CAP 19 CAP 19 - 3D Wireframe Submissions

paintseagull

pink wingull
is a Top Artistis a Forum Moderator Alumnus
Welcome to CAP's very first 3D modeling submission stage. These rules are a work-in-progress and may change as we go along. I expect this will be a learning experience, so your feedback and questions are appreciated and encouraged! It would be great if we could have a model ready for the playtest, but I expect that such a timeline would be far too ambitious. For now, this thread has an unspecified deadline as we work out how much interest there is, problems with the system, and how long it takes people to finalize submissions. We are planning to have a vote on wireframes only, after which the winning wireframe will be finished in a collaborative way. This way, the community has a say in what the final model will look like, but all the work is not put on the shoulders of an individual modeler.

If you're interested in rigging, UV mapping, or animating, please let us know, and stick around! Wireframe modelers will need your feedback when it comes to optimizing their models for these next steps.

Your 3D Coordinator is Yveltal

Rules
  • Wireframes will be created in 3D Modeling software such as Blender (open source), 3ds Max or Zbrush. The software must be able to export to .obj format.
  • Wireframes should be inspired by the winning design from the Art Poll and the Sprite Poll. It does not need to be an exact rendition of every detail of the design; "artistic license" is granted. However, drastic deviation from the selected art and sprite designs is discouraged.
  • Wireframes should have polygon counts to mimic the style of Pokemon models in Pokemon X and Y. These models have counts ranging between 2,000 and 10,000 depending on complexity.
  • Wireframes should be created using a neutral pose. Rigging and animation will need to be done at a later stage, so keep this in mind.
  • When posting WIPs, please post screenshots or renders of your wireframe in a compressed image format such as jpg and host them on external sites such as imgur or puush. Please keep these images under 640 pixels in either dimension and under 200 kB.
Final Submission

All modelers must make a final submission post conforming to the following:
  • The post must have "Final Submission" (in bold) as the first line, followed by three compressed images under 640 px in either dimension and under 200 kB. These images should give a front, side and 3/4 view at eye-level. The 3/4 view can be adjusted to best show off your model.
  • The post must also link to the .obj file of the wireframe, hosted externally on a site such as dropbox. No lighting, textures, colour, or UV maps should be applied to the wireframe in the .obj file. (Submitters have the option of sending the .obj file privately to the Wireframe Coordinator if they would prefer).
  • Wireframes that have extraneous details left out or which are still works in progress are acceptable submissions. The chosen submission will be edited later for detail and quality control. At this stage, the focus is on proportions, shape, and an appropriate pose. If their submission is chosen, a modeler may continue to work with the Coordinator on their model during the Quality Control stage.
  • By making a final submission you agree to allow the modification of your submitted wireframe and for it to have UV maps, rigging and animation applied to it.
  • Only make one (1) final submission post.
If the Coordinator feels that a wireframe is not suitable to be rigged or UV mapped in the future, it may be disqualified. Otherwise, all legal final submissions will be included in the poll.

3D Modeling Stages

  • 1. Wireframe Submissions <- you are here
  • 2. Wireframe Poll
  • 3. Wireframe Quality Control
  • 4. a) Rigging / b) UV mapping (these are concurrent)
  • 5. Animating (needs only Rigging to be finished to begin)
Advice for Modelers

- To be added -

Main Design


Additional supporting material by the design artist (Magistrum) can be found here. Modelers are NOT obligated to use any design elements of artwork not represented in the Main Design. This includes details shown only in back views or other action poses by the design artist. The only official reference artwork for this project is the Main Design. Spriters have priority in the interpretation of details not shown in the main design -- it is encouraged but not required to refer to the winning sprite when finalizing model details.

--------------------

Concept:
Yilx said:
Name: Einherjar ~Acta Est Fabula~

Description: A Pokemon that dissuades your opponent from fainting it, or can even leave it's presence on the field felt even after it faints.

