impressed, and cant wait to give this a run

Kewl! Can't wait to try her out and make some morphs.

Anyway, I modified Bill's supplied "Spots" node for pigment variation a little bit and added a second one to imitate that typical faint bluish network of veins below the skin.

I think this together with a lo-res color map could already produce some pretty awesome skin.

Excellent work
have actually been testing out veins beneath the skin in poser aswell
brilliant render and lighting is superb
can not wait till shes done :D

i find to really test a skin shader you have to test a lot more than very ambient lighting situations.  i've used PR2 in low light situations, but i think i did some adjusting to do it.  it's hard to say, because i haven't tried the same character in the same scene that way, so i keep changing texture and target skin tone, too.    also, there's a tutorial here on using fastscatter with a control map to get certain places to do backlighting SSS right.  i think i'm going to try to incorporate that into my adjustment of PR2 soon.  i've had pretty negative dealings with fastscatter in the past, but maybe it's not so bad if you limit it to just a few places (ears, nose, edge of the fingers, a touch on the lips, etc.). 

personally, i presently add veins in postwork because it's very important for me to control their visibility, orientation, length, and thickness based on where they are on the figure and (for artistic reasons) the lighting.  i mention it because it would probably be better with a control map.  i probably wouldn't handle veins procedurally myself, because i find i need to change veins when i paint them.  i would probably want more control than a node can give.

but, as i think you mentioned you want to, i eventually would like to be able to paint photoreal skins, as seems to be suggested and done by the experts at CG society.  perhaps with various shader-controlled layers (freckles and sunspots, veins, burned in AO, etc.).  i know from my PR2 experiments that i'll want to make good SSS maps.  i don't know when i'll ever get to that, because right now i'm experimenting  with modeling and UV mapping Blender (think a toddler with blocks) and wishing i knew python.  but i figure Antonia will give me the opportunity to try my hand at a lot of new 3d development tasks. 

colbaltdream: I definitely agree that control maps would help. Also, I would use that node only for that network of small veins, not the larger, more visible ones. Basically the point was to introduce some irregularities to simulate the local variation in skin tone that we expect to see on real skin.

It would be very exciting to get some state-of-the-art skin-rendering magic into Poser figures. I admit to being a complete amateur in that area. But that's fine, because I'm an amateur in any other area of CG, as well. :biggrin:

ODF when do you think Antonia is ready? i just cant wait to texture her:)

It's not easy to tell. Very soon, I hope. But the last bits of a project have that sneaky tendency of taking much longer than one thinks. :biggrin: I have no experience with the beta testing process, either. Sometime in February would be nice, but I can't promise that.

Any hopes of me making a figure rigged for Poser has just been blown out of the water.

What an ordeal. My hat's off to you odf.

Good luck!  I hope to have a live internet connection in the next month or so. The progress is impressive. I have followed this thread with eager excitement.  The evolution of this figure gives hope for the rest of us. Keep up the good work!

Quote - Any hopes of me making a figure rigged for Poser has just been blown out of the water.

What an ordeal. My hat's off to you odf.

Well, that's why I'm not doing it on my own. Without phantom3D doing the rigging I can't even begin to imagine the headaches I would have had. :biggrin: I mean, it was certainly a good thing that I did my own clunky rigging first and learned how it's done. But also a big relief to be able to hand it over to someone with more talent and experience.

Wow, really impressive work and a very educational read watching it all come together, really look forward to her release.

Now I've noticed that Poser actually doesn't care. If you apply a texture map, it basically treats it as a 1-by-1 tile and I get my figure textured as if the UVs where all done the conventional way. The only gotcha I've seen so far is that when you apply position-dependent nodes in the material room, you have to use the correct reference point. For example, when I tried out a procedural iris shader by bagginsbill, I had to move the UV center of the iris from (0.5, 0.5) to (0.5, 1.5), because that's where it is in my map. But you always have to make sure you use the correct reference points in such cases, and it shouldn't really matter that much whether those are in the [0,1] x [0,1] square or not.

With me so far? Well my question is: is there any good reason anyone can see why I should smash my UVs together into that one white square instead of leaving them the way they are?

Nice idea!
Does it cost the same on memory?
Also does the value of moving the applied map depends on the texture size or it is always the same number in this case it will be 0,33 for example? (sorry I haven't used that feuture much, although I know it's existance)

Other programs may have trouble with the UV coordinates lying outside the first unit square. Poser clearly doesn't and that's a neat thing.

As far as I know, for Poser, this has no performance or usability impact whatsoever, since everybody leaves Image_Maps tiled anyway in almost every shader.

