Hello I am working on Nea 3 and there are some black spot right across the edges where the groups of the hip and the Thigh meet Nea is based on Posette who also has that artifacts however after reshaping the body the problem seams to get worse Does anyone know why this happens and if somehow could be eliminated?

Might be in the UV map believe it or not. Just as an experiment subdivide the welded mesh and have a look again. I just encountered some weird dark artifacts on a figure I'm working on. Wasn't any kind of polygonal issue, turned out to be an artifact at a spot on the UV map where a cut stopped and edges flowed into 2 wings of the symmetrical chart. I fixed it by extending the seam cut up a couple of polys. Subdividing also seemed to help but the mesh was already at a final level of division so it wasn't really a fix option. The artifact looked like a dark grey bruise. Of course it could be something else but just pointing out that it's a possibility.

Just had a look at posette, you're talking about the dark grey smudges in the crotch where the hip meets the thighs? Those artifacts look about the same as I encountered. It seems to be a problem in rather low poly meshes, or at least that's my suspicion, as the mesh I'm working on is rather low poly too. It's something to do with how the UV is cut in relation to how the polygons are flowing in an area, in particular where a flow merges, not necessarily a pole but an area where say a tri or a triangular shaped quad merges and the proximity of the cut to that area. This is just my guesstimates but I was able to fix the problem as I said in my previous post. Not a biggie on your own mesh, early in the creation process. Probably a big issue with somebody elses mesh that requires content compatibility. Not sure how you would fix in a case like that.

The problem is that any change on polygons will make any morphs useless I used the weld command on uv mapper on every point that those two groups meet but no difference

BTW, I'm the guy who explained to you how to fix the smoothing issues on your arch model with control loops about 3 years ago (going by primorge then) so you might take into consideration my analysis of the problem. :) https://www.renderosity.com/mod/forumpro/?thread_id=2864872

Yeah, like I said if it's not your original mesh you're kinda stuck with the problem in terms of pre existing compatibility. Why I don't bother fussing with other peoples figures anymore in terms of public customization. Too much hoop jumping, I'd rather just DIY. But... this whole economy is basically built on slobbering all over somebody elses art/models and selling it sooooo. No offense, Pitklad. I love Posette and your NEA project.

Also, welding is just connecting 2 parts of an already troubled uv map, or perhaps depending on the cuts doing nothing at all in relation to the uv so I'm not surprised that your welding in UVMapper didn't fix anything. My suggestion was to weld and subdivide to see if that helped. Strictly as a narrowing down process.

Oh!!!! Thanks again for than and also thank again for now! You are right about making a new mesh etc sometimes the time to correct issues like this take longer But this is part of the NEA concept idea from day one It is a fun project that pretty much forced me try and create new thing in/for poser I don't work with poser those days but I want to finish this project at a point that satisfies my eye The help, comments, attention and kind words is just the motive for me :)

It seams that even the subdivided mesh still has the same problem, so this has to do with the vertices placement...! update for nudity flag...rndr.jpg

Keep at it man, I think NEA is an awesome project. You of anybody I know of in the Poser community has done the most to expand the possibilities of that figure. I'm sorry that I don't know of a viable solution to the problem.

Looked at your image. Sigh, well like I said, wasn't totally certain. Was only going by my experiences with a different mesh. Funny how good that model looks with another level of division. I still think it's a uv related issue though. Shame some more veteran figure creation people don't seem to frequent the forum anymore, curious myself about the exact where's and why's of the problem. Let us know if you find out anything please.

Well unless someone gives a solution I'll just have to get over it and move on the next problem :)

Which Poser version are you using?

Sometimes, there can be an issue with deformers with lower poly meshes. Also, from a discussion I had awhile back ago, Poser does something funky with the first vertices between groups that can prevent them from deforming in what some would deem an "appropriate" manner when morphed. It was highly technical (to me in regards to Poser's specific rigging) and what I basically took away from it was: At the borders of groups, if you encounter what appears to be a vertice anomaly and all else fails in the way of common fixes, this is probably what is causing the problem. :) Unfortunately, there is no fix for it, since it's intrinsic in how Poser handles multi-grouped figures. (The conversation was in the UVMapper forum, IIRC, a year ago+ or so?) If that is truly the problem, the issue is probably due to some shading anomalies caused by the confused border verts being told to do things that Poser will not let them do... or sumthin'..

