*stupid short edit time limit...

I should also add that the UV template you use should be the same resolution as the image maps you inted to use.

Quote - > Quote - > Quote - K, the new morphs are up on the dev site, however I was unable to upload the original unsquished morph files for some reason (I split them into two archives, just under 8 meg each). Not sure why that is, but if anyone needs them, I'll find a way to get them to you. I may need to put them on my own website :o).

Have fun.

Laurie

Don't worry. It's not too hard to pull a bunch of morphs out of a cr2 file with Dimension3D's PMD editor.

(It helps to know which morphs to look for, so common name prefixes are nice. But since I made all the original morphs, I should know which ones not to pick, shouldn't I. :laugh:)

I deleted all the ones I was able to delete that weren't mine odf....sorry bout that ;o).

Laurie

I guess I should also add that all my morphs are in the head and no where else. There's one that extends to the neck, but that's it. And all my morphs have LA on the end of the morph name ;o). That should make it easier.

Laurie

Wow Mike, that's a lot of explaining of not what I'm talking about. ;-)

The practical limit of a texture map in Poser is 4K.

If you arrange to use that space ineffectively, you will get pixellation and there's nothing to be done about it. For example, consider the uppermost thigh on v4 where it joins the hip/buttocks. Ther'e good 2 inches of thigh that is compresed into about 60 pixels on a 4K map. You can't put any kind of detail there at all. The effective pixels per inch comes out to like 10 in some places.

If you try to draw a smooth clean black/white transition, even a blurred one, in those spots, it will be obviously pixellated. This is also why the bump map on M4 on the upper thigh is total crap, due to UV crunch. Each tiny bump pixel covers a quarter inch sized chunk of his thigh.

Also, if the line we want to draw is less than 5 degrees off from the UV map horizontal/vertical axis it looks like crap. It looks stairstepped.

Well I'm talking about UV mapping and image application in general, not Poser specifically.

I'd have to take a look at those maps to see what you mean specifically, but by "ineffectively", to me that translates to "distorted".
Ideally, your UV shells should each be the same relative size as the polygons they correspond to. A UV map that has that done perfectly will be as effective and efficient and distortion-free as possible.

Then of course there's the issue of bending body parts and morph targets, both of which can cause texture stretching.

Mike, I'm talking about generally, too, not Poser.

Just draw something with some fine detail, like a tattoo, along V4's thigh/buttock using a 3D painting tool, where you don't have to worry about seams like in Photoshop. I don't have tools like that.

Stamp something detailed there and save it as a set of 4K image maps. Then try to render it up close.

I'm not seeing any real distortion in that area. Actually, I think it looks really good. I'm not thrilled with where the seams are, but the UV map is about as good as it gets.

Quote - Mike, I'm talking about generally, too, not Poser.

Just draw something with some fine detail, like a tattoo, along V4's thigh/buttock using a 3D painting tool, where you don't have to worry about seams like in Photoshop. I don't have tools like that.

Stamp something detailed there and save it as a set of 4K image maps. Then try to render it up close.

The problem with that is, 3D painting apps tend to interpolate the lines between vertices, not actually using the UV map at the same time.
But I'll give it a try in Modo, where you can paint in 3D and 2D at the same time.

Of course you're not seeing a problem - you're 30 inches away and you used big blocks.

Render close, with something that is detailed. Imagine you're trying to stamp the maker of a pair of stockings on the welt, like they used to do, in white. Try to read the words.

Or use a lace pattern. How does it look up close?

Well I don't have the time now to do anything elaborate.
You do realize you're questioning the very veracity of image maps and UVs, right? Perhaps I'm not the one to argue about it with. ;-)

I'll try something more elaborate later. I can paint fairly well when I take the time. ;-)

For now, I did something simple, at 1K, 2K and 4K. The image map used was the same for all three and is 4096x4096.
Rendered in LW, that is. Maybe Poser has other problems. I'll test that out with my more elaborate image I'll do later.
mj41865.home.comcast.net/TESTS/V4_Legs_1K.jpg
mj41865.home.comcast.net/TESTS/V4_Legs_2K.jpg
mj41865.home.comcast.net/TESTS/V4_Legs_4K.jpg

The "pixellation" you see in the small writing is a limitation of the brush. Drives me nuts, but it is what it is. I used Deep Paint 3D, because I remembered Modo sucks for painting across seams.

