New Clothing Body Style Converter Beta v0.89f (10-26-2014)

[ArgusSCCT]

I guess it can be done. But does it work with every armor? These armors are not like that mesh, they're fully covered, everywhere it counts. 

[AnnaFly]

I guess it can be done. But does it work with every armor? These armors are not like that mesh, they're fully covered, everywhere it counts. 

Greetings.

 

..." does it work with every armor ?..."  the only way to know that  is to try first !

 

I've tried this test for you , now it's time to test some meshes by yourself.

 

In the  second ".nif"  i've  posted , there was another upper body part ( in 2 pieces)  once converted.

 

I just wanted to know if i can replace those parts by the "upperbody" part i used to create the " Lattice ". 

 

So , in my opinion it 's possible to tweak the "settings"  ( or choose another Lattice ) to get a better result.

 

The purpose of your question was " ...is it possible to convert a male outfit ( and the model i used  was a "male" version) ... to a female model . 

 

So , yes , it's possible. How it 'll looks , that's another story.

 

Seriously , it's better to try by yourself.

 

Cheers.

[gerra6]

OK. I'm Back.

 

I was on a Alaska Summer Burn Out Sabatical. But yesterday was freaking cold, so it's probably time to start staring at computer screens again.

 

I'll be taking some time to look through the thread and get a handle of current issues/bugs/requests/etc. But it looks like things have been busy, so please, let me know about any bugs/requests/etc that you have relating to clothing converter.

 

It will help me to prioritize and figure out what I want to tackle first.

[myuhinny]

Cool Welcome back.

[Vaelorian]

OK. I'm Back.

 

Welcome back.

[ger4]

Good to see you again.

 

From what I remember, the current feature requests included a UV map conversion engine and some kind of a fix for "cleavage crash"  where the center of an outfit follows the breast line too closely.

 

I think this was hard to do, but I think some kind of anti-clipping algorithm was also being investigated.

[s.eden]

In case, an edit to the first post as to the latest release of the converter would be a good idea

Otherwise, yes, compilling the bugs since release of RC85a is mandatory.

[myuhinny]

I have found out by trial and error that a person can actually convert oblivion bodies to skyrim bodies. Some of the bigger breasted ones like BDM tend to get the breasts destroyed either they look like the model ran into a wall or the bottom part gets a nasty spot. I found a CBBE body that I took from a armor the body was already in the oblivion T pose. I changed the CBBE body textures to match the body textures of the armor that I was going to test on which was the 7BASE body non TBBP version from the zipsuit mod.

 

I first ran a lattice using the CBBE/7BASE as the to and from then converted the CBBE body to the 7BASE body. Next I took the converted CBBE body and ran a lattice using the converted CBBE body and BDM oblivion body after the conversion I made a copy of the body_0 from the zipsuit folder with the skyrim armors at the sides and  used the CBBE/BDM lattice it converted it pretty well for the most part a few spots that a person would have to use blender/3dmax to fix. If there were more bodies in the T post pose a person could convert oblivion bodies over to skyrim quite easily with some minor tweaking being needed for the mesh.

 

[gerra6]

If there were more bodies in the T post pose a person could convert oblivion bodies over to skyrim quite easily with some minor tweaking being needed for the mesh.[/font]

Gotcha Covered.

 

http://www.loverslab.com/topic/13140-pose-converter-076-992013/?p=721431

 

That's a link to my experimental stand-alone pose converter. It should be able to interconvert between any Oblivion, Skyrim, or Fallout poses.

 

Works great for everything but the fingers. The fingers are OK...I just need to spend some time tweaking my template poses in Blender to get a better effect (time consuming and, for me, mega boring).

 

Now, if anyone would like to play around with my template file to see if they can generate better conversion poses, send me a PM. I'll send the blend file and a brief explanation of what would be needed.

[myuhinny]

I tried it earlier today after reading your post. The first time that I ran it I thought it had crashed because it did it so fast but upon looking into the folder it showed the nif with the new pose. Seems both BBP and NON BBP have trade marks that convert well. BBP converted to a oblivion big breasted body the breasts will convert nicely but other places like the arms/shoulder will come out warped. NON BBP converted to a big breast oblivion bodies will become half flat but everything else will convert nicely now if only there was a way to combine both to make one body.

