have issues when using different character bodies. Even after updating the MtrCtl, I think there were still issues, so it looks that tool corrects that right? If you had mentioned that in the past I forgot since I had gone a long time without making progress.

There's been so much progress with the PC version compared to the Xbox version it's awesome. Even the stuff with .obj injection editing by others opened up a lot of potential for custom inserts. Other than the means of copying, adding, or re-parenting bones from other costumes, the sky is essentially the limit.

***
I just got a successful vertex converter written. Other than those rounding issues, the output looks good. I now see what you were saying about some information being stored weird. The tangent & UV coordinates aren't just endian flipped. A endian flip for UV would normally be [A,B,C,D] to [D,C,B,A] but for some reason to match PC it needs to be [B,A,D,C]... oh well, easy to fix.

This got me to thinking though. Do the Xbox Indices require a special expansion conversion as well?

Edit: nvm, I think I solved it.

[abcdefh]

You can use the alpha channel of texture as a specular map, it seems to manage to use as a transparent.

If you are able to explain which values should be changed to, I could see what I might able to fix.

Since I don't know how these sections work, could you show an example of the difference between one using the alpha as specular vs one using the alpha as transparency?

Example：AYANE_DLCU_001
Two of mesh in first ObjGeo is copied.(WGT_body、WGT_body_1)

・itable
1.The number of subobject is changed to 4 from 2.(Green part)
2.Copy source divides the index number of customp which was referring.(Blue part and Purple part)
3.boundary adjustment is performed.(Orange part)
4.Shift the offset of WGT_sweat_chest.(Light-blue part)

・matecp
1.Shift the offset of each customp.
2.The block size of matecp is increased.

・cpf
The block size of cpf is increased.

[SleepySea]

Wow, Ros is back. Ros has helped me a great deal in writing my own converter (and basically inspired me to get into this type of game modding). I hope you have had a chance to try the tool I sent you. It probably doesn't work as well as you'd like, but it's my first novice attempt at DOA5 modding, lol.

@ProtocolX27: Your tools and patience in helping me have been such indispensable resource as well. In fact my very first attempt at converting a model from 360 to PC was partially based on one of your earlier tools.

I managed to add some more stuff to my own converter such as allowing people to edit any of the four UVs, fix corrupted normal vectors/seams, and some other fixes/enhancements, but I kind of lost steam as of late as well. Good to see that tools are getting more and more advanced by the day.

[Rosalin]

@Spleey
I didn't test it because it looked it's working good enough.
it looked it's working good more than I imagined. maybe I can ask for the code about it for my tools.
and second, I didn't actually needed it with normal problems. I was working on nothing about it, I just wanted to inform about that issue.

And there was another issue with the vertex normal also.
It's not a general problem. it's just something about my mods only.

In my mods, the vertex normals on additional meshes doesn't always follow the shape of the mesh for better expressions.
So if you 'recalculate' the vertex normal depending on mesh shapes, those will be lost anyway.

for example, I know there are some people who prefers earlier version of my mods. (you know about the caps on the breast)
It has very complicated mesh shapes on the tip of the center, so if you initialize the vertex normal of the mesh, you will see ugly cracks or overshadings on there.
It will never be easy to refine the meshes. I used some tricks to get the clean vertex normal of that cap meshes.
so I made 'refined' version for that for easier handling.

It's kind of tricky parts to explain so I didn't informed about it actually..
You'll notice changes on the shades, with the meshes I added, after you recalculate the vertex normals.


@prot
Good to see that you made it m8

[Protocol X27]

@abcdefh Thanks for posting the explanation. After I tackle some more stuff, I'll look into updating the Reactor some more. Next thing I have to do is get the Xbox to PC automater working, so I unforunately have to suggest manual updates until then.

@Sleepy, glad that we've been able to help and that you've been learning! Sorry, that I've been too busy to give you valid feedback as well. I tend to be linear focused and work on one project at a time, but I really need to dive into yours since, I've noticed some normals on my own stuff that could use some adjustments.

@Ros, Thanks! Your info was huge! I have only written methods so far. Next I need to implement them and do some actual tests.

***

Made some light but important changes. No one wants to spend time doing the monotonous stuff so I keep trying to find ways to cut out the repeat stuff.

Toolkit Update: V2.3

MeshReactorDual replaces PC Version: Now works with Xbox files as well, so that theoretically a modified PC file can be 'Mesh Warped' to a modified Xbox file when combined with ...