Quote -
Of course you're not seeing a problem - you're 30 inches away and you used big blocks.

The block size makes no difference. That's a "UV Checker", what is generally one of a few styles of industry standard means of checking UVs.
The size makes no difference, What you look for is distortion.

I am a dumb ass when it comes to poser file creation somewheres I am doing something wrong. I am thinking my problem is something I am doing in binary morph editor which I bought yesterday. I was wonder if you could elaborate more on exactly what you did there.

Quote - aella: I've uploaded your morphs to the developers site as PMDs. It's as I thought : once loaded into Antonia-96 and converted to deltas, they work fine with the latest Antonia.

Here's what I did in detail:

1. Opened Poser and loaded Antonia-96.
2. Loaded each of your obj files as a morph target.
3. Saved the character under a new name.
4. Opened BinaryMorphEditor and loaded the character file created in step 3.
5. In the selection window that popped up, checked all morphs loaded in step 2 (here it paid off that I gave them all names starting with "aella-") and pressed "OK".
   6) Saved a PMD file.
   7) Saved binary injection and removal poses.

Quote - I am a dumb ass when it comes to poser file creation somewheres I am doing something wrong. I am thinking my problem is something I am doing in binary morph editor which I bought yesterday. I was wonder if you could elaborate more on exactly what you did there.

Quote - aella: I've uploaded your morphs to the developers site as PMDs. It's as I thought : once loaded into Antonia-96 and converted to deltas, they work fine with the latest Antonia.

Here's what I did in detail:

1. Opened Poser and loaded Antonia-96.
2. Loaded each of your obj files as a morph target.
3. Saved the character under a new name.
4. Opened BinaryMorphEditor and loaded the character file created in step 3.
5. In the selection window that popped up, checked all morphs loaded in step 2 (here it paid off that I gave them all names starting with "aella-") and pressed "OK".
   6) Saved a PMD file.
   7) Saved binary injection and removal poses.

Here's what I think odf was proposing:

Load your morphs into Antonia 96 and then export each morph as a new obj morph so that you create new deltas. Those new obj morphs should, in turn, work in the new Antonia. I'm 90% certain that's what he was suggesting. I think he meant for you to bypass pmd injection altogether until you create the new deltas and have them loaded into the newest version of Antonia.

Laurie

Quote - ...Imagine you're trying to stamp the maker of a pair of stockings on the welt, like they used to do, in white. Try to read the words.

Sounds like a Susan Sarandon line from Bull Durham...

Quote - > Quote - I am a dumb ass when it comes to poser file creation somewheres I am doing something wrong. I am thinking my problem is something I am doing in binary morph editor which I bought yesterday. I was wonder if you could elaborate more on exactly what you did there.

Quote - aella: I've uploaded your morphs to the developers site as PMDs. It's as I thought : once loaded into Antonia-96 and converted to deltas, they work fine with the latest Antonia.

Here's what I did in detail:

1. Opened Poser and loaded Antonia-96.
2. Loaded each of your obj files as a morph target.
3. Saved the character under a new name.
4. Opened BinaryMorphEditor and loaded the character file created in step 3.
5. In the selection window that popped up, checked all morphs loaded in step 2 (here it paid off that I gave them all names starting with "aella-") and pressed "OK".
   6) Saved a PMD file.
   7) Saved binary injection and removal poses.

Here's what I think odf was proposing:

Load your morphs into Antonia 96 and then export each morph as a new obj morph so that you create new deltas. Those new obj morphs should, in turn, work in the new Antonia. I'm 90% certain that's what he was suggesting. I think he meant for you to bypass pmd injection altogether until you create the new deltas and have them loaded into the newest version of Antonia.

