Jump to content
Sign in to follow this  
Fox '09

more weighting issues , soldiers

Recommended Posts

Hello, stalkergb (hehe) , I have another weighting problem. Sort of.. I have a feeling i can do this a few ways, but was wondering what is the best way to solve this problem..

g75DS.png

g7fCr.png

I have thought about:

re-weigh the points that don't match up to the vest the same way as the vest is weighted ...

merge points ...

re-weigh and merge points

cheers,

fox

Edited by Fox '09

Share this post


Link to post
Share on other sites

I would think that the best way of doing that would be to merge the points together first. Generally its easiest to have the model as "finished" as it can be before you start too much of the weighting. Otherwise you end up making more work for yourself (well sometimes).

To be honest, if you weighted the unmerged points the same as the vest there would probably still be a gap. So I'd definately merge straight away.

Also, I believe that merging will mean the points inherit an averaged weight, so as an example, if the vest has 0% weight on LeftArm and the other point had 100%, when you merge them, the new, single point, would have 50% weight on LeftArm. So infact you could probably get away with just merging. Although obviously you can weight it if it doesnt have the desired result.

Hope that helps mate. (and makes sense)

Share this post


Link to post
Share on other sites

ok , I'll give merging a try. On that note, my straps are more complex than the body itself, there are 2 polys for every one .. should I merge the closest or all ?

Share this post


Link to post
Share on other sites

Hmm, in that case, it may be worth creating a new face rather than merging, (so take 2 points on the strap and 1 on the body). You'll need to quickly UV map the new faces but it would only be a small job I guess.

As you can see in my epic drawing, the top red line is the vest, the bottom the body, the black dots are points and the green bits show where new faces would go. Basically whatever would fit best but I thought I'd do it just as an example.

points.jpg

Merging would probably yield some weird results but you could still have a quick try, would mess up UV's too I suppose. If you did add the faces, you'd then need to weight the points on the body to be similar to the ones on the vest (as you planned to do before merging earlier). The gap would be gones because of the new faces though.

Share this post


Link to post
Share on other sites

i could give that a try. My model is very closely fitted to the body , so i dont think that will work as well, or at least as efficiently . thanks once again

Share this post


Link to post
Share on other sites

EunhS.png

EupN0.png

i was thinking of bringing the shirt verts to match the vest ones, which would probably reduce stretching... what do you think?

Share this post


Link to post
Share on other sites

I find that verts that aren't attached to anything seem to move more so it may well reduce stretching/movement for you. Generally I try and keep everything connected and closed so I'd say its worth a go. Make a back up just in case though.

Share this post


Link to post
Share on other sites

It just looks like your weight transitions are too soft. You can try and weight everything as boldly as possible, where a vertex is only weighted to one bone, and then get the best results out of that, and then soften the boarders ever so slightly and see what happens.

Share this post


Link to post
Share on other sites

it's me again, and this time i'm confused on whats wrong. The bottom of this body is distorted in game, but only the vest. everything below it is OK.

http://filesmelt.com/dl/foxus_sog23.p3d (if you want perms just ask, but that's not the point in this anyway)

any ideas?

cheers,

fox

Edited by Fox '09

Share this post


Link to post
Share on other sites
it's me again' date=' and this time i'm confused on whats wrong. The bottom of this body is distorted in game, but only the vest. everything below it is OK.

[url']http://filesmelt.com/dl/foxus_sog23.p3d[/url] (if you want perms just ask, but that's not the point in this anyway)

any ideas?

cheers,

fox

Have you got a picture of the problem in game?

EDIT: Also did you still need that SOG guys weighting (the t-shirt on him)? Might have a bit of time next week if you do.

Edited by STALKERGB

Share this post


Link to post
Share on other sites

Pic:

arma2oa_2010_12_11_00_01_32_28.jpg

arma 2 is not starting right now, thats an older pic. I managed to fix it up to the bottom of the vest, where it still stretches. Around the Spine selection.

The way I fixed it is reassigning the vest to the appropriate selections, as before i also selected the t-shirt behind the vest, which assigned weighting only to the t shirt and not the vest..

As for the shirt, that would be great, I'm still not sure how to fix that one either.

edit:

got arma2 to start, here's the most recent pic:

arma2oa_2010_12_11_12_12_17_22.png

arma2oa_2010_12_11_12_12_33_26.png

Edited by Fox '09

Share this post


Link to post
Share on other sites

Stretching is from verts that are probably not part of 'any' animated selection and are left behind in the 'default' bind pose positions when the anims. are applied.

But this should be evident when you apply any standard .rtm to your model when testing in O2 prior to packaging.

As appears in pic 1 & 3.

In pic 2... that still could be a tiny little bit of a vert or 2 not being part of a selection. Or, possibly weighting or a little misplaced uv'ing. Or a combo. of all of those. Kinda hard to tell from here.

Ah... sorry... just saw the .p3d.

These verts...

th_foxus_sog23.png

should probably be allocated some weight to Pelvis, Spine & Spine1 at the very least... and probably the upper legs should influence them as well... currently they are not part of 'any' selection let alone having any weighting. Btw, that's the 'Pelvis' weight map being displayed above.

There are also other verts on the back of the vest that are not allocated any weight as well so are therefore not becoming part of the 'selections' in the .p3d and are therefore not animating when .rtm'd by either O2 or in-game.

What you should always do is once over in O2 apply a .rtm to the model to see if there are any glaring anomalies.

So, just to clarify the process you would go through here would be something like...

Open your current .p3d in O2.

Apply a small .rtm to it.

Move to frame 0.0.

'See' if there are any major looking deformaties.

If there are then save the .p3d. This will keep the applied .rtm in the .p3d as a 'point cache' animation chunk.

Start modo and in the plugin toggle that 'On Import Create MDD' in the P3D Assistant.

Then open this 'new' .p3d. The plugin will dump the included anim. data out to a .mdd file for that LoD during loading.

Apply this .mdd to the mesh in modo.

You should now be able to readily identify which 'verts' are not being moved and apply some appropriate weight to each for all weight map's that should logically influence them.

Cheers, Sy.

Edited by Synide

Share this post


Link to post
Share on other sites

excellent explanation, thanks. Where can i find sample rtm's or something? I couldn't find any the other day.

Edited by Fox '09

Share this post


Link to post
Share on other sites

You can use something like the ICP anims and extract the PBO to test on your model, I believe the ARMA 2 ones are binarised in a way which means they can;t be opened in O2

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
Sign in to follow this  

×