i finished modelinging my character and now i don't know if i should rig it first or texture it first...it seems like it is usually rigged before textured, am i right?
You could go either way. An advantage doing the rigging first is that you can modify the geometry to improve deformation if need be, without swearing like a sailor about how you just messed up your uvs and texture maps.
I rig at the end. I always modify the mesh when I texture, like 99,99999% of the time. And working under a physique or skin, is well hell. so nokthanks.
I rig at the end. I always modify the mesh when I texture, like 99,99999% of the time. And working under a physique or skin, is well hell. so nokthanks.
[/ QUOTE ]
Same. Though I keep Uvs in mind when I model, sometimes something will crop up that needs changing. Often I'll discover when texturing that something could have been mirrored, while other parts shouldn't have been, etc. As he said, working under physique is hell.
you can always create a data mesh and alter the geometry there and then bake the weights back on, or just transfer the weights over to the altered mesh, then tweak what weighting needs to be tweaked.
The way I work is:
Create said model.
Hand off model to animator without uvs.
Create uvs and texture while animator rigs
Give animator model with final textures and uvs.
Animator transfers weights to final model.
It's good to know if there is any changes to be done to your mesh before starting the uvs or at the earliest point possible.
I rig at the end. I always modify the mesh when I texture, like 99,99999% of the time. And working under a physique or skin, is well hell. so nokthanks.
Verm, which program are you using? Max is a bitch as people are pointing out. I cant tell how manytime I have saved out skin points only for it to not work even on the same unmodified model. Many times, you still have to almost redo most of the rigging. If we are doing it wrong, then can you list the methodology in max?
Also, what about mirrored portions of your model? If you rig then go back, twice the UVW work unless you go to an early version where it was still in symmetry mode. However, how many of us haven't gone back and tweaked and reduced poly counts more after symmetry has been collapsed?
Yarrr suffer and scorn, make thee Max walk the plank! If yee keep the skin modifier on top of the model the whole time all changes propogate up through the model, no need thar be to delete skin or save out points matey.
As far as symmetry, you can detach parts of the model reattach on said symmetrical side and skin will update when yee go back up the stack matey. Twas a tip from Poopbeard from long ago.
Actually I confused myself (easy enough). It was when I was trying to use the symmetry function in skin that it failed,and I still ended up having to weight each side individually.
So your saying that I can skin the model, then go back and detach say the lower legs which are symmetrical. Delete one, uvw the remaining, and then mirror it across. Reattach both and they will both still have the right skin weighting??! If so I didn't realize skin was this advanced.
Give me a list of your stack.
Mine usually is
Skin
MultiRes
UVW
Poly
Oh, and I encounter problems sometimes with skin confusing points when I reduce multires to about 75%. So I still have to go back up and manually adjust it. ( I use multires to help create LOD).
I've never rigged a model before texturing it. I adjust the geo and UVs after rigging if I have to. I've never had to adjust to such an extreme degree that I ended up with texture stretching.
rigging only after all texturing, modeling plarps are done, if i want to go back after rigging its pain so i basically am sure if i really want to rig or not.
[ QUOTE ]
Verm, which program are you using? Max is a bitch as people are pointing out. I cant tell how manytime I have saved out skin points only for it to not work even on the same unmodified model. Many times, you still have to almost redo most of the rigging. If we are doing it wrong, then can you list the methodology in max?
Also, what about mirrored portions of your model? If you rig then go back, twice the UVW work unless you go to an early version where it was still in symmetry mode. However, how many of us haven't gone back and tweaked and reduced poly counts more after symmetry has been collapsed?
[/ QUOTE ]
I'm a Max user, but my workflow might be rather unusual. I really don't ever use symmetry, as my model work always seems to be asymmetric (hands and faces being the exception) and I don't trust the damn thing. I do mirrored instances as I work and then weld everything together before I rig (yeah, not efficient, but I don't trust mirrored weighting much either).
I build the entire model, add an unwrap modifier, rig it, test it through a full range of animations and then clone the model. I hide everything but the clone and map it, save the .uvw data, hide the clone, get the rigged model. With that selected, I drop down to my unwrap modifier and load the data I saved off the clone. I've never had a problem loading the map data into the original model since the modifier is already in place. I guess it's like I start the UV process by assigning the modifier, but I don't touch it again until the rigging is done. I never actually work under Physique, since I don't apply it to my clone model, and moving map data is much more reliable than moving rig data. Hard to explain, I guess - did that make sense?
Anyway, I burned myself a lot in the good ol' days of <1000 polycount characters by building joints without a lot of accomodation and have trashed plenty of UV maps as a result once I got to animating them. Since I resovled to do the rigging first, it is never a problem.
[ QUOTE ]
Actually I confused myself (easy enough). It was when I was trying to use the symmetry function in skin that it failed,and I still ended up having to weight each side individually.
[/ QUOTE ]
This is why I don't use symmetry - she's a fickle whore who creates more work for me in the long run.
My workflow for characters is close to Verm's with a few differances. I do use symetry but only early on.
- Model
- Unwrap with temp texture
- Temp rig select areas, attempt to fix stretching, re-unwrap select areas
- Final rig
- Pass on to 2D and start animating
So to answer the question, yes. I do them at almost the same time. Unwrap an arm do a test rig or final rig check it and if it looks good move on if not fix it, and then move on.
