You know, I had not thought to check this at zero pose (and I should have).  I see the same thing.  In fact I see it on all Poser characters I've looked at (5 or 6 different just now)  I don't know if that's a real improvement for your situation though - I've rigged a few dozen conformers and I can see there is a real problem here.  Other figures' rigs do not have this kind of trouble.  Maybe I am way off base (I don't think so, but maybe) and someone else would like to rig some conformers for you and verify.

Sorry, I'm confused now. What kind of trouble is it that Antonia's rig has and other figure's don't? Do other figures show asymmetry in the zero pose or not?

By the way, I was wrong earlier. It's not just a uniform shift applied to the whole figure. I exported the figure in the zero pose as an object file, and I see that the centerline is still at x = 0, but the overall symmetry is broken.

Incidentally I checked this with an un-rigged OBJ - a few different ones including Antonia's base OBJ freshly imported - and they are perfectly aligned at 0 X and perfectly symmetrical with a check vs. a reference prop (cube), so it seems to be Poser's rigging that causes the small shift in position.

You may want to try rearranging the cr2. I've found that sometimes Poser works better if you have tree sections in certain orders.   It reads things from the top down in the cr2, so a little judicious rearranging can do interesting things.  I recently had an issue with a conforming robe because I had the left arm section before the right arm, and it caused no end of weird little issues.

Looking at the hierarchy tree in PFE, I see that you have the hip section, then both leg sections, than the torso and arms.  Most poser figures have it the other way round, hip, torso, arm cascades and then the legs.  You may find it changes if you move those areas around. Remember also to sort the figure Weld and AddActor sections the same way or poser may barf. Avoid the poser figure setup room where you can, its a screwy module. I mean, more than usual even for poser lol

as for the head scaling .. yes thats easy enough to do via a propagating scale dial (as long as you havent done the weird thing daz did with the back of the neck cut on v4) or with a simple morph. I'm almost inclined towards using a morph, since that will be easier to add into clothing ... but thats 6 of one half a dozen of the other.

As for cascading scale changes via ERC statements .. I've figured out a relatively painless way to incorporate them in clothing items so that they more or less work in poser and definately work in DS.  Gareees Lycanthrope Plus set has some extreme and quirky scaling built in to some morphs and I managed to transfer that nicely.

On the arm thing... once you have lengthened the upper arm, the rest should mostly be adding x amount to the xtrans value of all the child bodyparts. Not fun by any means, but certainly feasible.

Oh and on the alternate uvs question someone raised a while back ... I figured out a one-click pz2 solution in poser to swap differently mapped identical geometry live in poser. But you do need both obj files around, so it doesnt obviate the need for rt encoding, alas. I can send you an example pz2 if you are interested. Its less than 10 lines long :)

Lyrra

Quote - Sorry, I'm confused now. What kind of trouble is it that Antonia's rig has and other figure's don't? Do other figures show asymmetry in the zero pose or not?

Of the various clothing items I've rigged, all upper body pieces are normally fitted right against the chest.  I don't have this problem getting the shoulder bend to look symmetrical on other figures (even the DAZ Gen 3 ones, whose rigs are wacked out in a few different ways), just with Antonia.  Does this make it clearer?

So you mean other figures have the same asymmetry, but the clothing inherits it, so there's no problem?

Quote - You know, I had not thought to check this at zero pose (and I should have).  I see the same thing.  In fact I see it on all Poser characters I've looked at (5 or 6 different just now) 

I take this back, I can't verify any asymmetry in other figures except the EF ones (stupidly I trusted that the P6 and P7 EF figures were aligned correctly).  All the DAZ figures that I just checked (D3, M3, H3, M4, V4, SP3) do not show this alignment/symmetry offset.

actually v4 is slightly assymetrical in the mesh itself and v3 was more so. Very very minorly, but enough that while modeling in c4d using mirrored mesh I have to take it into account. DAZ thinks its an artifact of the scaling process needed to use poser mesh in other programs with larger worldspace

Lyrra

Do you mean asymmetrical in "shape", or asymmetrical in geometry/point order?  I remember that morphs do not mirror from left to right but I've never seen a difference in the shape from left to right that affects the way I model clothing, and I've done kind of a lot of precise bits all over the body, including gloves and a lot of torso bits.