Justification: When a Pokemon faints, it's usually thought of as the battle having gotten down to a 5-6. However, we've yet to discover if a Pokemon can leave a lasting impression on the battle even after having fainted; be it through moves like Healing Wish and Destiny Bond, placing hazards that the opponent can't remove as their removal has been taken care of, or by leaving an opponent's key member weakened and/or taken out.

Questions To Be Answered:
  • How can a Pokemon leave a long-lasting effect on the rest of the battle with just it's moves?
  • How the hell is it different from simply ramming a sacrificial martyr into your opponent's team and hoping it punches holes in it?
  • Building on the previous question, is it possible to build this Pokemon as a defensive threat rather than a "Glass Cannon"?
  • Is it even possible for a Pokemon to leave a lasting effect on on the battle, even after it faints?
  • Could changing your opponent's way of thinking even be plausible? From, "I need to take CAP X out!", to, "Damn, if I take CAP X out, I'll be in trouble...!"
Type: Electric/Poison
Abilities: Storm Drain / (Insomnia/Vital Spirit)
Stats: 60 HP / 57 Atk / 119 Def / 131 SpA / 98 SpD / 100 Spe

Leadership Team:

Pwnemon- Topic Leader
ginganinja - Ability Leader
srk1214 - Typing Leader
alexwolf - Movepool Leader
Deck Knight - Stats Leader
 

Magistrum

DOITSU NO KAGAKU WA SEKAI ICHI
is a Top Artist
Hiya! for those interested in making the 3d wireframes, paintseagull and Yveltal suggested I draw some other views of the design to help with some details. So yea here are some orthographics for reference:

Also a bit of a reminder, since some of the details aren't seen in the main design (like the base of the dorsal fin, which is on a yellow circle), modelers can still deviate as stated in the OP if they prefer. ^_^b
 
Top
Bottom
Side
Front

Thank you, Magistrum, for making those easy to square off and put into 3D space. I'll be updating with my own entry whenever I have free time.

To anyone else hoping to participate, I will also try to remain actively logged in both on the forum and in #cap and #smeargle on IRC, if any questions come up. The winning entry can get away with a lot of missing detail (the back-rings and gills do not HAVE to be modeled on to make a good entry, for instance) as long as it looks proportionally accurate by our deadline...whenever that is. A disqualification can come from a couple of different mistakes that would make your model to hard to work with, normally stemming from geometry issues. I don't anticipate having to ever disqualify somebody, but the best response for if it comes up is, "I'll know it when I see it"
 
There is a very few answers to this thread so I hope I'm not posting this too early.....
:toast:
There are some renders of my model:
(Special effects are only in the renders, I deleted them in the included file)

General poses:
Flat:
http://i.imgur.com/8dkkLPf.jpg
http://i.imgur.com/5T7i7r0.jpg
Smooth:
http://i.imgur.com/H4xf2DS.jpg
http://i.imgur.com/A7tNYN0.jpg

Top
Flat: http://i.imgur.com/4pkOXZB.jpg
Smooth: http://i.imgur.com/lDWbZkp.jpg

Side:
Flat: http://i.imgur.com/ijWRox4.jpg
Smooth: http://i.imgur.com/jHkzOqf.jpg

Front:
Flat: http://i.imgur.com/qwUUPkR.jpg
Smooth: http://i.imgur.com/n2nMLRb.jpg

And the model itself: https://www.dropbox.com/s/vnvyxa6no44emys/StingRaySmooth.obj?dl=0

EDIT:
A little update. This time flat and smooth versions are just different render options, but the model is the same. This model is previous smooth model with sharper details (rings, stings, etc.). I don't think I will be modeling in details like polka dot pattern or gills, because I think it's better to implement them in textures....
(somebody correct me if I'm wrong...)

Model: about 3500 faces.

It's my first somewhat serious model and I don't know if its good enough for the CAP.
Please tell me if there is something wrong with this model, so I can change it.
 
Last edited:
You're right about the grill being better off implemented with the texture maps. Also nice work on that base mesh. Although I believe for this model it would be worth making the poly count a little higher just because you have quite a bit of room left and when the time for animation comes, it will also look smooth.