Is the problem area on a group border? If not and not on a material border, then it's likely a deformer issue, which you can fix, if you can edit the deformers, that is. If on a material border, then it's a materials issue, obviously. I don't suppose you remapped it, so it's not a UV issue as long as the morphs where not made in a program that will alter the UV based on topology changes, which... would be very weird.

The pic doesn't tell me anything. I have an old version of NEA and could look at it for clues. Just make sure that any deformers are properly set up for any new morph, the rigging is right (reworked if necessary), UVMaps haven't been stretched by some evil do-nasty program and your material settings are correct across all materials to prevent material border issues. Try setting deformer strength in the main Object menu to zero, just in case any built-in deformers are causing the issue - That'll give you quick feedback for those.

pitklad posted at 12:35AM Fri, 04 December 2015 - #4242298

It seams that even the subdivided mesh still has the same problem, so this has to do with the vertices placement...! update for nudity flag...rndr.jpg

Doubtful. If you're experiencing the exact same issue with a new sub-d, then what could it be? It's not the verts, since you just changed them with the sub-d. What didn't change? The deformers/materials/rigging/uv/etc... Sub-D it x3 and see what happens. If the anomaly is still there, then it's not just a shading effect due to verts being placed in a certain way relative to the lighting/camera - It's something else.

@Teyon The render was made in Poser 6 The issue is visible both in preview and render, less visible if the light is not on the side of the model

@Morkonan no material border and no deformers on the figure just group border The same problem exists on posette also so you can see it even if you don't have Nea

Hi, apart of all tips given, you can try , if you are using a ibl light or with ambient oclusion for render, go to light properties, then click in the "Scene AO options" and try to put a upper value for "Bias", then try it.

@adh3d No the lights are not ibl and no AO activated

Maybe slightly out of alignment vertices that when welded in poser cause shading artifacts. I encountered the same thing with Akatora's free Decoco figure, I worked out a geometry switching version with a new chest, abdomen and hip. Same figure, same morphs, dial in altgeom with the fixes... no broken morphs (these area's didn't have any to speak of but with geometry switching you can dial in completely different geometry with more loops and it's own morphs). I used Wings 3d's vertice function 'absolute commands; snap' and meticulously realigned each body part group seam vertice. It worked as the fix. You could bring Nea's mesh into wings and check along those borders to see if the vertices that are meant to be identical (occupy the same space) are out of alignment. You'll be able to tell by selecting a vertice in the hip seam, view aim, zoom in on the vertice, and move your mouse a little... if It's twin vertice is out of alignment it will select easily and you'll see if it's not occupying the same position as its weld counterpart in the neighboring group. Run this check on the unwelded mesh of course, which is easy because wings forces you to jump through hoops to get a grouped welded mesh in one piece into it, lol.

If the problem ends up being out of alignment vertices you could realign them and apply that fix as a fbm, and you could leave it dialed as always 1, hide the dial, mess with the initial values or force limits or change the key value 0 0 1 in the .cr2. Little fuzzy on that last bit, I'd have to sit down with the .cr2 text and fuss with it to refresh my memory.

Another way to tell if the vertices are out of wack is to export the mesh with all options unchecked, reimport with weld identical vertices checked. After you've done this, run the smooth brush over the former seam areas. If the seams part then the weld didn't take, which indicates that the vertices do not occupy identical v coordinates and weld identical could not weld those vertices. As an aside, I find weld body part seams on export to be pretty useless or at least not as revealing as weld identical on import.

I'll run Posette through some tests later on tonight when I'm Posering... I'll let you know if I find anything odd.

Well one thing is for certain, the way that the Poser4 nude woman's groups are spread out through the .obj text is an absolute mess. Importing her into wings showed an import group count of 4815 groups, probably because all of her body parts are sectioned into countless same named mini groups. Your Nea .obj is alot neater (file says it was run through UVMapper so that's probably why it is). All those groups are making wings lag like crazy, which is funny cause it's a low poly mesh and I've gotten wings to handle upward to 800000 polys before. Anyway, brought the mesh into both wings and zbrush and everything looks ok with the vertices as far as the seams go. I don't know, I'm totally stumped on this one. One thing you should be aware of, Zbrush is indicating 26 Ngons in your Nea .obj, whereas an import of the original Poser4 Nude Woman .obj indicates only tris and quads.

pitklad posted at 9:34PM Fri, 04 December 2015 - #4242433

@Teyon The render was made in Poser 6 The issue is visible both in preview and render, less visible if the light is not on the side of the model

@Morkonan no material border and no deformers on the figure just group border The same problem exists on posette also so you can see it even if you don't have Nea

I took a look at an old copy of NEA.