Quote - > Quote - You know, I had not thought to check this at zero pose (and I should have).  I see the same thing.  In fact I see it on all Poser characters I've looked at (5 or 6 different just now) 

I take this back, I can't verify any asymmetry in other figures except the EF ones (stupidly I trusted that the P6 and P7 EF figures were aligned correctly).  All the DAZ figures that I just checked (D3, M3, H3, M4, V4, SP3) do not show this alignment/symmetry offset.

Ah, okay! That makes more sense.

I'm trying the reordering thing that Lyrra suggested. Interestingly, D|S displays the tree with the upper body first, but exports it with the lower body first.

Quote - Do you mean asymmetrical in "shape", or asymmetrical in geometry/point order?  I remember that morphs do not mirror from left to right but I've never seen a difference in the shape from left to right that affects the way I model clothing, and I've done kind of a lot of precise bits all over the body, including gloves and a lot of torso bits.

What I mean is, Antonia's mesh is perfectly symmetric, but the rigged figure is not, even in the zero pose.

And yes, I mean the shape. I assume by geometry you mean the mesh topology? I normally understand geometry to mean shape, so I don't see the difference there.

Actually I was responding to Lyrra, sorry for the confusion.  I understood what you were saying earlier.

Quote - Actually I was responding to Lyrra, sorry for the confusion.  I understood what you were saying earlier.

General confusion everywhere!  I'm pretty sure though that Lyrra meant shape as well. Rounding artifacts that can change mesh topology would be quite strange. :biggrin:

am I butting in on your conversation? thats me ..sowing confusion all over heh sorry?

pjz - its a very very small asymetry in the mesh, and so far only a few other modelers I've spoke to has seen it. The DAZ guys say its impossible with how v4 was made. But you load hger into c4d and lo and behold .. a slight shift L vs r. So they went and looked in various programs and it shows for a few. Very odd.  So I'm not losing my marbles (no really!). Its slight enough that generally its only an issue with super skin tight clothing which you dont want anyways cause the joints will be a *****

odf - heres an obnoxious excercise for you. export an object from poser and reimport the new one. Now repeat a few times, like say, ten. Then check it agaisnt the original.  You will find that the continued rounding off on import and export has melted the detail a little. Weird eh? EF told me over and over it wasn't so even when I showed them the results. SIlly humans.

Any rate I think the rigging thing might be partially caused by the cr2 load order. I'll ask Nerd next time I talk to him .. he might know.

Lyrra

edit: heh I see the directory structure was recorded wrong in the previous version anyway, it wouldn't have loaded

I changed the order of actors in the CR2, to no avail. I removed all the handles, disabled bulges and inner/outer spheres - no effect. But I've just had a look at the original mesh and the one Poser exports in Wings. The difference is really, really tiny. I understand that it bugs you, pjz99, but I can't imagine that it would matter that much in practice.

Well, I guess I've failed to convince you this is a problem.  I won't repeat it any more, but don't be surprised if other clothing riggers notice this.

Poser is completely, entirely, rational and explainable..

Except when it isn't

Some thngs you just have to shrug and move on

Lyrra

Here is a pic of what I mean.  The angles are listed in a different order, this does not matter, but what does matter is that each angles counterpart from left to right should have the same numerical setting only reversed + or-. Thry are not the same, so obviously they are not getting mirrored correctly.

Edit:  This is the shoulder zrot (up-down)

For the mc6, I want to use the vss shader set but I ran into problems.
When I use the vss prop , no matter what lighting I use I get seams everywhere.

If I apply SF's Toni tex and replace all her textures with mine, no seams.
This image shows the result with SF's set altered with my texture.

Now how would I change the skin tones here, what nodes would I change?

Dunno what nodes to play with, but damn that's some good work!

The only suggestion I would have (if they haven't been done already - no way to tell from the image) is to make sure the palms and soles of the feet are lighter in tone than the rest of the body, which is common in darker African/dark skin tones.

I still need to do a bit of tweaking on the arms so the lighter tone blends better.

This was created using a photo from 3d.sk of an Afro-American woman.
And yes, her feet were really that color. :)

Quote - I have found something in the cr2 which might explain the asymetry problem.  The numbers for the mat spheres are all the same but reversed as you would expect. But the numbers for the angles are not the same only reversed, the numbers are actually different from side to side. They should not be since the rigg has been mirrored.  The angle settings being off could definitely cause differences in the way the mesh deforms.

