I was going to try and create some morphs for Sarah but when I export her (and with the original geometry) I find that her body has been slightly (around z axis) very slightly twisted around the hip area, which has made her unsymmetrical about the z-y plane... which means MTMirror sees her as unsymmetric (which Im pretty sure she isnt) :( which is a bit of a pain. Did my .obj file get mangled somehow or is this the same for everyone ? Sorry if this sounds like a gripe.. because it isn't, Im just enquiring if this is just me and if its not, then is there a way to rectify Sarah ( a great model ) so I can pass her through MTMirror ? Best Wishes, Barrie.

Yeah, I saw that too. Neftoon looks even worse for MTs mirroring.

wonder if its a purposeful thing to stop anyone but the owner of the original mesh having an easy time ?

Un-symmetric? Wow... just like real people...grin... - Jack

Why do you export her ? When you're modelling or morhing, copy the obj. file - the exported file will have been posed ever so slightly. See if that helps :-) ~S

Sorry, I did mention (ever so softly :) ) I did try the original geometry (.obj file) , and its the same. "Real" people un-symmetric ? no way. :)

oops, forgot to mention (well I just remembered actually) the reason why I had to use Export, was becasue COMPOSE (excelent program as it is) wouldnt read the original.. Best wishes.

Interesting. I'm working on something for her in Max right now, and I don't have that problem. I've noticed that sometimes, when copyright holder data is inserted in the .obj file, that some apps have problems reading them, btw And yes - you said original object file, and I'm sorry - I didnt see that. I couldn't read the original in Compose either, so I took the original object file, opened it in UVMapper, saved it to a different location, and now Compose can open it. ~S

Thanx, uv-mappers a good idea , just need something to filter out bad syntax (or new syntax: Copyright Holder ?) and rewrite a standard obj file.. I can work on her in MAX, ok, i can compose ok... its just when I goto MTMirror, that i find out shes not as symmetric as she should be, try for yourself, when you zoom in like i have in the FRONT view, is the central line of her head ... central ? ... if so, it may be an interesting side effect of uvmapper that its rounding down the errors or something.. maybe. Barrie

I just looked at the 3dusara2.obj in wordpad ... usual layout for a .obj file is V ertex coords, Vt cooords... then f aces .. that reference the V's and Vt's by index .. Sara (is it sara ? I thought it was sarah) does this wierd thing: vt 0.158630 0.176237 vt 0.165828 0.168407 vt 0.152074 0.154748 vt 0.143574 0.164016 f 33804/-1 33774/-2 33719/-3 33762/-4 it uses negatives to reference Vt coords.. I think thats why compose cant read it.. its quite an old program and prob doesnt recognize it. Barrie.

I imported the original obj into max via Habware's plugin, didnt go near anything else ;-) This is probably the culprit - this is at the beginning of all the sarafiles, with a variance: #### # # OBJ File Generated by LightWave3D # LightWave3D OBJ Export v2.2 # #### # Object: sympseudofix4.lwo # # Vertices: 35139 # Points: 0 # Lines: 0 # Faces: 33124 # Materials: 27 # #### UV Mapper removes that, and inserts: # file generated by UVMapper # NumVerts/NumTVerts/NumVNormals/NumFacets 35139/35170/0/33124 # NumGroups/NumMaterials/NumRegions 72/27/0 # x/y/color/ppu 512/512/0/50.00000000 instead. I'll get back to you on the linechecking - I'm in the middle of a render :-) ~S

#'s are just comment lines, I doubt they would cause any problems, Im more inclined to go with my guess, but I could be wrong :). Thanx for checking the symmetry but dont worry about it too much I'm resigned to the fact that God doesnt want me messing with Sara / Sarah .. whatever .. lol Barrie.

Reason I'm going for the cpmment files, is that I had a beast of a time earlier this year trying to import a commented object file into the Deep UV demo. I couldnt for the life of me figure out what it was, I looked at the verts, I renamed, I tried this, I tried that... Until I in a fairly last ditch attempt removed all the comments at the top of the file - and then it worked imported just fine. Wasn't even a Lightwave generated object. And it usually does the trick in stuff like Compose too. I have no clue why - just know it works :-) ~S

Then we've probably had different probs with the objects - and know there is a workaround :-) ~S

Hi,

all my (freebie) software rejects the .obj too. However I don't think that the age of the interpreting software has anything to do with it!

Obviously you can't have a negative index into an array of anything, or can you? According to my (ancient) obj reference, wellll, maybe! Have to say that a quick analysis of other .obj files on my HD shows this is the only file I have that actually does this, but;

f 33804/-1 33774/-2 33719/-3 33762/-4 means:

face at absolute index into array of vertices (33804,33774,33719,33762) maps to the texture vertices at - RELATIVE index, in terms of lines into the file of - (-1,-2,-3,-4)

In simpler terms: the "/-" parts of the (f)ace chunk reference the preceding 4 lines, in inverse order.

John

Yeah I agree.. it just references going backwards. Maybe its just a little used way (rather than new) of storing the info in the .obj file, and because its so little used not many apps have made allowances for it. Barrie.

Sorry, Im always missreading things :( ... you said that -1 , -2 .. refer to previous lines in the file, I'd personally take a guess at them refering to the last vt entry.. I mean -2 means the second from last vt entry(rather than file line) , -n being the nth from last vt etc... maybe you can apply negatives to the v coords as well, although ( I hope not :( ). Barrie

The reference I have states: "Each element references the vertices stored immediately above it in the file." As the only file I've with minus indeces in it is this one, and the preceeding 4 lines are vts, don't know what else I can infer from that! The file I'm reading "obj_spec.pdf" doesn't seem to be copyrighted in any way, so I can mail it to you if you want. It probably came from Wotsit.org, that's where I get most such stuff, it might be worth searching there for something more up to date. John PS of course the order doesn't have to be inversed, it would depend on the face normal.

Thanx for the info :) I think I have the spec somewhere as well I was just too lazy to find it, It seems like UV mapper will reindex the faces though according to Spinner, so its not a big deal really. Its this symmetry thats really bugging me. Barrie

I've noticed some symmetry problems as well with Neftoon Gal, and others.. I'd think the authors would want to encourage as much developement as possible for their figures!

anyone know of a lightwave script to correct symmetry?

Attached Link: http://www.m2estudios.com/plugins.html

I don;t think I wanna select 500 points and try to fix them, but there other VERY cool scripts at that link! Thanks!