As I work on my own mesh, I'm thinking that the golden rings and purple spikes that protrude out of his fine may be best as separate meshes as to avoid it looking like it will blend with the surface of his body too much when it comes to texturing. Normal maps may make this less of an issue, but I think the poly count limit leaves room to do that and just play it safe.
 

Pikachu315111

JAPE Judge!
is a Community Contributoris a Smogon Media Contributor
If no one minds me commentating, now that there's two models I'd like to say I like how Koldrage did the body BUT I like viiragon's tail design. I know both aren't complete (and G0dsl4yer hasn't shown his model yet), but thought I would say that to maybe give any of you some ideas to improve your designs. :)
 
If no one minds me commentating, now that there's two models I'd like to say I like how Koldrage did the body BUT I like viiragon's tail design. I know both aren't complete (and G0dsl4yer hasn't shown his model yet), but thought I would say that to maybe give any of you some ideas to improve your designs. :)
Thanks, but don't get too attached to how the body looks. Once I get to do the retopoly some detail will be lost, sadly :(
Speaking of retopology, it'll take a while longer. I have to make 2 mini comics for this week and totally I'm losing my head.

Later guys!
 

Pikachu315111

JAPE Judge!
is a Community Contributoris a Smogon Media Contributor
Thanks, but don't get too attached to how the body looks. Once I get to do the retopoly some detail will be lost, sadly :(
Speaking of retopology, it'll take a while longer. I have to make 2 mini comics for this week and totally I'm losing my head.

Later guys!
Don't worry, with 3D modeling that's bound to happen. Details lost can be fixed or covered up with later texturing. In the end as long as the final model looks complete that's all that really matters.
 
Thanks, but don't get too attached to how the body looks. Once I get to do the retopoly some detail will be lost, sadly :(
Speaking of retopology, it'll take a while longer. I have to make 2 mini comics for this week and totally I'm losing my head.

Later guys!
Well I think the point was to show your re topology/wireframe version anyway. I don't think we needed to see the hipoly version in the first place. But it's still nice to see how you're going about it.
 
~Coordinator post~

I'd like to talk about a few things. Keep in mind, the "best" model will be the one that not only makes the design look the best in 3D space, but who's topology is most ideal for texturing an animating. I will try to express my criticisms in a way that explains what I believe to be best practices:

viiragon: Proportionally, the body looks very nice to start. the back spike, fin spikes/claws and facial extremities don't feel entirely on-point, though. That doesn't mean you shouldn't keep editing the main body, though. It's true that, while smoothed, it would be passable from a modeling standpoint, but texturing and rigging your submission would be a horrendous experience. You have an unsettling presentation of diamond-shaped quadrilaterals all along the main body, as if you put a very minimal turbosmooth on it just to round everything out and get a definitive shape. Your seam down the middle, likely a result from mirroring your geometry, could dare to be softened just a bit. The ear spikes and gill spikes also look very tacked-on at the moment. Hopefully you'll find a way to tailor the mesh around his neck to make those features look more intentional.
Meanwhile, the extremities and tail look like they were modeled by a completely different person, with the only major exceptions to the quality of the tail being 1; the glaring transition in polygonal layout from the very base of the body to the base of the tail, and 2; the uncomfortable looking split that the tail has at the base of its fork. Really, to make your submission closer to acceptible, the vertices on the body should feel a bit more uniform. You won't lose the organic feel that your model has if you make it appear more structured and less...generated.

koldrage: Depending on how your retopology phase goes, you may have less to worry about as far as the layout of your polygons goes, as we can probably expect to see it generated as squared-off triangles. You don't have to worry so much about the detailed details, which means you can smooth out the mouth and gills/pointy bits on cap19's underside. Other than that, your proportions worry me. The eyes look a little sunken, and the tail looks completely separated from everything else. Part of that fault goes to the curves of the mesh (as I see them through your screenshots) for looking a little too dramatic as they move on from the body to the tail, and the rest of the blame is on the harsh and inorganic-feeling placement of that ring. I'm also displeased with the spikes/claws on his fins. They don't look cartoony enough, if that makes sense. Maybe scale their width to make them look a little more plump, and see if that helps them.