Thanks Mike! I've noticed this problem, too, and it'll be fixed in the next CR2. Unfortunately, restoring the symmetry of the angles didn't seem to change anything.

pjz99: I didn't mean to be dismissive. I realize that there's a problem, and I'm grateful to you for making me aware of it. I've spend many hours trying to get to the bottom of this, and it's not unlikely that I will spend many more. So to a casual observer, it might seem as if I was taking this somewhat seriously. All I'm saying is that if you look at the wireframes, you'll notice that the displacement is really tiny, so it might not be quite as bad as it looks at first.

BluEcho: That looks really beautiful. I'm tempted to drop everything else and make some morphs to go with that skin.

Beautiful!  Her feet are perfect, hands look like they need a little more work, but you know that.  =)  I'll probably try making Tony into an elf if I can find the time..  want to buy 100 hours in a day..

Quote - > Quote - I have found something in the cr2 which might explain the asymetry problem.  The numbers for the mat spheres are all the same but reversed as you would expect. But the numbers for the angles are not the same only reversed, the numbers are actually different from side to side. They should not be since the rigg has been mirrored.  The angle settings being off could definitely cause differences in the way the mesh deforms.

Thanks Mike! I've noticed this problem, too, and it'll be fixed in the next CR2. Unfortunately, restoring the symmetry of the angles didn't seem to change anything.

**Oh dang it  I Thought I was going to win the prize. I'll keep looking. :)
**
puts on glasses and goes back to looking at numbers

so I did a little more with one of the clothing items i made and tested it in poser this one i am calling fae. So far there appears to be no poke through. I used the wings from sadie fae. Now all I need is to get my hands on one of those skin textures for antonia. The only postwork was the addition of the background. Not really my best render but i just wanted to test the outfit out really.

http://renderosity.com/mod/gallery/index.php?image_id=1910509

p.s. that texture is looking very good blue

aella: Yay faeries!

You are not on the developers site, yet, are you? According to my made-up rules, you just earned yourself a pass, so send my your email address vie sitemail or to Antonia dot Polygon at gmail dot com, and I'll let you in.

Antonia's first, of course, and here's the explanation of what you see: each mesh is duplicated and flipped along the x-Axis. The original is shown in cyan, the flipped copy in magenta. If everything were perfectly symmetric, we would only see cyan, because the original is rendered first, and the magenta from the flipped copy shows up only if it gets to lie in front of the original.

The image at the top is from the original mesh. That's the one Poser loads each time it reads the CR2, and as you can see, it's darn near perfectly symmetric. The image at the bottom is what comes out of Poser when the figure is loaded and then exported as an OBJ file in the zero pose. Not quite so symmetric anymore.

So, the conclusion so far is this: Poser always creates asymmetries in our meshes, but the degree varies between figures and between body parts. Your mission, if you choose to accept it, is to find out what makes V2 so special.

I still think that in the grand scheme of things, the asymmetry that Poser produces in Antonia is so small that it will only ever matter for really, really tight clothes. But one has to be weary of cumulative effects as Lyrra mentioned. Also, if this effect can be reduced, it can also be amplified, and since we don't know how it happens, we can't be certain that the next edit on the rig will make it really, really bad. So I'd really like to solve this.

By the way, I'll still have to check if the asymmetry gets worse or stays pretty much the same when the figure is posed. I'll use V2 as my benchmark on this.

Quote -
For the mc6, I want to use the vss shader set but I ran into problems.
When I use the vss prop , no matter what lighting I use I get seams everywhere.

Is the VSS changing the textures in the materials room to Filtering Quality or Fast instead of None?

The filtering is set to Quality.

Turn texture filtering off for no seams - it's only good on transmapped hairs anyway:
http://digitalbabes.jp/dload/texturefiltering.html
For other textures it is leading to seams.

Hello there,

I am no Poser expert but I think the differences that ODF showed between the original object file and the file that Poser exports can be explained when you open both obj-files in a text editor.

