start with a cylinder with as many sides as the banana has do some loops and scale into a straight banana shape detach and deform the lower parts to match reference ffd modifier to curve the upper part
I'm curious how would you guys pull off modeling a cartoony character that looks like this?
I've done some more realistic character modeling in the past but nothing heavily stylized like this. Capturing the expressions of the mouth and eyes is honestly stumping me.
This is what I've done so far:
I've been having a lot of trouble with the mouth. I thought I'd just have it as a black line like in my picture above and move the cheeks with soft selection to give it expression but it doesn't look great at all. Link of it here: https://i.imgur.com/tJXPQDl.gif You also can't have the mouth open the way I'm doing it right now. Should I actually model the lips? Would opting for animating textures give me more the look I'm going for? Same questions also apply to the eyes as well I guess. Any insight would be appreciated!
I got myself in a situation where I merge different meshes to one mesh so I can UV map them. After flattening, relaxing and packing the clusters into the 0,1 boundary, I then detach parts of the geometry from that one mesh.
However, along the way I realize that I need to fix some of the polygons on certain meshes; this usually involves fixing vertex positions; sometimes deleting and creating polygons.
So that means I have to fix those clusters in UV space to scale and match proportionally; not only that, I find that I have to reattach those meshes with other meshes so I can pack those clusters again.
Hopefully you can see where I'm getting at, is there a better way of dealing with this tedious process? (that is avoid repetitive reattachment of those meshes and packing clusters).
You don't have to merge meshes to edit their UVs together.
Select all components and add a UV Unwrap modifier and you'll see it in the stack with italics, this signifies it's an instanced modifier i.e. it's used in more than one location.
This will allow you to edit UVs on multiple meshes without merging them
The issue with this isn't that you can't UV multiple objects under an instance. It's that UV normalisation and a few other tools don't work under instance.
I got myself in a situation where I merge different meshes to one mesh so I can UV map them. After flattening, relaxing and packing the clusters into the 0,1 boundary, I then detach parts of the geometry from that one mesh.
However, along the way I realize that I need to fix some of the polygons on certain meshes; this usually involves fixing vertex positions; sometimes deleting and creating polygons.
So that means I have to fix those clusters in UV space to scale and match proportionally; not only that, I find that I have to reattach those meshes with other meshes so I can pack those clusters again.
Hopefully you can see where I'm getting at, is there a better way of dealing with this tedious process? (that is avoid repetitive reattachment of those meshes and packing clusters).
You don't have to merge meshes to edit their UVs together.
Select all components and add a UV Unwrap modifier and you'll see it in the stack with italics, this signifies it's an instanced modifier i.e. it's used in more than one location.
This will allow you to edit UVs on multiple meshes without merging them
The issue with this isn't that you can't UV multiple objects under an instance. It's that UV normalisation and a few other tools don't work under instance.
Yeah that's one issue, you'll have to use select one object and either use textools get and set texels on the other one or I think Advanced UV Normalizer worked across objects though. Not sure about other tools failing you but it would be nice if Adsk fixed that.
I'm pretty sure I'm overcomplicated this a shit ton! I've tried using 2 different approaches: 1. model flat, subdiv and bend. 2. model on a cylinder with lots of sides. would love to see a cleaner way of doing this shape, please.
I'm pretty sure I'm overcomplicated this a shit ton! I've tried using 2 different approaches: 1. model flat, subdiv and bend. 2. model on a cylinder with lots of sides. would love to see a cleaner way of doing this shape, please.
That looks clean to me! Perhaps a tad too many edge rings around that central cutout but I don't see the need to improve it. You can obviously carve a lot out of it to create a low poly asset of it but for a high poly that seems perfect.
Hey guys, I'm currently modelling the barrel of a shotgun here is what I've got so far:
This will be a game model so I will bake the high poly down to a low poly mesh but as it stands now the edge looks a bit soft because there is no loop running on the side where there is a highlighted face because as you guys know, it will give a hard edge along the vertical cylnder edges. I'm worried if I leave it as is, I will get baking issues because there is a bit of a difference in the high and lowpoly meshes.
What do you guys think? I tried insetting the side but that gave me other issues because the support edge did not run all the way down.
If you first line up the cuts, then decide on amount of segments for the cylinder shape so their geometry match up, you'll get the cuts spaced evenly. Also consider having double the segments in the cylinder, as those chamfered edges turned out to not have a curve in them anymore.