-On to general concerns:
You're right about the grill being better off implemented with the texture maps. Also nice work on that base mesh. Although I believe for this model it would be worth making the poly count a little higher just because you have quite a bit of room left and when the time for animation comes, it will also look smooth.

As I work on my own mesh, I'm thinking that the golden rings and purple spikes that protrude out of his fine may be best as separate meshes as to avoid it looking like it will blend with the surface of his body too much when it comes to texturing. Normal maps may make this less of an issue, but I think the poly count limit leaves room to do that and just play it safe.
Having more polygons for ease of animation is both true and false. While you put yourself at less risk of breaking the model just because you rotated your animation controls too far one way, once you add in a rig, skin weights, a cel shader and a texture that references an image, you're going to have a bigger file to open, so having less polygons in the interest of reducing your chances of crashing would be more ideal. That's really what we hope to accomplish by having multiple entries and quality control just for modeling.

Separate meshes would be tricky and unnecessary as well. When laying out the UVs, you can relocate and prioritize the UVs however you like. For example, if you want to do something special with the claws/nails hanging off of the fins but you're afraid that having the UVs where the body meets the extremity connected in your texture map would make it hard to outline, you can move the claw/nail UVs away from the main body and mess with the color however you like. Which means, you should just make it all one object. Helps with rigging.
This isn't really a fix to what you were trying to say, but you made an incorrect implication that made me want to say that.
The back-rings can be painted on to a perfectly smooth back, and we shouldn't have to worry about it beyond that. Your implication would involve having a hi-poly and a low-poly model to bake geometry definition. Not only do we not have the time, but the end result would be visually unfit to be in XY/ORAS. Because it would look "too detailed".