Laurie

Edit: if you need help with that just let me know :o).

Laurie

Click on this one and make it bigger, cause that smaller version looks a little wonky. The larger one gives a better representation.

Laurie

That's a nice body shape, Laurie! Love your work on Antonia.

Quote - > Quote - I am a dumb ass when it comes to poser file creation somewheres I am doing something wrong. I am thinking my problem is something I am doing in binary morph editor which I bought yesterday. I was wonder if you could elaborate more on exactly what you did there.

Quote - aella: I've uploaded your morphs to the developers site as PMDs. It's as I thought : once loaded into Antonia-96 and converted to deltas, they work fine with the latest Antonia.

Here's what I did in detail:

1. Opened Poser and loaded Antonia-96.
2. Loaded each of your obj files as a morph target.
3. Saved the character under a new name.
4. Opened BinaryMorphEditor and loaded the character file created in step 3.
5. In the selection window that popped up, checked all morphs loaded in step 2 (here it paid off that I gave them all names starting with "aella-") and pressed "OK".
   6) Saved a PMD file.
   7) Saved binary injection and removal poses.

Here's what I think odf was proposing:

Load your morphs into Antonia 96 and then export each morph as a new obj morph so that you create new deltas. Those new obj morphs should, in turn, work in the new Antonia. I'm 90% certain that's what he was suggesting. I think he meant for you to bypass pmd injection altogether until you create the new deltas and have them loaded into the newest version of Antonia.

Laurie

Yes, I think that's basically what I was suggesting. The workflow I described above is one way to achieve that. But I'm confused as what you mean by 'obj morphs'. Exporting new obj files will not work. One needs to create deltas based on Antonia-96 and export those deltas. One way to do that is to load all those morphs into Poser and create a new cr2. The new cr2 will then contain the deltas that can be extracted into INJ/REM poses.

For new morphs based on Antonia-114, the workflow is basically the same. The only thing to remember is to load the obj files as morph targets into whatever version of the figure they were made from. Once the deltas are created, they should be freely exchangeable (except maybe for very sensitive ones like nipple morphs - no pun intended).

aella, which step were you struggling with?

Quote - That's a nice body shape, Laurie! Love your work on Antonia.

Thankee ;o).

I have not yet begun to morph! lol

Laurie

odf, you asked about texture baking. I don't suppose you have DAZ Studio 3 Advanced? It has that feature.

Quote - odf, you asked about texture baking. I don't suppose you have DAZ Studio 3 Advanced? It has that feature.

Yes, I have it, although I'm still using 2.3 for my ongoing work. Good to know! That feature could come in quite handy.

@ odf

I've resumed my UV work, having lost interest in discussing V4 textures...

I have to say, you really laid this model out nicely.
I imagine a lot of people who aren't having to dig down deep into every nook and cranny, having to scrutinize every last edge, may not notice, but I really do like they way you put her together.

And the newer, more symmetric version is working out perfectly. I was really dreading mapping the inner mouth and teeth, but it's going far better than I was expecting it to, because of how neat it all is in there.

:thumbupboth:

Thanks! :blushing:

Well you're welcome. You definitely deserve high marks for this model.

By the way Laurie, I totally forgot about trying to do anything about mouth raise/lower morphs today.
Being that I'm playing dentist on the UVs right now, I suddenly remembered. ;-)

for one figuring out the .cr2 business sorry I am a real noob when it comes to poser files and content creation.

@ MikeJ that "hello world" was freaking hilarious, too much time spent learning programming languages?

I applaud your work, its great that someone is tackling an independent figure, but I have some serious reservations about one of the directions you are going.
I took a look at this figure for the first time today, with a mind to how easy it would be to create conforming clothing for, and no offense but it's going to be a nightmare. The JCM work hides a lot of the bending flaws but that's the major problem, you relied heavily on morphs to optimize the joints, and the amount of sculpting you used in those jcms is massive.  for clothing to work it needs to share the skeleton, so it's going to share those same bending flaws, and is going to need the same amount of massive resculpting to get it to work on the figure and look smooth. Than when you get into matching full body morphs it becomes even more of a nightmare, because calculating the directions those deltas go with both the JCM and the full body morph and getting the clothing to bend in the same way with the FBMs will be torturous. It's doable I'm sure with MLP and reverse deformations and a lot of patience but it won't be easy.