Quote - Other programs may have trouble with the UV coordinates lying outside the first unit square. Poser clearly doesn't and that's a neat thing.

As far as I know, for Poser, this has no performance or usability impact whatsoever, since everybody leaves Image_Maps tiled anyway in almost every shader.

Thanks! That's what I thought.

Wings3D, Blender and UVLayout are fine with the UV coordinates as I have them now. Obviously, I need to check UVMapper and ask around some to see what other people might be using for texturing and such. I can easily write a Python script that shifts all the UV coordinates back into the first unit square and writes the result out as a new geometry file. Restoring the original coordinates so that all the shaders will still work is a tiny bit more tricky, but can be done. So unless I hear about some commonly used program that can't handle my current layout, I will probably leave it as it is.

pitklad: Are you referring to the square positions in the picture I posted? Those are in UV coordinates, so every square is one unit wide and one unit high. If you apply a texture map, it will normally be mapped onto the first unit square, independent of its resolution. Or in tiled mode, it will be mapped onto each unit square. That's the nice thing about UV coordinates. They don't care how many pixels you have in your texture maps.

So this also means that if some people want they can make one texture for head and body and use a factor number to adjust the map on the right position of the uvs?

Quote - So this also means that if some people want they can make one texture for head and body and use a factor number to adjust the map on the right position of the uvs?

I think that should work, although I haven't actually tried it yet. But as far as I know, it would require to set up a texture2d node in the material room, so I presume Poser 4 and D/S wouldn't be able to use that texture properly.

Which reminds me, I should probably check how D/S deals with this.

By the way, is anyone still using Poser 4? I use Poser 6 myself and was planning to set up Antonia so that Poser 5 and up are quite happy with her, and nothing breaks too badly in D/S. Do I need to include P4 in the "nothing breaks too badly" list, or can I safely assume that its time is past?

is there a benefit to this at all?  because what it seems like is anyone who wants to work with her textures has to memorize coordinates just to apply a texture or work with any UV based procedurals. why make it harder to work with her than any other figure ever created for Poser?

Quote - is there a benefit to this at all?  because what it seems like is anyone who wants to work with her textures has to memorize coordinates just to apply a texture or work with any UV based procedurals. why make it harder to work with her than any other figure ever created for Poser?

That's what I thought, too. But as I explained earlier, I could not see any differences between working with this kind of UVs and the traditional ones, except in very, very special circumstances. More precisely, unless someone tried to do wacky things like make procedural iris or lash textures, everything would just work normally. And people who do that should be able to deal with a slightly unusual UV layout.

The benefits are also very clear: it would be much easier to view and edit the UVs. For instance, if you are a texture maker, you can very easily open the geometry file with Wings or Blender and print out your own texture guides without having to hide or move overlap charts out of the way by hand. As I said, I haven't checked UVMapper yet, and of course I won't go ahead with this layout if I can identify any practical drawbacks at all.

Quote - For example, when I tried out a procedural iris shader by bagginsbill, I had to move the UV center of the iris from (0.5, 0.5) to (0.5, 1.5), because that's where it is in my map.

for me that doesn't seem like a very special case, but if you don't have to do the same thing for a regular image, then only people like me who like to play with the material room are affected.

oh, and i think it would be a better idea to give out templates with guidelines (see SnowSultan's freebies for DAZ products) to texture artists anyway. if you can.

Quote - > Quote - For example, when I tried out a procedural iris shader by bagginsbill, I had to move the UV center of the iris from (0.5, 0.5) to (0.5, 1.5), because that's where it is in my map.

for me that doesn't seem like a very special case, but if you don't have to do the same thing for a regular image, then only people like me who like to play with the material room are affected.

That's what it looks like. And if you're doing a procedural iris texture for a new figure, you have to take a look at the UVs and determine the center and extent of the iris, anyway - or of course, use measurements kindly provided by the maker of the figure.

I'm not feeling very strongly about this. If any practical drawbacks surface either in this thread or during beta testing, I'll bury the idea immediately. It just seemed foolish not to at least run it by the think tank. :biggrin: I think I'll check what UVMapper Classic makes of it tonight, and if that looks good, ask for opinions in some of the more specialized groups.

Quote - oh, and i think it would be a better idea to give out templates with guidelines (see SnowSultan's freebies for DAZ products) to texture artists anyway. if you can.

Yes, I'll definitely do that, too.

centers are generally in the middle, half way down and half way across.  if someone's managed to do an eye map where that's not true, but it's a single eye, it's kind of nuts.  but i was thinking more about general UV usage, and things needing to start somewhere different, rather than just, say, using a U or V node raw.  i wouldn't have to know any co-ordinates to, say, create a gradient on most faces.  i would if i needed to know the top was 1.5 instead of 0. i have used U and V nodes for various tricks before, and since i often have to keep other numbers in my head to remember to apply or transpose or whatever among materials, i personally and individually don't really want more of them.  that said, i think i'm not your average user. 