Don't worry, with 3D modeling that's bound to happen. Details lost can be fixed or covered up with later texturing. In the end as long as the final model looks complete that's all that really matters.
That doesn't really happen in the professional world?
Material/textures exist to embellish on detail that you've already given your model most of the time, by adding color and alpha maps to dictate how the character looks in a lit environment. If you're going to have a large number of objects in the screen all at once, having a lowpoly character with a normal map and a bump map that references hipoly modeling prevents a loss of visual quality by making the character appear to be more detailed than it should be.
As the submissions get closer to being finished, I'll have more chances to talk about what the best solution will be for creating the material.

~~~~

WIP:

It's a couple of screenshots all clumped together, I didn't have much time after I got what I wanted. I'm trying to model it based on how I see the UVs laying out in a photoshop document, along with what I believe a rig would be able to deform the best, since I still haven't decided how I'd want to animate an idle position for this guy.
 
Last edited:
~Coordinator post~

koldrage: Depending on how your retopology phase goes, you may have less to worry about as far as the layout of your polygons goes, as we can probably expect to see it generated as squared-off triangles. You don't have to worry so much about the detailed details, which means you can smooth out the mouth and gills/pointy bits on cap19's underside. Other than that, your proportions worry me. The eyes look a little sunken, and the tail looks completely separated from everything else. Part of that fault goes to the curves of the mesh (as I see them through your screenshots) for looking a little too dramatic as they move on from the body to the tail, and the rest of the blame is on the harsh and inorganic-feeling placement of that ring. I'm also displeased with the spikes/claws on his fins. They don't look cartoony enough, if that makes sense. Maybe scale their width to make them look a little more plump, and see if that helps them.
Yeah the retopology and details and smoothing won't be much of a worry.

Now about the eyes, maybe I did pushed them inward too much, which I can totally fix. The rings, honestly they looked to me coin thin. At least on the concept. Then on the orthgraphic views, they are inflated, which is sligthly confusing. I picked the "coin" ones, since they looked more like the ones on the concept. Finally I'm not too sure about more cartoony claws. I mean I can curve them more, since they are more curved. The thickness is fine to me. I was using the top view to make them wide enough. Maybe when I sharpened and polished the sides a little that may have thinned them a little. Anyways, all of it is fixable, since I haven't gotten around to do the retopology.

Also, thanks for the comment, man. GOtta work work work now! :)
 
Hi Yveltal. Thanks for the response and I can't wait to see how your base mesh turns out.

First and foremost, here are the links to my WIP wireframes

Top - Bottom http://imgur.com/lVcwPiz,Ls3JDZM#0

Side - Perspective http://imgur.com/lVcwPiz,Ls3JDZM#1

Now in response to what you said:

My point for saying the use of more polygons wasn't to make the animation any easier. Quite frankly, that could go either way, but I'm not gonna get into that. The total polygon count that I am working with is 5277. The difference between my polycount and Viiragon's polycount is negligible when it comes to cpu performance. To be honest, even with the skin, rig, and a whole set of texture maps, it's not going to put the software at risk for crashing.

http://i.imgur.com/BMFXNRP.png

I've modeled, rigged, textured, and animated all of these pokemon from scratch. Volcarona is by far the biggest polygon count among all of those pokemon by about 3k polygons, hitting roughly a little less than 10k and that's just because I wanted the wings to have some thickness. With that said, my 3ds Max file containing the finalized Volcarona is 4.3 mbs. A single mp3 file will easily exceed this limit. It is a fact that having less polygons in your scene will reduce the chance of it crashing at any given point, but generally that is the case if your scene has numerous objects of containing a high number of polygons or even an extremely hipoly object. I was simply suggested adding another 500-1k polys so when it comes to animation, it won't look blocky around interest points. This model has a lot of curves and the last thing you're going to want to see are straight edges, which is what my advice derived from, not the ease of animation.

Also, I am very familiar with the process of creating UV maps, especially for game characters. Unwrapping a mesh with multiple objects is just the same as unwrapping one of a single object. When I was in college, my professors always told me that I should keep my mesh a single object, but they never had any foundation as to why that was the case. Upon doing my own research and experimenting with many character's I've created, I found out it isn't at all necessary to keep the mesh one object. In the image with those pokemon, Glalie and I believe Hawlucha are the only 2 pokemon that are one object, but the rest are composed of more than one. There are even tricks you can do with using multiple objects in a mesh to make your life easier later when it comes to animation and even texturing. Regardless of the end result, it is all about the illusion of completeness within the confines of efficiency. My Cap19 model as you can see, the spikes that comes from outside the fin are separate objects and believe me, it will not be any worse that one that is all the same object.