Quote - @ MikeJ that "hello world" was freaking hilarious, too much time spent learning programming languages?

Was the first thing in HTML I ever learned to do, same with the first Python tutorial I ever followed. ;-)

wdupre: I agree with everything you said. That's the direction I'm going, and if you'd prefer a different direction, you're very welcome to take the figure and give her a rigging more to your taste.

Poser - realistic bending - ease of use. Pick any two.

Edit: That said, if you have any suggestions as how to improve the bends pre-JCM, I'm happy to hear them. phantom3D did most of the rigging. He's one of the best riggers for Poser figures I know, and I wouldn't dream of correcting his work. But if you have any particular ideas that sound promising, I might give them a try.

I just did a full-body morph on Antonia, just wanted to be a little bit creative/ambitious.  I hope to post a picture of her shortly.

JB

P.S.: I used the Antonia114

wdupre: Another thing. You don't have to use the JCMs, by the way. My plan was to add some dials or poses that would disable certain groups of JCMs. I people want to go that way for conforming clothes, I can go and reduce bulges pre-JCM and things like that to reduce poke-through. If the bend is hidden under clothing anyway, there's no point in making it super-realistic. It is then up to the clothes maker to decide how realistic they want to go.

The next potential step would be to go to 8192x8192, cut out just the bits we are interested in and place them on one or more 4096x4096 map. I imagine the scale and offset parameters of the image map node could be used in connection with such a map.

I'll post the new height map to the developers site, so you can play with it if you want. I might be able to do a 8192x8192 overnight if you're interested.

PS: I'm not sure I understand what you're trying to say re V4's mapping. Could you post a render that show's what you mean?

Over at RDNA, they're saying that the 4096x4096 limitation isn't the case since P7, which is why the TerraDome sky dome uses 8000px maps.

Quote -
PS: I'm not sure I understand what you're trying to say re V4's mapping. Could you post a render that show's what you mean?

Well good, it's not just me then. I was going to ask the same thing, about posting an example. I have no idea what he was trying to say.

Quote - Over at RDNA, they're saying that the 4096x4096 limitation isn't the case since P7, which is why the TerraDome sky dome uses 8000px maps.

Woot! I'll have to try that.

Anyway, I just noticed that I had rendered the height map with antialiasing facepalm. Filtering of any kind will distort the details big time. So I'm re-rendering a the moment and will post the new results in a bit.

Okay, the new height map is on the developers site, under textures.

I had to remove some old files because we hit the memory limit.

Edit: speaking of limits, I think I exhausted my 3GB bandwidth on the file locker for this month. Anyone who's waiting to download Antonia should wait until September or find someone who already has her.

If she stays this popular, I'll see if I can set up a torrent. Also, HostingCompanyIShouldProbablyNotMentionByName, here I come. :laugh:

I have noticed while using SVDL's MorphFromOBJP6 python, it doesn't save the INJ/REM outright, so I don't know about that.

I'm still playing with the possibilities, and I'm liking it.

Thanks for viewing

Quote - wdupre: I agree with everything you said. That's the direction I'm going, and if you'd prefer a different direction, you're very welcome to take the figure and give her a rigging more to your taste.

Poser - realistic bending - ease of use. Pick any two.

Edit: That said, if you have any suggestions as how to improve the bends pre-JCM, I'm happy to hear them. phantom3D did most of the rigging. He's one of the best riggers for Poser figures I know, and I wouldn't dream of correcting his work. But if you have any particular ideas that sound promising, I might give them a try.

No worries, I figured I would give my input from a content creator's point of view, not knowing if anyone has actually assessed the figure based on how easily it could be supported yet.
as far as any rigging suggestions I will take a look as time permits and see if there is anything that catches my eye.