As long as you're using the same groups/CR2, you're going to have this problem. The grouping on this figure is poopies... (At least, today it is. Way back when, it was probably fine.) For Poser, that's exactly the worst place to create group separations for that joint. (IF you don't want butt-bumps, that is. :) )You are dealing with the issue that I mentioned, above. (The single-vertice group-border issue. Search the UVMapper subforum where this was discussed, briefly.) You can subdivide all day, mostly, and the CR2/grouping/rigging is going to cause this problem. (However, using Snarlgribblies sub-d, inside Poser, effectively just creates a single-grouped, sub-d'd, object, which renders just fine, since it doesn't have this problem - No groups.) However, extreme sub-d might reduce this problem to a finer area, but it could also emphasize the effect in that area, yielding something like a black line in a high sub-d version, where you only have some slight smooth-shading defects, right now.

Unless you retopo'd it, which you can't release directly, since it's Posette, IIRC, there's nothing you can easily do to remove the effect you're seeing. Even then, you'd also have to regroup it, basically creating an entirely new figure. That's not impossible to do, really. But, it is likely impossible to do for sale, since you'd have to require the original and use the watchamackalit file-compare-unlocker-thingie to unpack/prove-prior-ownership... stuffs.

My advice? Regroup the figure, retopo it, and include instructions for users on how to apply your new .cvs file to a COPY of the original figure, using free UVMapper, and how to redirect the CR2 to read the new object file, using a text-editor. That isn't difficult to do and there is precedent. (IIRC, this same process was given out on the DAZ forums, along with files, in order to provide texture compatibility for "Kids" figures and V4 texture sets. Easy peasy, no cheesy. ;) )

Of course, that's a bunch of work. But, if you want to do it right and correct the issues with the original figure and its groups, that's what you will have to do, else some of your dedication towards resculpting and improving NEA is going to be wasted.

@EldritchCellar I 've tried already moving the vertices of the borders on Wing so they have the same position, still the same results.
Also the mesh is brushed on Zbrush so this should have been smoothed.
The image on my previous post made it clear to me that this is not a mesh problem because the same obj grouped the same way looks fine as a prop, but when poser reads it as a cr2 file than the spots appear, so the problem is how poser welds those parts that have a sawtooth border

@Morkonan I'm sorry I should have introduced the whole project on my first message. Nea is a figure based on Posette , redistributed using the encode method for copyright reasons.
Backward compatibility is required so that all content for Posette will work for her and also all the morphs that I've created for the previous Nea versions
I totally understand that regrouping is the easiest way to correct -many- not only this issue but this would move the whole project to a different direction

I just realized that when I edited my first post to add the nudity flag I deleted the image
It's been a long time since I posted so many things have changed in the forum apperance-code
I know it is late but I'll upload it anyway

Dunno what version of poser you are using but this issue dissapears if you change it to Poser Unimesh. So yes, it's a cr2/rigging issue.

@ghostman the examples are all done in Poser 6 Excuse me for my ignorance but what a Poser Unimesh is?

Didn't realize it was called unimesh but it's a newer skinning method where it doesn't rely on welded groups but acts more like vertex weighting. Or at least that's how I understand it. In any case a definite improvement over the legacy weld/group/spherical falloff method from my understanding. Teyon discusses it briefly in a thread buried round here somewhere.

So here's the deal. It seems there's a bug in the way Poser sometimes handles normals. It doesn't always show up but it has been around awhile. We've recently discovered it thanks to a similar issue on the new figures. It's something we're working to resolve but it has been my experience that in many cases the issue is not with the model (though sometimes it is, which makes it harder to narrow down). So if you're seeing a black spot on your mesh in Poser but not in any other program, it may be a Poser error...or not. Which isn't really helpful, I know but I wanted to let you know our programmers are looking into it.

The plot thickens.