D|S, on the other hand, is a big deal.  its users are plentiful and vociferous.

and just to say, because i don't think it comes out in text, "if you can" isn't a challenge.  it's just recognizing that seam guides take time to make and you may not have that time.

Quote - centers are generally in the middle, half way down and half way across.  if someone's managed to do an eye map where that's not true, but it's a single eye, it's kind of nuts.  but i was thinking more about general UV usage, and things needing to start somewhere different, rather than just, say, using a U or V node raw.  i wouldn't have to know any co-ordinates to, say, create a gradient on most faces.  i would if i needed to know the top was 1.5 instead of 0. i have used U and V nodes for various tricks before, and since i often have to keep other numbers in my head to remember to apply or transpose or whatever among materials, i personally and individually don't really want more of them.  that said, i think i'm not your average user.

Fair enough. I have to admit it took me a while myself to figure out why the bagginsbill iris shader I tried on Antonia didn't work the way I expected. I think that is in fact a serious argument, because any advantages of mapping a figure that way will not be visible to the average user, either. And for my personal entertainment, I can always use scripts to untangle the charts.

Quote - D|S, on the other hand, is a big deal.  its users are plentiful and vociferous.

As far as I am concerned, D|S users should stick to - or possibly with - DAZ content. It's a promotional tool to sell more millennium figures and accessories. :biggrin:

KEEDING!!! Of course it would be foolish to release a new figure that didn't work on D|S, or at least downgraded nicely. Also, I hear that D|S can do subD, which would make the lo-res version of Antonia much more interesting.

I just learned a new word, by the way: vociferous. That doesn't happen all that often, anymore.

Quote - and just to say, because i don't think it comes out in text, "if you can" isn't a challenge.  it's just recognizing that seam guides take time to make and you may not have that time.

Sure, I didn't read it as a challenge. Well, I guess I could always outsource that task. :biggrin: Or do the first release with some quick-and-dirty texture templates and put up the more detailed versions later on.

Quote -

Quote - and just to say, because i don't think it comes out in text, "if you can" isn't a challenge.  it's just recognizing that seam guides take time to make and you may not have that time.

Sure, I didn't read it as a challenge. Well, I guess I could always outsource that task. :biggrin: Or do the first release with some quick-and-dirty texture templates and put up the more detailed versions later on.

StitchWitch can do seams guide, the program is for texturing clothes but can extract textures templates from any object, but i don't know how it handle the unusual UV space

I own the program so i can give a try i you want, no need to spend time on this if a program  give the same result in one mouse click.

Well, UVMapper does not like 'out of range' UVs, so that's that. I mean, I don't have the faintest idea what that program is good for, but lots of people seem to use it, so I guess I can't ignore its existence, as much as I'd like to. :biggrin:

Quote -

StitchWitch can do seams guide, the program is for texturing clothes but can extract textures templates from any object, but i don't know how it handle the unusual UV space

I own the program so i can give a try i you want, no need to spend time on this if a program  give the same result in one mouse click.

Sounds awesome.

Quote - Well, UVMapper does not like 'out of range' UVs, so that's that. I mean, I don't have the faintest idea what that program is good for, but lots of people seem to use it, so I guess I can't ignore its existence, as much as I'd like to. :biggrin:

most often i use UVMapper to correct the grouping, it's easy to select some polygons and assign to a group, a least more easier than the poser grouping tools :biggrin:

Quote - > Quote - Well, UVMapper does not like 'out of range' UVs, so that's that. I mean, I don't have the faintest idea what that program is good for, but lots of people seem to use it, so I guess I can't ignore its existence, as much as I'd like to. :biggrin:

most often i use UVMapper to correct the grouping, it's easy to select some polygons and assign to a group, a least more easier than the poser grouping tools :biggrin:

Then again, a text editor and a blindfold are easier to use than the Poser grouping tool. :biggrin:

tried on Olivia G2, work seamlessly :biggrin:  a few mistakes but overall the seam guide is correct

The chest is textured, the head isn't - just in case you couldn't tell.

I'd help out here with the lashes but I'm about to leave for a long ski weekend.

See ya next week.

Here's what I have after an extra hour of so of fiddling.

the second one looks much better
but why not just use a trans map for the lashes instead of shaders???

i wouldn't know how to do it with shaders, but the eyelashes should be more random and unruly. also, there should be less  bottom lashes. if you want, i can do a trans-map texture for the lashes.