Since I discovered pelt mapping I am not as nearly hesitent as I used to be about tweaking the mesh for better animation after I have unwraped the model the first time. As long as I haven't passed it off yet I'll unwrap and tweak a model a few times. It only takes a few hours to unwrap most of the models I make and most of the time I only have to reunwrap one or two areas. The models I work on are 3,000-4,000 polys.
Vig's comment about remapping just a few areas reminds me of a good point. Since I am teh old skool, I think of remapping issues being slightly catastrophic - having to alter entire limbs if the joints don't deform well, etc. The models I build these days have ten times as many polygons as my original work, and I can generally tweak things instead of rebuilding (adjusting a knee joint without having to modify/optimize the entire leg).
That probably makes rigging after texturing less of an issue, since the fixes aren't usually as major. I still would prefer to rig first - it's more satisfying for me to see a model moving than one painted, I guess.
Replies
thanks.
I rig at the end. I always modify the mesh when I texture, like 99,99999% of the time. And working under a physique or skin, is well hell. so nokthanks.
no thanks.
I rig at the end. I always modify the mesh when I texture, like 99,99999% of the time. And working under a physique or skin, is well hell. so nokthanks.
[/ QUOTE ]
Same. Though I keep Uvs in mind when I model, sometimes something will crop up that needs changing. Often I'll discover when texturing that something could have been mirrored, while other parts shouldn't have been, etc. As he said, working under physique is hell.
The way I work is:
Create said model.
Hand off model to animator without uvs.
Create uvs and texture while animator rigs
Give animator model with final textures and uvs.
Animator transfers weights to final model.
It's good to know if there is any changes to be done to your mesh before starting the uvs or at the earliest point possible.
no thanks.
I rig at the end. I always modify the mesh when I texture, like 99,99999% of the time. And working under a physique or skin, is well hell. so nokthanks.
[/ QUOTE ]
Clearly you are doing it wrong.
FACT.
CLEARLY!
All thoses year. I have been a secret noob.
Also, what about mirrored portions of your model? If you rig then go back, twice the UVW work unless you go to an early version where it was still in symmetry mode. However, how many of us haven't gone back and tweaked and reduced poly counts more after symmetry has been collapsed?
As far as symmetry, you can detach parts of the model reattach on said symmetrical side and skin will update when yee go back up the stack matey. Twas a tip from Poopbeard from long ago.
YaaaRRRR!
So your saying that I can skin the model, then go back and detach say the lower legs which are symmetrical. Delete one, uvw the remaining, and then mirror it across. Reattach both and they will both still have the right skin weighting??! If so I didn't realize skin was this advanced.
Give me a list of your stack.
Mine usually is
Skin
MultiRes
UVW
Poly
Oh, and I encounter problems sometimes with skin confusing points when I reduce multires to about 75%. So I still have to go back up and manually adjust it. ( I use multires to help create LOD).
Yaarrr!
Verm, which program are you using? Max is a bitch as people are pointing out. I cant tell how manytime I have saved out skin points only for it to not work even on the same unmodified model. Many times, you still have to almost redo most of the rigging. If we are doing it wrong, then can you list the methodology in max?
Also, what about mirrored portions of your model? If you rig then go back, twice the UVW work unless you go to an early version where it was still in symmetry mode. However, how many of us haven't gone back and tweaked and reduced poly counts more after symmetry has been collapsed?
[/ QUOTE ]
I'm a Max user, but my workflow might be rather unusual. I really don't ever use symmetry, as my model work always seems to be asymmetric (hands and faces being the exception) and I don't trust the damn thing. I do mirrored instances as I work and then weld everything together before I rig (yeah, not efficient, but I don't trust mirrored weighting much either).
I build the entire model, add an unwrap modifier, rig it, test it through a full range of animations and then clone the model. I hide everything but the clone and map it, save the .uvw data, hide the clone, get the rigged model. With that selected, I drop down to my unwrap modifier and load the data I saved off the clone. I've never had a problem loading the map data into the original model since the modifier is already in place. I guess it's like I start the UV process by assigning the modifier, but I don't touch it again until the rigging is done. I never actually work under Physique, since I don't apply it to my clone model, and moving map data is much more reliable than moving rig data. Hard to explain, I guess - did that make sense?
Anyway, I burned myself a lot in the good ol' days of <1000 polycount characters by building joints without a lot of accomodation and have trashed plenty of UV maps as a result once I got to animating them. Since I resovled to do the rigging first, it is never a problem.
Actually I confused myself (easy enough). It was when I was trying to use the symmetry function in skin that it failed,and I still ended up having to weight each side individually.
[/ QUOTE ]
This is why I don't use symmetry - she's a fickle whore who creates more work for me in the long run.
- Model
- Unwrap with temp texture
- Temp rig select areas, attempt to fix stretching, re-unwrap select areas
- Final rig
- Pass on to 2D and start animating
So to answer the question, yes. I do them at almost the same time. Unwrap an arm do a test rig or final rig check it and if it looks good move on if not fix it, and then move on.
Since I discovered pelt mapping I am not as nearly hesitent as I used to be about tweaking the mesh for better animation after I have unwraped the model the first time. As long as I haven't passed it off yet I'll unwrap and tweak a model a few times. It only takes a few hours to unwrap most of the models I make and most of the time I only have to reunwrap one or two areas. The models I work on are 3,000-4,000 polys.
That probably makes rigging after texturing less of an issue, since the fixes aren't usually as major. I still would prefer to rig first - it's more satisfying for me to see a model moving than one painted, I guess.
my framerate drops when I use symmetry, so I never got used to it in my pipeline.