Xtaz Posted December 18, 2004 Posted December 18, 2004 Hi ... I modeled a new character. I don't want to model a low resolution version ... Can i use same model to use as lo-res ?? if I did it, which do the parameters need to be changed ? Peace Xtaz Quote
John Bigboote Posted December 18, 2004 Posted December 18, 2004 THIS is a good question and I'm not offering an answer rather 'chiming-in' on the question. When I was fiddlin with WM I had my char modelled and then did NOT want to go to the trouble of REmodelling a low-patch version of it so I just grabbed 'Thom' off the hash CD and my experiment was 'less than successful'... What is a recommended course of action to use as a proper 'low-res' proxy? Thanks Xtaz, hope I din't step on your toes! Quote
Frank Silas Posted December 19, 2004 Posted December 19, 2004 This is pasted from an email originally sent to the WieghtMover Email list on Wed August 27, 2003 8:20AM. The answer is yes you can do it. What it does is give your model a uniform smoothness so that it bends well and predictably: Using a copy of the high res version as the lo-res model for WeightMover. (pasted email follows) 1. I did not want to have WeightMover affect the head of my model since I have pose/relationships there for phonemes and facial expressions. (I tried that it totally wanks out that stuff if applied before running through WeightMover. I am really excited to run a model through WeightMover before creating the facial expressions and phonemes! I will try that next.) 2. So then I made a copy of the model from the neck down. Named the copy hi-res (since it was not attached to any bones) then named the original mesh (which is boned) from the neck down lo-res. Created my exclusion groups. Then ran through WeightMover. 3. The trickey part was connecting the lo-res version to the neck under the head. Originally I was getting a lot of wrinkling in the shoulder area after deleting the lo-res torso and connecting the hi-res torso to the head of the character. To solve this I made sure that all of the verticies of the neck where part of the lo-res and hi-res groups before running through WeightMover. To connect the lo-res model to the head I "cut" the top ring from the neck on the lo-res group and connected all splines on lo-res to matching splines on the bottom of the head model. There were some dangling splines so I connected them to the head using patch hooks. All of the splines I connected where from the lo-res group to the head. (Don't create any new splines and don't connected any splines from the head to the lo-res group....don't select a vertici from the head group and use the ~ key to connect it to the lo-res group. Only select a vertici from the lo-res group and use the ~ key to connect it to the head.) This is important because it would not work for me if I did it the other way. I would still get wrinkling in the shoulder area. 4. Now when I animate the character he has all of his facial expressions and full set of phonemes and his torso from the top of his neck down has been run through WeightMover and bends incredibly well. WeightMover is awesome. I won't be animating characters in A:M without it. Frank Silas 3D Artist Rockstar Games SD http://www.franksilas.com Quote
luckbat Posted December 19, 2004 Posted December 19, 2004 Psst-- Frank, the singular of "vertices" is "vertex." Quote
Hash Fellow robcat2075 Posted December 19, 2004 Hash Fellow Posted December 19, 2004 Ok, I don't understand the question. I thought the whole point of WM was to transfer weighting and smartskinning done on a lo-res proxy (because that's easier) to a high-res version. If you use the High-res version as your lo-res proxy... how does that simplify anything? If you're going to weight a high-res proxy ... why not just weight the original high-res model? And as far as making a lo-res proxy... how about just deleting spline rings from a copy of the original until you have the bare amount to define the shape of the character? Quote
anzovin Posted December 19, 2004 Posted December 19, 2004 Robcat is correct--if you use your hi-res model as the proxy, it works, but you do lose the advantage of being able to manipulate a lo-res version instead of a hi-res one. For most character models, taking the original hi-res model and deleting spline rings, etc., to create a lo-res version is indeed the way to go. It might take only a few minutes to create the necessary proxy, depending on the complexity of your character. And then you will also have a proxy to use for animation. Steve Anzovin Studio Quote
Morgan Posted December 19, 2004 Posted December 19, 2004 Robert, That's definitely the best way to make a proxy, in most cases. That said, the advantage of using an exact duplicate of your hi-res group as your "lo-res" group is smoothing. In addition to the initial weight transfer, which is based entirely on spatial positioning of individual points, WeightMover applies a secondary weight smoothing along the splines of the hi-res model. For models with moderate spline-density, this can actually be a fast and easy way to get your joints bending nicely. John, I second Robert's advice of taking your pre-existing model and deleting some of the spline rings. (Be sure to leave at least one per bone, obviously.) This will give you a better match-up for weight transfer, and it's usually faster than modifying an existing lo-res model like Thom. But if you still can't get the results you're looking for, feel free to e-mail me with screenshots or model files. Xtaz, The only parameter that will affect a one-to-one transfer like that is "Final Smoothing," and the default setting (1.0) should be fine. Quote
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.