As I said, just proof of concept. It's fun to see what can be done just with shaders. Don't worry, I don't think hand-painted lash transmaps will become an endangered species any time soon.

In the meantime, I made a little bit of progress with my first texture. As usual, there are some fairly easy bits and some tricky bits. I'll certainly not produce outstanding quality here, but at least there's a fair chance she won't have to go skinless when I release her.

I think I'm done with the geometry now. The UVs have all been moved to square one, I've repaired a little glitch in the lash geometries and put the materials in a sensible order for Poser's dropdown menu. Once I get the rig from phantom3D, we should be able to go into beta testing.

this is looking very nice.

I guess this render probably shows the texture itself a bit better, without the shader-generated spots I had in my material.

look pretty for a first try, :thumbupboth: a least far better than my own experiment in texturing

Quote - look pretty for a first try, :thumbupboth: a least far better than my own experiment in texturing

nods I agree I think your skin is already looking very nice.

Impressive skin texture. I must admit I was fooled at first with the pic showing the head. I thought it was textured as well until I read your post.

I really like that your not adding specular highlights, shadows etc...I think some texture makers get carried away in detail and add too much photo reference into the texture which of course has specular highlights and shadows. It winds up looking off with different lighting than the references. I prefer a basic texture with subtle variations and let shaders drive other effects.
You can always add more detail to a base texture if need be. Not so easy the other way around.

Quote - Impressive skin texture. I must admit I was fooled at first with the pic showing the head. I thought it was textured as well until I read your post.

I really like that your not adding specular highlights, shadows etc...I think some texture makers get carried away in detail and add too much photo reference into the texture which of course has specular highlights and shadows. It winds up looking off with different lighting than the references. I prefer a basic texture with subtle variations and let shaders drive other effects.
You can always add more detail to a base texture if need be. Not so easy the other way around.

Well, this is made from photo references, but it's only 2000x2000, and I'm taking pains to edit out as much shadow as I can. I hate it when people make "high-res, photo-realistic" textures that only look right if the lighting is the same as in the photo references. As you said, a diffuse map should have neither shadows nor speculars. It shouldn't show pores, either. That's what bump maps or displacement maps are for.

And, yeah, bump maps and the like need to be hand-painted. You can't extract them from a photo.

I've never seen proper photo refs for cgi skins. What sells them or makes them believable to the casual buyer is the very thing that makes them unusable to a texturemaker with professional  intent. Sure looks cool though... photo in,, photo out I mean.

2000x seems just right for this low res mesh.

Speaking of low res: I was just wondering how much interest there would be in good multi-resolution support for Antonia. As I mentioned earlier, I'm basically developing her as a low(-ish) res figure and subdivide once to obtain the default Poser version. The grouping, material zones and UVs are all getting back-ported to the low-res version, so that poses and textures should pretty much work the same way on both versions.

Now I'm working on some software to transfer morph targets. By "working on" I mean that it's pretty much done and functioning, except for some special details I still need to address (specifically, "hard" edges and 3-poles, for those interested in the nitty-gritty). But I'm pretty sure I know how to handle those. Of course, on the other side of the medal, I could also use this to automatically generate super-high res versions that would take all the poses, textures and morphs just as long as Poser doesn't choke on them.

The thing is, I've been using Scala for these things, which is a language based on the Java virtual machine. So either I'd have to provide a - probably rather large - binary that would run on any Java installation, or just the raw source code that people would have to compile themselves after installing Scala on their machines. Which one it is would obviously depend on how much demand there is. Or, and that would obviously be the most practical solution from a user's point of view, I could take my existing code, translate it into Python and make it run from within Poser as a script. That way, it should eventually be possible to load a high-res morph and apply it to the low-res figure directly within Poser, or vice versa.

My feeling is that there'd probably not be a lot of demand for the low-res figure, though. So, even if the integrated Poser solution would be extremely cool, it's probably not worth the effort to translate a thousand lines of code and add all the extra stuff needed to deal with actors, Poser's internal geometry encoding etc.

So finally, here's my question: do you guys agree, or am I vastly underestimating the interest in low-res (or super-high res) figures?

Super Low was never my favourite since poser can make even mid resolution figures looks much higher poly if you use the smooth render option

Low Res would be for sure a very useful tool not only for background figure but also for setting up the scene, make simulations and everything and than replace with higher resolution figure for final render

I've done that with Victoria 2 low res and Victoria 2 before Poser 5, afterwards due to the smooth option  I started using lower poly figures for my scenes :biggrin:

I would even suggest a ultra low resolution figure if that would be possible ( with no great expectation for bending and rigging just something like a mannequin)