MeshWarpAuto: Added the ability to process multiple ObjGeo mesh items at once. This saves a ton of time instead of processing, change value, copy/paste output, process, change value, copy/paste, etc. Now provide an array of mesh pieces like [1,2,3] instead of 1.

https://www.mediafire.com/?kx7wkcf8ed6qdkr

[SleepySea]

@Ros -

Yes I noticed the strange shading after I attempted to get the "correct" normal vectors for your models and realized that you managed to hide certain imperfections with normal vectors, but it really is already very good and the imperfections are very hard to tell even without hiding them.

This is such a neat a trick that I ended up employing in some of my stuff later on. It's a convenient way to make a complex/high density parts look smooth and hide "dark spots" on models. I've had to use it to smooth out some high density nipples by "borrowing" normal vectors.

@Prot -

Sorry I didn't read your previous post carefully, so I am not sure if this stuff is still relevant - I had some experience dealing with UV while writing my converter. The UV representation in TMC file for each UV coordinate is a short float (16 bit half precision floating point format) in four groups, so to convert Endianness you would need to just flip 2 bytes at a time. This unfamiliar data type problem gave me a bit of an issue as if I type-cast it to a short and then to a float, it would be a very large number as opposed to a decimal. If you don't need the actual UV coordinates, this probably isn't an issue.

As Ros mentioned earlier, it's totally normal to lose some precision when casting/converting floating points, so typically in 3D applications people do a "close enough" match as opposed to exact match when determining equality of two floating points. In 3DS MAX, when outputting to OBJ format, it actually drops a lot of precision and use only 4 decimal places (can be adjusted to a higher number). It appears to be good enough for this particular application.

Thanks again for the updated tool!

[th0rough]

I have the new dlc files, but archivarius isn't decrypting them properly. What gives?

[sf234]

Updated DLC Tool.
viewtopic.php?p=111891#p111891

• Added optional compression of files (thanks to immi from FSD for bit flags information). Use it with caution.
• "---C"/"--P" files now can be skipped (if they are not added built-in stubs will be used instead).
• Hairstyles now limited to 8("H") per character/slot (9+ crashes the game).
• dlc5lr.dat updated to match 1.04 version of the game (some entries are changed, so former version (and DLCs built with it) is not fully compatible with the current game version).

[bbbmods]

Updated DLC Tool.
http://rghost.net/private/6cXpqKQrT/cd2 ... b7b527e49a

• Added optional compression of files (thanks to immi from FSD for bit flags information). Use it with caution.
• "---C"/"--P" files now can be skipped (if they are not added built-in stubs will be used instead).
• Hairstyles now limited to 8("H") per character/slot (9+ crashes the game).
• dlc5lr.dat updated to match 1.04 version of the game (some entries are changed, so former version (and DLCs built with it) is not fully compatible with the current game version).

Thanks for the update . I noticed the swimsuit dlc still can't be opened. any chance for a fix for that? I would like to add the missing hair styles to it.
also, for point 4, not sure I understand. does that mean dlc packs made with the older dlc tool wont work with 1.04? I have not updated to 1.04 yet because I do not wish to lose lnk_reshuffle.

[sf234]

Updated doa5LRarchivarius.dat (should support all DLCs except 358142g).
viewtopic.php?p=110757#p110757

Free/used costume slots information.
viewtopic.php?p=111057#p111057

th0rough
Try to replace doa5LRarchivarius.dat with the version linked in this post.

bbbmods
DLC 358142 uses some strange character/slot definition for first 12 entries. You can use attached version to make changes that you want, then copy hex (from offset 0x8 to the end) into original file (at offset 0x20).
As for 1.04, they changed some entries in names table, e.g. MARIE_DLCU_018.TMC was listed in previous version as

Code: Select all

2Y512@1S1T1@CH10]]J183H42DO1G187QR7H1O{S76171J

and now become

Code: Select all

2S512@1S161L1T1@DI1S11]]L1B3J42GQ1I187RS7A1K1R1R{J981Y1A1N]1

There is bunch of similar changes, but everything else remains same. In the end, some combinations of character/slot will work with both versions, and others will work only with one of them. If you want to stay on 1.02 then just use old dlc5lr.dat.

[bbbmods]

Updated doa5LRarchivarius.dat (should support all DLCs except 358142g).
-deleted-

Free/used costume slots information.
http://rghost.net/private/8GFsKcYJh/688 ... 7ea5c28d1b

th0rough
Try to replace doa5LRarchivarius.dat with the version linked in this post.

bbbmods
DLC 358142 uses some strange character/slot definition for first 12 entries. You can use attached version to make changes that you want, then copy hex (from offset 0x8 to the end) into original file (at offset 0x20).
As for 1.04, they changed some entries in names table, e.g. MARIE_DLCU_018.TMC was listed in previous version as

Code: Select all

2Y512@1S1T1@CH10]]J183H42DO1G187QR7H1O{S76171J

and now become

Code: Select all

2S512@1S161L1T1@DI1S11]]L1B3J42GQ1I187RS7A1K1R1R{J981Y1A1N]1

There is bunch of similar changes, but everything else remains same. In the end, some combinations of character/slot will work with both versions, and others will work only with one of them. If you want to stay on 1.02 then just use old dlc5lr.dat.