While what you suggested does work, it doesn't bring about the full effect that separating the spikes can achieve. I made the rings of the Capmon all part of the same object because it didn't affect the silhouette too much as I thought it would have, but I have another version of the Capmon where it is all one object, and it just doesn't look as good when it comes to the purple spikes. And I guess I wasn't clear with what I meant about the gold rings. It seems like you thought I was referring to the design on the top fin, but I was referring to the rings that you can find around it's tail and "whiskers". But to comment on that anyway, you say that you don't have time, but to me it seems like these segments of developing a capmon go on for a while before the next phase. Even if this wireframe phase lasted for a 2 weeks, it doesn't take long at all to create a highpoly pokemon and create a lowpoly mesh based off of that. In fact, I was practically finished with it a couple days after the rules were posted up. And all of the pokemon in that image has taken me no longer than a day(per pokemon) to make from start to finish (hipoly sculpt - animation). Although I believe I work fast, I don't think it would take anyone too much longer. Then again, I'd be assuming that we all have the same workflow/resources. So, I would say the time issue is a valid point.

And one final thing, you've seen the Malaconda sprite from a while ago, you know that I know how the pokemon should look in game because and that my intentions were not to put in something that was going beyond what XY/ORAS delivers.
 

paintseagull

pink wingull
is a Top Artistis a Forum Moderator Alumnus
I just want to pop in and say thanks to everyone for the enthusiastic contributions to the first CAP 3D project so far! This discussion that you guys are having is way over my head, but I really appreciate the thought and effort that's going into this from everyone. All of the model work is looking awesome to me, I really look forward to our progress on this :)
 

Pikachu315111

JAPE Judge!
is a Community Contributoris a Smogon Media Contributor
Also, I am very familiar with the process of creating UV maps, especially for game characters. Unwrapping a mesh with multiple objects is just the same as unwrapping one of a single object. When I was in college, my professors always told me that I should keep my mesh a single object, but they never had any foundation as to why that was the case. Upon doing my own research and experimenting with many character's I've created, I found out it isn't at all necessary to keep the mesh one object. In the image with those pokemon, Glalie and I believe Hawlucha are the only 2 pokemon that are one object, but the rest are composed of more than one. There are even tricks you can do with using multiple objects in a mesh to make your life easier later when it comes to animation and even texturing. Regardless of the end result, it is all about the illusion of completeness within the confines of efficiency. My Cap19 model as you can see, the spikes that comes from outside the fin are separate objects and believe me, it will not be any worse that one that is all the same object.
Hmm, that's odd, in my Maya class I was taught that one the most important thing to do with a model with multiple objects it to make sure they were properly parented so that the child objects moved along with the parent (say a person's limbs (child) moved if the body (parent) moved. In addition the limbs can move without effecting the body unless it's a parent to other objects).
 
Hmm, that's odd, in my Maya class I was taught that one the most important thing to do with a model with multiple objects it to make sure they were properly parented so that the child objects moved along with the parent (say a person's limbs (child) moved if the body (parent) moved. In addition the limbs can move without effecting the body unless it's a parent to other objects).
Yes, in a lot of cases when an object is split up into multiple objects, but are a part of the same object, it is important to make sure that all the child objects are correctly linked to the parented object. This is pretty much key in animation when using a custom rig/bone setup, however there are other rigging systems that allow you to build your rig with those things in place already such as the CAT rigging system in 3dsMax which is what I use.

When it comes to skinning an object, if your bone system is properly set up (with the child/parent in place), there will be no need to set up your base mesh in that way (making the spikes a child of the body). Your rig will exhibit those qualities you mentioned.
 
Hi Yveltal. Thanks for the response and I can't wait to see how your base mesh turns out.

First and foremost, here are the links to my WIP wireframes

Top - Bottom http://imgur.com/lVcwPiz,Ls3JDZM#0

Side - Perspective http://imgur.com/lVcwPiz,Ls3JDZM#1

Now in response to what you said:

My point for saying the use of more polygons wasn't to make the animation any easier. Quite frankly, that could go either way, but I'm not gonna get into that. The total polygon count that I am working with is 5277. The difference between my polycount and Viiragon's polycount is negligible when it comes to cpu performance. To be honest, even with the skin, rig, and a whole set of texture maps, it's not going to put the software at risk for crashing.

http://i.imgur.com/BMFXNRP.png

I've modeled, rigged, textured, and animated all of these pokemon from scratch. Volcarona is by far the biggest polygon count among all of those pokemon by about 3k polygons, hitting roughly a little less than 10k and that's just because I wanted the wings to have some thickness. With that said, my 3ds Max file containing the finalized Volcarona is 4.3 mbs. A single mp3 file will easily exceed this limit. It is a fact that having less polygons in your scene will reduce the chance of it crashing at any given point, but generally that is the case if your scene has numerous objects of containing a high number of polygons or even an extremely hipoly object. I was simply suggested adding another 500-1k polys so when it comes to animation, it won't look blocky around interest points. This model has a lot of curves and the last thing you're going to want to see are straight edges, which is what my advice derived from, not the ease of animation.