[gerra6]

Hmm. I think I have an idea about what might be going on.

 

Could you send me the files you were working with. I have an idea, but I'll need to test it (and view the armatures/rigging) to be certain.

[Alwayon]

After two days of learning & trying this tool, I have two questions to ask for your kind help.

 

1) I am wondering whether I understand this right ?

 

Suppose that I generate a lattice between SourceFemaleBody_0.nif and TargetFemaleBody_0.nif ... then proceed to convert clothing ..

 

When converting a morph meshes pair, this tool will convert Armor_0.nif using the above Source0 to Target0 lattice. So we get ConvertedArmor_0.nif

 

Then it will find a morph info between Armor_0.nif and Armor_1.nif, then apply this morph ON TOP OF ConvertedArmor_0.nif to make

ConvertedArmor_1.nif

 

Am I understand this correctly ?

 

2) I am wondering whether there is a way to use this tool, not to convert between different bodies, but to generate weight 1 out of weight 0 armor or vice versa ?

 

I've tried generating a lattice between Femalebody_0.nif and Femalebody_1.nif and use it to convert Armor_0.  The result seem to be almost perfect in every way, except 2-3 vertices seam on both inner wrists. (Look like no feet or neck seams)

 

Aside from these slight problem or confusion which is very likely on my side, ..... I found this tools work likes a dream. The result is precise and amazing when I got a lattice right.  Thanks so much !!!

 

Edit : The 2) question has been solved.  I re-run the tool with Vertex Index Search / Node target = 1 and the result was perfect. 99.9% match of the original Femalebody_1.nif too, no seams at all.  I guess a very small difference around breast area come from skeleton model between the armor set and a newly rigged body.

 

[hemmicane]

I have downloaded the portable package 72 and the upgrade. Followed the directions. When making the "Lattice". select my from file, then my to file. Click okay. Program seems like it's doing what it's suppose to do. When done, check my "Lattice" file. 0KB in size. I don't think this is correct. This is from the "Make Lattice" windows batch icon.  Think that also makes it so that running convert clothing batch does nothing.

[myuhinny]

@hemmicane

 

Don't use the .72 version use the .85 version found in one of the later pages as that is the up to date one.

 

*Edit

 

Post 409 page 21 .85 build of the clothing converter. http://www.loverslab.com/topic/27557-new-clothing-body-style-converter-release-candidate-v072-3-30-2014/?p=759447

 

[myuhinny]

@Gerra6 

 

Seems I made a few errors in my post guess that is what happens when you post late at night just before going to bed. Non BBP and BBP versions most convert quite well and look the same it is just the breasts that look flat/sometimes warped. The one I tested on was the bouncing bodices UNP and UNP BBP version. The one that I did that had where the breasts converted nicely while other areas did not was with the zip suit non TBBP version.

 

 

 

This is how bodies usually look if you try to convert anything with breasts BDM size or bigger when trying to convert the oblivion body to a skyrim body.

 

 

This is how the zipsuit body looks like.

 

 

Not sure why skyrim bodies don't like bodies like the oblivion BDM maybe it's a mesh thing or maybe a lattice thing. 

 

 

 

 

[hemmicane]

Okay. Downloaded .85. I'll give this a try, fingers crossed.

[gerra6]

Not sure why skyrim bodies don't like bodies like the oblivion BDM maybe it's a mesh thing or maybe a lattice thing.

Well, it depends a bit on what's going on.

 

Is that the result of running only Pose Converter on the meshes, or Pose Converter followed by Clothing Converter?

 

The reason I ask, the distribution of vertex distortion looks like what I would expect if I were to apply the Nearest Point on Surface Search between two bodies that had significant differences in the distorted regions. (Default mode of Clothing Converter if UV search is disabled)

 

That's why conversions between significantly different bodies with different UV maps is such a clusterfuck right now.

 

One approach that *should* work is a bit tricky, but hopefully might do the trick. It comes down to generating a good template.

 

I apologize in advance for the IKEA-esque instructions.

 

Let's say uv_1 is the UV of the body type that you want to convert from, and uv_2 is the UV of the body type that you want to convert to.

 

Let's call the body you want to convert from body_a1 (uv_1) and the body you want to convert to is body_a2 (uv_2)

 

First, find two bodies that are extremely similar. body_b1 uses uv_1, body_b2 uses uv_2.