Thanks it worked So I guess its the hot getaway costumes that are the problem since they are the only ones I couldn't edit, but fortunately those ones already had all their main styles, so only a couple special ones are missing for those like senran ayane, headband hitomi, rad band pigtail leifang. but no biggie

So if I understand correctly, certain costumes that had their names changed, that we used in custom dlc before wont work in 1.04? but the official packs that had them would still work?

also, one more question. would it be possible to open up official dlc packs with dlc tool and delete certain costumes? I ask this because I made a pack of all nyotengu costumes to have no wings, and have glasses instead of her mask. i replaced her original with lnk_reshuffle, but for dlc, I had to make a new pack for them, so I have 2 versions of those costumes in my game, and have ran out of nyotengu slots to include wingless versions of the new dlc. so I wish to remove the original wing version which I don't use, to have free slots and no unnecessary duplicates.

[sf234]

Updated link for doa5LRarchivarius.dat:
http://rghost.net/private/6CLSCrcBM/86a ... 72c0300c0a

So if I understand correctly, certain costumes that had their names changed, that we used in custom dlc before wont work in 1.04? but the official packs that had them would still work?

As far as I can see name entries was changed for costumes from new DLCs (that was released together with 1.04 patch, also for Kagura Costume Set (MARIE_DLCU_018's new name used here)). Names that used in old DLCs, and that not used at all, remained same.
I should mention that changes for Kagura Costume Set, that most likely was made in 1.02A patch, are not reflected in previous dlc5lr.dat version (it was dumped from version 1.02). So those names/slots are not available if you will use previous version.

also, one more question. would it be possible to open up official dlc packs with dlc tool and delete certain costumes? I ask this because I made a pack of all nyotengu costumes to have no wings, and have glasses instead of her mask. i replaced her original with lnk_reshuffle, but for dlc, I had to make a new pack for them, so I have 2 versions of those costumes in my game, and have ran out of nyotengu slots to include wingless versions of the new dlc. so I wish to remove the original wing version which I don't use, to have free slots and no unnecessary duplicates.

You can use archivarius' "swap selected" function and just swap needed models in original DLCs, just don't forget to backup them. As for deleting characters from existing DLC (also adding new characters, changing costume slot, etc) - maybe it will be implemented in future releases. The thing is that there is no easy way to do it. And I'm just lazy.

[bbbmods]

Updated link for doa5LRarchivarius.dat:
http://rghost.net/private/6CLSCrcBM/86a ... 72c0300c0a

So if I understand correctly, certain costumes that had their names changed, that we used in custom dlc before wont work in 1.04? but the official packs that had them would still work?

As far as I can see name entries was changed for costumes from new DLCs (that was released together with 1.04 patch, also for Kagura Costume Set (MARIE_DLCU_018's new name used here)). Names that used in old DLCs, and that not used at all, remained same.
I should mention that changes for Kagura Costume Set, that most likely was made in 1.02A patch, are not reflected in previous dlc5lr.dat version (it was dumped from version 1.02). So those names/slots are not available if you will use previous version.

also, one more question. would it be possible to open up official dlc packs with dlc tool and delete certain costumes? I ask this because I made a pack of all nyotengu costumes to have no wings, and have glasses instead of her mask. i replaced her original with lnk_reshuffle, but for dlc, I had to make a new pack for them, so I have 2 versions of those costumes in my game, and have ran out of nyotengu slots to include wingless versions of the new dlc. so I wish to remove the original wing version which I don't use, to have free slots and no unnecessary duplicates.

You can use archivarius' "swap selected" function and just swap needed models in original DLCs, just don't forget to backup them. As for deleting characters from existing DLC (also adding new characters, changing costume slot, etc) - maybe it will be implemented in future releases. The thing is that there is no easy way to do it. And I'm just lazy.

when I click the swap selected button in the archavarious, I always get this error

[sf234]

bbbmods
Which file in which dlc? What version of archivarius you use? Latest is either this or this.
The logic behind BLP entries detection (which is used in swap function) usually fail on small files (---C/--P/--H), but it should work for TMC/TMCL. Anyway, I didn't spent too much time fixing/testing swap function, and looks like b0ny himself didn't touch it either.