I made a small Poser model. It includes a segment "ray", whose parent is the root segment. In its .CR2 file I accidentally omitted the Weld statement welding "ray" to its parent. When I tried to give "ray" a morph, my Poser 4.0.3 (without PPP) always said that the morph target had the wring number of vertexes, although the number of vertexes was correct. The trouble went away when I went into the model's .CR2 file with a text editor and inserted the missing Weld instruction. Ref. also my previous messages about severe rendering faults happening in Poser when the user follows the Renderosity member Bloodsong's advice to insert extra Weld statements when parts need to touch when one is not parent of the other. It seems that the process of welding shares facilities with other Poser functions, perhaps to save space. It would be useful if the welding process could be made completely independent of other processes, and if the welding process was not relied on to produce values or information that other processes need.

When you 'weld' items some of the verts become shared, and those shared verts are allocated to one of the groups (the parent I think) If it didn't do this, then for things like full-body-morphs, you would get the verts of seams between groups being morphed twice, which makes an unseemly mess :) Haven't encountered the rendering faults you mention for welding body parts to more than one other part. This was standard practise in older Poser models. If you look at some of the cr2's for version 2 models for example, you will see that the collars are welded to the chest, neck, and each other. Since the designers did this them selves, you would think they would have made it work :) Maybe they had problems themselves though, which is why the newer models don't do such things.

When you 'weld' items some of the verts become shared, and those shared verts are allocated to one of the groups (the parent I think) If it didn't do this, then for things like full-body-morphs, you would get the verts of seams between groups being morphed twice, which makes an unseemly mess. Not if those points are morphed before they are welded.

Can I throw in a slightly off-topic question here? I'm building a lot of furniture in cr2 form and removing all the weld statements, along with the jointx/y/z lines from the cr2s. Can this cause any kind of problem? I was led to believe that in a figure which doesn't use the Bend option, these lines were all superfluous. Is this the case? mac

i suspect that, when Poser runs, obeying Weld has side-effects which are relied on by the process of rendering and the process of adding a morph target. So, omitting Weld lines or adding extra Weld lines messes things up.

along with the jointx/y/z lines The channels to remove are joint... and twist... and smoothScale....

Sorry, Anthony. I actually meant joint/twist/smooth. To be more specific, SmoothScaleX, JointZ, JointY and TwistX. Also Taper, which is no use to my figures whatsoever. So you reckon removing the Weld lines causes problems? I must admit, I've never found any. I must double-check this, Thanks mac

Whether SmoothScale and Twist have X or Y or Z, depends on whether that part's line in the .PHI file had xyz or xzy or yxz or yzx or zyx or zxy.

From Anthony Hernandez's notes on cr2s. "This (weld) parameter makes a smooth transition between adjacent body parts in conjunction with the Bend option" I'm assuming that if Bend is off, Weld doesn't apply. Every one of my figures has xyz in the PHI file. Since these channels and the Weld commands control the smoothing between seams in poser, if there are no seams to smooth (as in furniture), I don't see that they have any relevance. Certainly, I know a few people who habitually delete them and have no problems. mac

I'm assuming that if Bend is off, Weld doesn't apply. I found that if one part is parent of another, and one has bend off and one has bend on, on the part which is bend off, its vertexes at the join go halfway to the matching points on the non-bending part. LIkeliest Poser does all the Weld calculations for all parts; but then for non-bend parts it does not use the resulting values to move the edge points.

Thanks for the replies, Anthony. If that's the case, I'm ok, since EVERY part of my figures has Bend off. I've double-checked my stuff and there really seem to be no probs. mac