Quote -
No worries, I figured I would give my input from a content creator's point of view, not knowing if anyone has actually assessed the figure based on how easily it could be supported yet.
as far as any rigging suggestions I will take a look as time permits and see if there is anything that catches my eye.

It did occur to me that the extensive JCMs might pose a problem when making conforming clothes. I don't see any way to achieve realistic bends without JCMs, no matter how good the rigging. Since my main interest is in realistic nude figures, that's what I'm pursuing. But as I said earlier, if I can make things easier for clothes creators without compromising that goal too much, I'm happy to do it.

Ultimately, I hope that we'd find ways to get a greater variability into our Poser figures. If the JCMs don't work for you, throw them out or replace them with another set. If you don't like the UV mapping, use a different one. And so on. As it is, these things are tedious. You need to create a new cr2 and possibly hand-edit it. Pose files can only do certain things, and Poser only saves certain things into them, so more hand-editing is required. Then of course, the greater the variablities, the harder it will become to keep track of dependencies between different packages. There are methods to overcome most of these problems, but they are somewhat scattered and there does not seem to be a useable, unified system anywhere on the horizon.

I think the big piece of progress we're making here is that now we have a figure which does not have to hide behind the ones made by DAZ, Smith Micro and others, but which is completely free of copyright restriction, so anyone can bake their own version and we can potentially have a lot of experimentation and innovation.

Even, I`m pretty sure there will be no problem at the end to provide a set of magnets that "transfer" the JCM's to a certain degree to clothes so clothes creators are not forced to rebuild the "deformations" made by JCM's, somehow like the magnetize poses for some figures.

Silly technical question here...

But Poser only needs vertex group information in the OBJ file when substituting one figure's OBJ for another, right? No polygon groups needed?

Quote - Silly technical question here...

But Poser only needs vertex group information in the OBJ file when substituting one figure's OBJ for another, right? No polygon groups needed?

No, it's exactly the other way round. You need polygon groups, not vertex groups.

You also need the proper 'usemtl' lines.

OK, thank you Olaf.

I may be wrong, but I don't think the obj format even supports vertex groups. Or if it does, they are not commonly used.

Oh, it most definitely does. Vertex groups and polygon groups.
In Poser objects I've noticed the vertex groups are almost always the same as the polygon groups - same names, same corresponding points. In the DAZ models that's mostly true.

But OBJ can also support "overlapping" vertex groups, while poly groups can't overlap. Meaning a point can belong to any number of groups, but a polygon can only belong to one group.

The Poser figures made by E Frontier have overlapping groups. My theory is they did it like that to prevent people from making their own morph targets so they had to buy from them. Maybe I'm just paranoid. ;-)
It did seem kinda sloppy to me though, if not intentional.
And I still was able to make MTs, after I figured  out what they had done.

I don't know if the new ...err... "humans"... included with Poser 8 have that going on though.

BTW, your Antonia model OBJ shows up in LW with both polygon and vertex groups, and the groups correspond to each other, as in the case of the DAZ models.
Modo 302 as well.
As for commonly used or not, I use them all the time. Excellent way to make selections, for one thing.

Quote - BTW, your Antonia model OBJ shows up in LW with both polygon and vertex groups, and the groups correspond to each other, as in the case of the DAZ models.

Those vertex groups are made up by LightWave. They don't exist in the OBJ file. I happen to know what's in my OBJ files, because I make them with a program I've written myself.

Well I don't know the technical details of how it all works, but I can see that.
Still though, the OBJ file does support vertex groups. Obviously it's not just LW making it up when some models import with those "overlapping" groups while others don't. If they didn't exist in the OBJ to begin with, then LW would simply make all poly groups into corresponding vertex groups and there would be no overlapping.

So whatever they at E Frontier used to model must have been exporting their vertex groups along with it.

If your exported model doesn't contain vertex groups, then LW is making all the poly groups into corresponding vertex groups on import, for whatever reason. Fine with me, it's damn convenient. ;-)