I've been combing through the forum to try to figure out my problem but I can't seem to figure it out; I'm modeling in Maya. It seems like I'm not planning something right when it comes to getting the shape, but I'm not sure how to approach it. It's a similar problem to some previous ones but just can't wrap my brain around how to do the geometry to get it to not pinch and have an irregular surface. I'm also not sure if it's better to add edge loops or to use bevel to keep edges. Is there an industry recommended or is it kinda personal preference?
start very low, let the program do your work and done.
You could also start with more geo, other than to use my workflow. You just need to let your edges "flow" around your model. Dont let multiple edges end in one corner right next to your curve.
As you can see at the bottom right, I want a rectangular divet in which I will insert a hinge mechanism. The heart outline did not used to have so many points: I added more to accomodate the divet and still give a roughly even density of ege loops.
I am trying to get the divet to maintain its rectangular shape in subdivision. Here is what I'm working with before any attempts are made. Obviously it's tending towards more of a semi-circular shape in profile because there aren't enough edges supporting the rectangular structure:
Here are a couple of attempts I've made to provide these loops:
It's not really taking. I think I'm going about this the wrong way, conceptually. Any advice?
I have become frustrated in trying to model this mouse:
i cannot even create the base mesh...you would think something so small would be easy to create...Nope. Can someone help me with this? i use Blender 3D...
What have you tried so far? The mesh is basically a subdivided cube with some hard edges, apply some subdivisions and then split the mesh in parts.
i have tried using a cube, and i create the top part, but then to create the side part is very hard. @.Wiki
Create a rough basemesh first, dont try to model only one part first.
Something like this should be a good start. Just subdivide it, add some bevels and split it in parts. But first you need to get the basic shape of the object done.
Making the locket again from scratch for better topology, and this time I'm trying to simply extrude out of the perimeter for the hinge that connects each half. Still gives me unfortunate shading.
Making the locket again from scratch for better topology, and this time I'm trying to simply extrude out of the perimeter for the hinge that connects each half. Still gives me unfortunate shading.
The Importance of having the basic shapes is getting lost. Model the big shapes first, and see if your extrusion can hold the topology, add extra GEO, IF NEEDED. if not and no huge manual tweaks are necessary, go ahead. started with 18 sided cylinder.
1. 18 side cylinder. Made a Box to see that i hade Automatically Edge support for the box. 2. connected my edges. 3. Cut the box in (could've used boolean or simply extrude it out after inserting) 4. added support edges to support my "Box" shape. 5. Done here's a gif with one with a higher subd.
Hello, first post. If there is a better place for this, please let me know. I'm making the Con-Am shuttle from "Outland" as an exercise to get better at hard surface modeling. I've blocked it out and am starting on the more detailed version. I am faced right away with the question of how much to "connect" details vs. just intersecting them on the surface. In particular, making this piece
so far I have this, with diff material on the objects I'm not sure If I should bother connecting or not...
I plan on subD'ing this model. Not sure If I want to bake a low res version, as it's just practice. Oh, also, any comments on edge flow, (and anything else I might be missing) would be appreciated as well. thanks.
The Importance of having the basic shapes is getting lost. Model the big shapes first, and see if your extrusion can hold the topology, add extra GEO, IF NEEDED. if not and no huge manual tweaks are necessary, go ahead. started with 18 sided cylinder.
1. 18 side cylinder. Made a Box to see that i hade Automatically Edge support for the box. 2. connected my edges. 3. Cut the box in (could've used boolean or simply extrude it out after inserting) 4. added support edges to support my "Box" shape. 5. Done here's a gif with one with a higher subd.
The importance of getting the initial shape isn't lost on me, that's kind of just it: the cage I'm using for the extrusion is already three subdivisions on top of the cage I'm using for the basic shape:
Was therefore kind of reluctant to add even more subdivisions to support this extrusion, just because the polycount would get that much higher, and unlike a cylinder, adding additional loops around the perimeter where I want to support the extrusion will deform the curve pretty noticably. But it sounds like there's no way around it, so I'll have to subdivide more.
@AstralZombie First question is "Do I need to connect these parts?" which is a tossup with greebles like this. An easy trick for making these look connected is to keep them as separate floaters, but extrude the bottom of each part to make a little skirt that gives the same look. This is mostly a baking technique however.
@throttlekittyThanks for the tip! yes, have done that for baking, don't see why It couldn't be used for a hi-res model as well, to keep the poly count a bit lighter.
The Importance of having the basic shapes is getting lost. Model the big shapes first, and see if your extrusion can hold the topology, add extra GEO, IF NEEDED. if not and no huge manual tweaks are necessary, go ahead. started with 18 sided cylinder.