So I guess the answer is that there is nothing I can do at the moment... Thanks everyone for the suggestions-replies and hope this will be solved some-way in the future

pitklad posted at 9:53PM Tue, 08 December 2015 - #4242917

So I guess the answer is that there is nothing I can do at the moment... Thanks everyone for the suggestions-replies and hope this will be solved some-way in the future

It's just a WAAAAY out-there suggestion, but you could try regrouping the object to move the group edge-borders to a place where this wouldn't show up as badly, preferably to a region with as many coplanar faces shared between bordering vertices as possible. It's not a standard "normals" problem, in my unschooled opinion, but a vertice normals issue with locked-in verts, due to what the CR2 rigging does, that get confused. CR2 says one thing, geometry says another, renderer goes splat... :)

You can try the following that doesn't require a remap, but it may not make her butt look any better, since it's comparatively low-poly:

Make all vertices coplanar relative to the group's bordering faces. :)

Basically, try to make sure that all four vertices for any given face on the group border of either group are all coplanar relative to one vector. To do that, you'll have to check whatever tools you have access to in your chosen modeling application. In Hexagon, for instance, you'd select the Face Tool, choose the orientation for the axes as "Selection" and then choose one border face. Then, you would use the Scale tool on the axis that represents as closely as possible the averaged face normal. (Pointing something around 90deg relative to the face normal.) Then, you would scale that axis down, to reduce the differences in position along that axis for all vertices for that face, yielding a truly "flat" face.

Other applications may have a "Make Coplanar" tool, but to keep your sculpts, you'd want to make the face's vertices coplanar, so use on single selections or use a toggle, if supplied. If you selected multiple faces and used such a tool, it might move them all coplanar to one axis, which wouldn't be good, since they'd all just flatten out relative to each other.

Basically, you want to eliminate any distortion to the face at all. Make each face completely "flat" relative to one vector AND coplanar with the bordering face in the other group. Why? To keep the renderer from looking at shading issues if at all possible. Will it work? I dunno... Try it, a little bit, on at least 12 contiguous faces, six in each group, so that at least two border faces will be surrounded by faces that, themselves, have verts all coplanar to one vector.

Obviously, if it was sub-d'd, you'd have a less noticeably geometric look at that border edge, since you'd have more real-estate to work with.

Question: What sort of renders are the ones that you have posted? Are they Preview Window renders or Firefly/other renders? Do these effects display in the Preview window? Do they display in both OpenGL and SReed Preview Window rendering options? That could lead to a quicker solution, depending on what sort of normal rendering they're using in the OpenGL mode. Can you make them "go away" using any other rendering mode?

Wings3d flatten by normal. You can also select by non planar faces, which in an organic model will be alot. I use flatten by normal quite a bit, but it's difficult to do it while maintaining a curve. There's also a distance tolerance advanced command. Have to do it a few polys at a time and perhaps a little tighten verts command. I've found that zbrush's smooth brush wreaks havoc on planar faces, it seems to basically sag the fill of a poly like a sagging canopy. Just something I've noticed, being conversational.

EldritchCellar posted at 1:48AM Fri, 11 December 2015 - #4243299

Wings3d flatten by normal. You can also select by non planar faces, which in an organic model will be alot. I use flatten by normal quite a bit, but it's difficult to do it while maintaining a curve. There's also a distance tolerance advanced command. Have to do it a few polys at a time and perhaps a little tighten verts command. I've found that zbrush's smooth brush wreaks havoc on planar faces, it seems to basically sag the fill of a poly like a sagging canopy. Just something I've noticed, being conversational.

Hmm... Thanks for that. I haven't used Wings3D in ages. It's time to redownload it and install it on this box. It's one of the nicest free proggies out there.

Those are P6 firefly renders, I tested the figure on Poser Pro and there are no black spots there. So it is an older poser version issue

pitklad posted at 11:32PM Mon, 14 December 2015 - #4243902

Those are P6 firefly renders, I tested the figure on Poser Pro and there are no black spots there. So it is an older poser version issue

...

Really? Hmmm, that's weird. Maybe it's a Smooth Shading issue with Poser6's Firefly as in it may have interpreted the constraints as out of bounds or something.

Is the old NEA-P4 figure the exact same geometry in the region you're working with? (My file display showing cr2 creation date of 8/27/2006.)

Sorry, I just rendered it and couldn't see the errors in PP2012Pro, either. I should have done that sooner. It's probably some smooth-shading dealing with P6 or a group border materials rendering issue, I would think.

Well, glad it's all worked out! :)

@Morkonan yes the old NEA-P4 figure probably has the same problem just like posette and the newest WIP Nea 3, as it seams on Poser 6 and maybe 7,8 haven't tested her there

Uhm... at the risk of seeming incredibly lame... did you look in the materials tab and make sure that the box for "normals forward" is checked? I mean, I've found with a lot of older Poser stuff, I have to do that if I want it to render properly.

@SeanMartin thanks for the suggestion but it doesn't seam to improve things