 

Notee: The greater the similarity between the bodies, the better the results of this method will be.

 

Create a lattice that converts one body_b1 to body_b2 . Convert body_b1 into body_b12.

 

Create a lattice that converts body_b2 into body_a2. Convert body_b12 into body_b12a2

 

Now, create a lattice that converts body_a1 to body_b12a2 (same shape as body_a2 but with UV_1).

 

That should be your conversion lattice that should give decent results to convert from.

 

Now, if the problem is just Pose Converter, then it's a bug and I'll probably need to take a look at the nifs and figure out WTF is going on.

 

Note: I am currently in the early stages of refactoring my tool framework, but I am considering generating a tool that will automate the outlined process.

[myuhinny]

Your pose converter converts them just fine with no problems.

 

It happens when trying to run  a skyrim body oblivion body lattice. Bodies that have big breasts like BDM F cup and bigger will result in the above mess. Pregnant bodies if the body has a small belly bump converts alright but if the pregnant body has a big/huge belly the belly will implode into itself.

 

I have a CBBE body that I use when skyrim to oblivion bodies convert weird. I convert the CBBE body to the body that is in the armor then take the converted CBBE body that was converted to the armors body and place it into the from folder and use the armors body for the to file.

 

Then I run it again it works on some but the bigger breasted ones it doesn't as I have yet to find a body that can handle a body like BDM and other bigger breasted oblivion bodies other then the zip suit one and that only converts the breast nice but the rest of the body is a mess.

 

Maybe the zip suit body has more lattices then many of the other body and that is the reason the breast convert better or maybe I need to increase the lattice search when doing them. 

 

I have also tried using the _1 body but it converts the same as the _0 bodies do.

[myuhinny]

Think I have finally found a body that works alright for skyrim to oblivion BDM body.  Though the zip suit one has some many defects in the body I had to convert it to the other body then use the other bodies lattices or it just keeps coming out nasty looking.

 

http://www.loverslab.com/topic/15224-show-your-skyrim-counterpart/?p=924956

 

I did try the big pregnant body again with the above body but still get a implode of the belly. Maybe it needs to be done with a body mesh made for pregnancy will have to try that later this week and see.

[ratrace]

A little late (my greetings, that is - not your return), but glad to see you back here. And I hope you recovered well. Take it slow, though!

[Alwayon]

@Gerra6

 

I think I might found a small bug in this tool.

 

It seems like when I tried to convert SOME outfits that has very low mesh detail and cut-out body part underneath.  ie. Vanilla miner clothes. The edge of underlying body will fly out as in the picture.

 

It's quite strange since my lattice is working well with 95% of these outfit / body type.

Full body -> no problem, various cut-out body shape -> no problem, just few type of cutting out and boom ! the edge fly out.

 

[gerra6]

That's probably not so much a bug as a consequence of the low poly mesh being unable to be properly shaped to neatly fit the high poly mesh.

 

First, try with v.0.85.a http://www.loverslab.com/topic/27557-new-clothing-body-style-converter-release-candidate-v072-3-30-2014/?p=759661

 

It has a few improvements to the nearest point on surface algorithm that *might* help.

 

However, clipping is going to be a problem if you are converting a mesh with a low poly density to work on a body with a high poly density. It's the difference between wrapping a ball in a box vs wrapping a ball in a sphere.

 

The main available solution is to make a bigger box. Run Convert Clothing. Enable 'customize mesh' and run the tool.

 

Each time the tool processes a mesh in your nif, it will pop up a menu that allows you to control how much the vertices on the mesh get moved. The specific controls are a bit tricky, and require some explanation.

 

Padding: This is the most straightforward. Slightly oversimplified, it simply causes the entire mesh to expand by the selected amount (given in nif distance units).

Specifically, it moves each vertex the specified distance along that vertex' normal vector.

 

 

Magnify Movement: This is a bit more complicated, but it has its uses. When a given mesh is converted normally, each vertex is moved an appropriate distance to match the new body type. Some vertices move a great deal, others won't move at all. The magnify movement setting is a multiplier that is directly applied to any vertex movement. This can be used to exaggerate the transformation, or make is more subtle.