1. 18 side cylinder. Made a Box to see that i hade Automatically Edge support for the box. 2. connected my edges. 3. Cut the box in (could've used boolean or simply extrude it out after inserting) 4. added support edges to support my "Box" shape. 5. Done here's a gif with one with a higher subd.
The importance of getting the initial shape isn't lost on me, that's kind of just it: the cage I'm using for the extrusion is already three subdivisions on top of the cage I'm using for the basic shape:
Was therefore kind of reluctant to add even more subdivisions to support this extrusion, just because the polycount would get that much higher, and unlike a cylinder, adding additional loops around the perimeter where I want to support the extrusion will deform the curve pretty noticably. But it sounds like there's no way around it, so I'll have to subdivide more.
@AstralZombie First question is "Do I need to connect these parts?" which is a tossup with greebles like this. An easy trick for making these look connected is to keep them as separate floaters, but extrude the bottom of each part to make a little skirt that gives the same look. This is mostly a baking technique however.
second that, even though it's not for baking, for HS like that, it can save time to do it that way instead of modelling it in. But if it's for a HS practice, go ahead. Start connecting the Bigger shapes, and then slowly go to medium shapes and, if you have enough of geo to hald the smaller parts, go ahead. It's a great practice.
@AstroZombie If you intend for the final piece to look like a real object, going beyond object intersection and actually combining parts that are combined in a real world scenario, will give you a better result. If its lost because of distance and not even noticed because its so small then go ahead. I usually make sure I merge anything that is merged IRL and if it floats on the surface, I give it a skirt. So at least it has an actual intersection between the geo, not some phasing wizardry.
Obviously one takes longer and is much more effort, the other, simpler and faster. The results usually speak for themselves though.
Left (floating Intersections, turbo and edgeloops) Right (Merged shapes, turbo, and edgeloops)
@Mossbros@wirrexx thanks much. yea, decided to go ahead and try to do it right. I'll post results later tonight to hopefully get some input on edge flow.
ok, had some things come up that had to be dealt with ... and, this took me WAY longer than I expected. but, here is progress. The white pieces aren't attached yet, as this is as far as I can go and still use symmetry, so, if there are major things I can fix edge-flow-wise, I can do it before adding the non-symmetric parts.
Thanks so much guys. Still struggleing with it a bit but think I understand a little better now. Kind og similar to the first attempt but think I improeved it in a few areas, lacks the roundless in area A from above though.
Thanks so much guys. Still struggleing with it a bit but think I understand a little better now. Kind og similar to the first attempt but think I improeved it in a few areas, lacks the roundless in area A from above though.
still not correct though, trying to comprehend what Perna says is not that easy, because his IQ is far beyond anybody elses. (love you Per). Try to break down the meshes without welding them together . Like pernas picture, i mean it would make more sense if he would do the mesh, but honestly, his been doing this for far to long and i think it tires him somehow! :P
"RIGHT: Define plane A as the widest point of the shape. This allows us from B and downards to create a perfect curve like in the concept. In order to make the upper, symmetrical, part of the curve to happen we need a triangular shape (BLUE), and lo and behold, such a shape is already indicated in the concept! So it's possible to fix this concept without destroying its look. The previously suggested concepts deviate too much." <-- This part, importante!
Replies
I've done some more realistic character modeling in the past but nothing heavily stylized like this. Capturing the expressions of the mouth and eyes is honestly stumping me.
This is what I've done so far:
I've been having a lot of trouble with the mouth. I thought I'd just have it as a black line like in my picture above and move the cheeks with soft selection to give it expression but it doesn't look great at all. Link of it here: https://i.imgur.com/tJXPQDl.gif You also can't have the mouth open the way I'm doing it right now. Should I actually model the lips? Would opting for animating textures give me more the look I'm going for? Same questions also apply to the eyes as well I guess. Any insight would be appreciated!
2 questions
1) How I can create this rope with requirement that I can easily change it's length and width later?
https://i.gyazo.com/e7421bb9af08377bc52c78e8fe56046f.png
2) On screenshot below you can see folds on the balloon. It happens because of the ropes.
https://i.gyazo.com/d59aff704cdb1863e6c4559931371216.png
How I can create folds like this but for this kind of ropes (on screen below).
https://i.gyazo.com/c1a95b53d84ce1ad42ab4978e7bd21ec.png
Again, preferably method which allow change size of the dolds later.
Small tip for the ropes. If you want to define the length you could get procedural, the way you do it depends on the software.
It's that UV normalisation and a few other tools don't work under instance.
Also you know I'm working from memory on these
Hi!