V4:
Original:
v  0.00149163  0.65661299  0.00175520
re-exported:
v 0.00149163 0.656613 0.0017552
-> poser limits the second column to 6 digits

Antonia:
Original:
v 7.3631111e-3 0.65570511 2.3961111e-2
re-exported:
v 0.00736311 0.655705 0.0239611
(slightly changed a vertex)
-> poser limits the second column to 6 digits

V2:
If you throw a look in the V2-file you see this:
Original:
v -0.000692 0.644490 0.036993
re-exported:
v -0.000692 0.64449 0.036993
-> V2 only has 6 digits

I don't know if that is related to the slightly different deformations, but I think that it explains
at least the differences ODF spotted between orginal and re-exported file.

Not a bad idea. But the differences I see are in the fourth digit, so unfortunately they can't be explained by round-off errors. I don't think Poser would round the numbers on import, anyway, only on export. So this wouldn't explain the difference we see in Poser itself.

At any rate, it probably won't hurt to make Antonia's OBJ file use some consistent fixed-point format.
From what you posted, it looks like Poser uses 8 digits, but trims off trailing zeros.

Hi - only have a second as I'm going sailing for the day.

VSS has texture filtering None/Off. If it is showing as Quality for you, then somebody else did it.

I believe it should be None.

I distribute the VSS skin shader with it set to None.

**Thank you Kerya, thank you Bagginsbill.

I did read about that and forgot, guess my brain is getting too slow these days. Lol.

It isn't the vss that's causing the prob because when I start fresh with everything new it works ok.
Must be the pz3 file became corrupted somehow...sigh.
**

Here is another image using vss and my changes to it on Antonia.

Last image for a bit, got an infection in my mouth and may have to go see the doctor...ugh.

I can not eat solid food because my mouth is so sore and has blisters in it.

So I will take a few days off and rest.

Poor BluEcho! If you want to correct your pz3:
SVDL has a free python script to do that for all items in a scene.
http://www.renderosity.com/mod/freestuff/index.php?user_id=99327

Kerya,

BluEcho is using VSS which explicitly creates all the nodes in the figure shader with known values. Sometimes a corrupted scene makes Poser ignore the values - it's a bug in Poser.

I've seen this before, both with regard to the value of Texture Filtering, as well as the value of the image gamma in Poser Pro. You can set these manually, with a script, by hacking the file - doesn't matter. Once Poser gets them wrong, they stay wrong.

That's what BluEcho meant about it works starting with a fresh scene.

Speaking of corrupt: I've butchered Antonia's mesh and rig until I had only 10 vertices, 12 faces and 3 bodyparts left. Poser still manages to turn a perfectly symmetric mesh into an asymmetric figure.

There's not much more I can rip out of the CR2 except joint and transform channels, and I'm not sure that's allowed. And I still can't see where Poser gets those displacements from. Fascinating!

You're basing a lot of work on how accurate the OBJ import/export really is and that may have absolutely no impact on how the joints will deform in Poser.  While it is certainly bad that Poser exports an asymmetric mesh, it may be a separate bug entirely.

Quote - Kerya,

BluEcho is using VSS which explicitly creates all the nodes in the figure shader with known values. Sometimes a corrupted scene makes Poser ignore the values - it's a bug in Poser.

I've seen this before, both with regard to the value of Texture Filtering, as well as the value of the image gamma in Poser Pro. You can set these manually, with a script, by hacking the file - doesn't matter. Once Poser gets them wrong, they stay wrong.

That's what BluEcho meant about it works starting with a fresh scene.

I didn't know that. Opening the scene, correcting things and saving under a new name doesn't help?

Quote - You're basing a lot of work on how accurate the OBJ import/export really is and that may have absolutely no impact on how the joints will deform in Poser.  While it is certainly bad that Poser exports an asymmetric mesh, it may be a separate bug entirely.

Nope, I can see the asymmetry in Poser, as well.

It may not do any good in the short run but you probably should open a ticket with Smith Micro support about this to help motivate them to fix it.

I'm not worried about saving that pz3, it was just for testing the texture anyhow.
Starting fresh is okay but it meant having to reload the figure and all the texture stuff since I hadn't created the mat file yet.

Now I have it all sorted out and mat files created.
Just need to finish my material room settings so I can make mc6 files. :)

vss gives the texture a much better look nad I'm learning to adjust the settings too.