 

Growth Bias: This is the trickiest of all. And, it's probably not very useful. When this is enabled, the tool will attempt to determing whether or not a given vertex transformation represents expansion or contraction. The bias can be used to selectively favor expansion or contraction. Honestly, I'm not sure how useful this is. I may remove it when I finish rebuilding the tool platform.

[Alwayon]

Thank you so much for your detailed reply, Gerra.  It's highly appreciated. I'll keep trying it.

 

Please kindly forgive my English which may sound weird sometime ...

 

The reason I believe it might be a bug is that ..

 

1) I am already using 0.85a

 

2) I made my lattice by ... wrapping my destination body mesh around the source mesh first. So I have two exact meshes but only different shape , then I only use vertex index search to make a lattice.

Theoretically, they are almost like skin-wrapping morphing in 3ds  and should result in almost 1 to 1 conversion, which your excellent tools always give out the expected result.

ie. The converted NIF always has underlying bodies which is almost 1:1 match to my destination body, except a bit of minimal distortion in breast and nipple area ??

 

3) Today, I tried to do some further experiment by getting a bunch of vanilla bodies embedded in various vanilla NIF ... and put them into the same NIF file.  So some of them are full, some are only small piece of neck and wrists+angles, and some almost full but has tip of breast cut-out.

 

4) All of them is perfectly aligned as they come from the exact same body

 

5) Then I convert this single NIF using the lattice made in 2). And found that each pieces were converted differently ..

 

6) And the log file seems to says that the reason some under-the-neck vertices look jumping-out may come from the fact that they are identified as neck seam.  So, they say put, not morphing along at all.

 

7) I tried to confirm this by loading the result NIF into 3ds Max. And found that these weird-looking edge vertices seem to stay exactly where they used to be in source NIF.

 

8) This ever happened to ANY armor/clothes piece, even in the same NIF. Every other shapes always morphed along well. Some of them may have small problem area if the mesh resolution is too low. Only body mesh has this problem because it does require special seam processing, I guess ??

 

I am so sorry for long post, Gerra.  And would like to thank you again for this wonderful piece of tools. It is totally godsend.

 

Edit : I'll try to attach the before and after NIF, along with log and lattice file if these will help in anyway. Thanks !!

[gerra6]

Ah. Now I understand. Thanks for the detailed description.

 

Yeah, the cut out thing is probably a bug.

 

Basically, my tool has a non-manifold edge detection algorithm that does a pretty good job. The basic function is to precisely preserve pre-existing inter-mesh seams for meshes identified as skin (by material type, rendering info, etc). For the most part, it does a fairly good job.

 

However, when dealing with cut out meshes, things can get a bit weird. Essentially, only the neck, writs, and ankle seams should be preserved, so I added an algorithm that attempts to categorize each seam. Any seams that don't match a category are ignored.

 

Unfortunately, it sounds like it's miscategorizing the cut-out seams.

 

Would you be willing to send me the files? I need to take a look at what's going wrong with my categorization algorithm.

 

*Edit*

 

Hmm. I took another look at your log file screen shot. It's definitely treating the entire edge of the cut out as a neck seam. Now, one thing that the tool *should* be doing is only allowing a single seam (chain of directly connected non manifold vertices) in a given mesh to be categorized as a neck.

 

I'm thinking it through and I don't see any reliable generic methods I can use to consistently identify a neck seam vs a neck seam linked cut-out vs cut-out within the general upper torso region.

 

So...here's what I think I'll try.

 

Short term fix, I'll update the menu with an option to disable seam detection on a mesh by mesh basis. This should be a fairly simple patch that I can upload to this thread.

 

Bigger fix...I can update the back-end seam detection algorithm with an optional 'strict' mode that will detect the most common seams found in Oblivion, Skyrim, and Fallout nifs, but won't properly detect non-standard seams. This will take a bit of research, time, and testing.

 

Anyways, once I take a look at your nifs I should have a better handle on what's going wrong and how to deal with it.

[Alwayon]

Thank you very much for your time and patience, Gerra.

 

Here they are ..

 

 

Original and Result NIF.rar

Log file and Lattice.rar

 

 

Edit : And here are some actual Vanilla armor/clothes NIF conversion 

 

 

 

Some Vanilla ORIGINAL and RESULT.rar