I'm using creases for this example. unselected cylinder is set to 20 faces, half cylinder is 20 face cylinder cut in half and one side extruded.
This will be a game model so I will bake the high poly down to a low poly mesh but as it stands now the edge looks a bit soft because there is no loop running on the side where there is a highlighted face because as you guys know, it will give a hard edge along the vertical cylnder edges. I'm worried if I leave it as is, I will get baking issues because there is a bit of a difference in the high and lowpoly meshes.
What do you guys think? I tried insetting the side but that gave me other issues because the support edge did not run all the way down.
Also consider having double the segments in the cylinder, as those chamfered edges turned out to not have a curve in them anymore.
I've been combing through the forum to try to figure out my problem but I can't seem to figure it out; I'm modeling in Maya. It seems like I'm not planning something right when it comes to getting the shape, but I'm not sure how to approach it. It's a similar problem to some previous ones but just can't wrap my brain around how to do the geometry to get it to not pinch and have an irregular surface. I'm also not sure if it's better to add edge loops or to use bevel to keep edges. Is there an industry recommended or is it kinda personal preference?
Any advice would be really appreciated. Thanks!
start very low, let the program do your work and done.
You could also start with more geo, other than to use my workflow. You just need to let your edges "flow" around your model. Dont let multiple edges end in one corner right next to your curve.
Like that:
As you can see at the bottom right, I want a rectangular divet in which I will insert a hinge mechanism. The heart outline did not used to have so many points: I added more to accomodate the divet and still give a roughly even density of ege loops.
I am trying to get the divet to maintain its rectangular shape in subdivision. Here is what I'm working with before any attempts are made. Obviously it's tending towards more of a semi-circular shape in profile because there aren't enough edges supporting the rectangular structure:
Here are a couple of attempts I've made to provide these loops:
It's not really taking. I think I'm going about this the wrong way, conceptually. Any advice?
Something like this should be a good start. Just subdivide it, add some bevels and split it in parts. But first you need to get the basic shape of the object done.
Take a screenshot of what you literally have now and let us see what you have.
1. 18 side cylinder. Made a Box to see that i hade Automatically Edge support for the box.
2. connected my edges.
3. Cut the box in (could've used boolean or simply extrude it out after inserting)
4. added support edges to support my "Box" shape.
5. Done
here's a gif with one with a higher subd.
I'm making the Con-Am shuttle from "Outland" as an exercise to get better at hard surface modeling. I've blocked it out and am starting on the more detailed version. I am faced right away with the question of how much to "connect" details vs. just intersecting them on the surface. In particular, making this piece
so far I have this, with diff material on the objects I'm not sure If I should bother connecting or not...
I plan on subD'ing this model. Not sure If I want to bake a low res version, as it's just practice. Oh, also, any comments on edge flow, (and anything else I might be missing) would be appreciated as well. thanks.
Was therefore kind of reluctant to add even more subdivisions to support this extrusion, just because the polycount would get that much higher, and unlike a cylinder, adding additional loops around the perimeter where I want to support the extrusion will deform the curve pretty noticably. But it sounds like there's no way around it, so I'll have to subdivide more.
here i try to move it but it get's distorted
Here is a video to help you understand how it's suppose to function https://youtu.be/QRY53mvTTFM
So not just up and right but also to the sides.
But if it's for a HS practice, go ahead. Start connecting the Bigger shapes, and then slowly go to medium shapes and, if you have enough of geo to hald the smaller parts, go ahead. It's a great practice.
If you intend for the final piece to look like a real object, going beyond object intersection and actually combining parts that are combined in a real world scenario, will give you a better result.
If its lost because of distance and not even noticed because its so small then go ahead.
I usually make sure I merge anything that is merged IRL and if it floats on the surface, I give it a skirt. So at least it has an actual intersection between the geo, not some phasing wizardry.
Obviously one takes longer and is much more effort, the other, simpler and faster.
The results usually speak for themselves though.
Left (floating Intersections, turbo and edgeloops) Right (Merged shapes, turbo, and edgeloops)
any crits extremely welcome, thanks
Try to break down the meshes without welding them together . Like pernas picture, i mean it would make more sense if he would do the mesh, but honestly, his been doing this for far to long and i think it tires him somehow! :P
"RIGHT:
Define plane A as the widest point of the shape. This allows us from B and downards to create a perfect curve like in the concept. In order to make the upper, symmetrical, part of the curve to happen we need a triangular shape (BLUE), and lo and behold, such a shape is already indicated in the concept! So it's possible to fix this concept without destroying its look. The previously suggested concepts deviate too much." <-- This part, importante!