Dead or Alive series formats and tools

[Lilstormcloud]

I haven't had cause.

[Rosalin]

prot.
Great, that's great. your app worked.
And that will definitely reduce time to make base body.

It seems that it can't handle with nails or toenails, but it's OK. I can handle it full enough.
I will provide a tip for that.

And I have a suggestion for your app.
You know there's only 2 numbers are required to pick out the body vertex buffer from TMCL.

The starting address of body buffer and the vertex count (in object properties).

How about making the app to provide text-box for inputting those 2 numbers for source and target each?
Then the app will do the process with original TMCL files, not with the user-made vertex buffer files.
And that will make the app much more user-friendly.

It will going to have no problem with reading or writing entire size of TMCL as an array or whatever.
or you can make additional '44 * vertex_count length' array to make calculation on your app easier.

***

About the neck seam..
It seems that matching 1st neck ring's material values is not enough for making perfect neck connection.
I could see some weird wrinkle in the 2nd(maybe) neck ring.

So I could guess the weight values of 2nd neck ring should be copied from original for more perfect connection.
or 3rd also maybe.

[Protocol X27]

Cool, Ros glad it worked. I know I could allow for manual inputs, but I want to focus on getting away from that since my earlier endeavors having been manual heavy. I want to invest the time into bypassing that so future apps won't even require it.

I'll try to start on the neck vert thing next. I'm thinking it'll require exact vert coordinates, so it has something to match on. The neck verts don't seem to match on UV or bone Id / weight, so they'll need something else.

LSC, I know dirt can be removed by switching the index to the all black dummy images, but I'm not sure how to actually break the object from thinking it has that multi layering altogether. :-/

[Rosalin]

I made a simple tutorial about Morphing & Partial Morphing.
http://youtu.be/9Ztw3sqybKI

**
http://youtu.be/5Faq2A8_quQ

**

prot,
yeh, I got your plan and I know you did it on the hielay matching app.
Actually, what I've told is not about what your app should be like,
it's just my personal request.
I'm already almost done with marie with your app, and I'm about to produce base bodies of all, like a factory.
If you are not about to make it automatic right now, I hope you can add that manual option temporarily. that will save my time a lot and you'll see what I've done on google drive also.

[Protocol X27]

Thanks for the tut, Ros.

I understand what you're asking. The downside is that I don't get a lot of time to work on apps, so I'm trying to maximize what little time I get. If I can I'll see what I can do, but I cannot make any promises just yet.

***

Had a few minutes today to test some things. The Ayane/Kasumi body style does not snap very well, it usually messes up the finger nails. Instead I finally tried using Alpha 152's body as an injection and with buffer adjust options it works just fine. Although filling with all 00's does negate the wet & dirt I believe. That will give me reason to attempt some util stuff for the vertex buffer like mentioned before.

[Lilstormcloud]

hair and faces.

[Rosalin]

lol.. I tried 'neck seam remover' to multiple length object group with some manual calculation,

I realized that I had wrong concepts about that.

Well.. it seems it's still ok with all 44 group mesh..
but the way to calculate it for multiple length object group I told before was totally wrong.

I think I should do some modifications on the app...

Code: Select all

length  Verts	ln*Vrts   >Actual Adress >My Calculation
44      1607    70708      6926352        6926352 
44      551     24244      6997060        6997060 (Perevious Adress + ln*Vrts)
44      916     40304      7021304        7021304
36      112     4032       7198896        7061608	???
40      160     6400       7202928        7202928	???
44      3120    137280     7061608        7209328	???

**
lil, I've seen your Excel macros.
It was so much more complicated than I imagined.
Why didn't you make it as a programming language?

I could sure you had no problem if you made it as C or Python or whatever.
In my case, after learning 'how to read and write files as bytes', it wasn't hard after all.
All it matters were how much I know about TMC & TMCL after that.

[Lilstormcloud]

I'm just more familiar with them.

[Rosalin]

I fixed my Neck Seam Remover app.
Previous version has some serious problem..
Please delete it. (Once it worked, files should be ok.)

Now it will work with multiple buffer length objects, like {44,38,40,44} whatever. (But body vertex buffer must be 44.)

NeckSeamRemover Fix_1.1.zip

Matching MtrCol was so stressful. because I've checked them in game.
But now I know how to find exact buffer data pretty easily.

****

Can I get some tuto about copying weight values?
I'm about to upgrade my app for weight value synchronizing.
It seems that some unnatural wrinkles appears behind the neck when weight values are not copied.

Let me set an example for that.

Code: Select all

LISA DLC 004  [7361]
BC 1F E8 68 3E D7 08 A9  3C 43 FC 44 77 3C 56 A0
00 54 9F 0C 00 1B 1A 06  F7 69 37 1D 25 43 38 08
25 43 38 08 25 43 38 08  34 EF 26 B0

MR DLC 004  [4807]
BC 1F EA 36 3E D7 08 A9  3C 43 FD 30 75 3D 16 74
00 0C A6 4D 00 23 22 21  F6 E9 6B 09 25 43 38 08
25 43 38 08 25 43 38 08  3A F9 36 0C

What I want to change is lisa's one.
Two vertices are in same geometrical position.

[Lilstormcloud]

it's the 5th set

[Protocol X27]

Sorry for the absence, been soooo swamped ... :-/

Will a direct copy of the bone ID work even across same character models though? Ros, when I get a chance I'll have to get you something to use for that.

With any of the neck information, did you find values that were an exact match?

Before I got detoured I had the util close for swapping any of the 11 values in the buffer. It'll be a quick finish, but I don't know when I'll get a chance just yet.

[Rosalin]

Well, I did tried.
Pic's on core.

It doesn't seem to be working with just copying weight values directly.
It got worse.
Do I need additional calculation for that?

I'll take time on that.
help me whenever you feel free.

and.. it seems that bone IDs are 4 bytes right?
Well.. I thought It was just 2.

[Lilstormcloud]

will look into it if not already resolved by weekend.

[Rosalin]

Now I got the reason and.. it's pretty simple and quite annoying.
it's about lisa. not momiji.

all lisa's body has been made like that originally.(tina's too)
it's just hardly noticable because of her hair.

just fixing faults was not that hard but,
refining it more naturally was so, so.. annoying.
it seems to be required to replicate 3rd or more neck ring's bones and weight values.

***

I fixed them all.

[Lilstormcloud]

Hair and ayane face up on google,