Also, I am very familiar with the process of creating UV maps, especially for game characters. Unwrapping a mesh with multiple objects is just the same as unwrapping one of a single object. When I was in college, my professors always told me that I should keep my mesh a single object, but they never had any foundation as to why that was the case. Upon doing my own research and experimenting with many character's I've created, I found out it isn't at all necessary to keep the mesh one object. In the image with those pokemon, Glalie and I believe Hawlucha are the only 2 pokemon that are one object, but the rest are composed of more than one. There are even tricks you can do with using multiple objects in a mesh to make your life easier later when it comes to animation and even texturing. Regardless of the end result, it is all about the illusion of completeness within the confines of efficiency. My Cap19 model as you can see, the spikes that comes from outside the fin are separate objects and believe me, it will not be any worse that one that is all the same object.

While what you suggested does work, it doesn't bring about the full effect that separating the spikes can achieve. I made the rings of the Capmon all part of the same object because it didn't affect the silhouette too much as I thought it would have, but I have another version of the Capmon where it is all one object, and it just doesn't look as good when it comes to the purple spikes. And I guess I wasn't clear with what I meant about the gold rings. It seems like you thought I was referring to the design on the top fin, but I was referring to the rings that you can find around it's tail and "whiskers". But to comment on that anyway, you say that you don't have time, but to me it seems like these segments of developing a capmon go on for a while before the next phase. Even if this wireframe phase lasted for a 2 weeks, it doesn't take long at all to create a highpoly pokemon and create a lowpoly mesh based off of that. In fact, I was practically finished with it a couple days after the rules were posted up. And all of the pokemon in that image has taken me no longer than a day(per pokemon) to make from start to finish (hipoly sculpt - animation). Although I believe I work fast, I don't think it would take anyone too much longer. Then again, I'd be assuming that we all have the same workflow/resources. So, I would say the time issue is a valid point.

And one final thing, you've seen the Malaconda sprite from a while ago, you know that I know how the pokemon should look in game because and that my intentions were not to put in something that was going beyond what XY/ORAS delivers.
When you first said "add more polygons", the first thing I thought was "Oh god, he's going to turn in some turbosmoothed nightmare of 50k polys as a first pass", so I convinced myself that you weren't going to stay withing the 5k range, give or take a few thousand. Not sure why...

My educated guess about using one mesh per model/character, after finishing an animation degree and doing personal projects like this, would be that it makes a game engine run a little easier. A game like Prototype, for example, was probably set up in pre-production to apply only one material to each model, because you could have a multitude of pedestrians and mutants all on-screen at once, each one a couple thousand polygons. In that case, having more than one material on a character means accessing double the amount of memory it takes to load just one color map (forgiving normals, AOs and bumps for the sake of the example), multiplied out by every object in the players line of sight. A game developer who made one of the Ghostbusters games once told me that the engine he was using was tailored to play/load the game at an optimal speed when a character is rigged with 32 joints/bones, and to go over that amount would require a costly reprogramming. For those in the game field, using as little as possible is generally a good practice. We know that, in XY, Pokemon can go up to 12k polygons because you can see, at most, 6 models plus a background, which can add up to a maximum of roughly 80k polys on the top screen (during a hoard battle between 6 Volcanions) Typically, there's less then 20k to have to render all at once. That, plus particles effects, lighting, shadows and extra attributes doesn't hurt the game card because everything is modeled so efficiently.

To banter on the topic of models in XY for a minute, it's clear that some of them are built out of multiple sub objects. One of my favorites, Granbull:
(sorry, not enough good screenshots of what I wanted to show)
was modeled with his ears separate from the rest of his head. You can tell because the lighting cast on his head shows on the underside of his ears (look closely at the bottom of his right ear, where there should be a shadow cast by the ear). Sub-Objects don't cast shadows over each other when being rendered in-game, so you can still see where the models break face.

The point of that: Yes, I know it's doable to make a working, riggable model out of multiple objects and textures, but I came into this knowing that entries could come from people of all different backgrounds and experience. Since I didn't know where the general skill level was starting out, I'd much rather have low, meetable standards for everyone and be wowed when I'm shown what talent we have hanging out on smogon than have high expectations for participants. However, since we have at least two people (you and myself) who are aware of the requirements, I'm less opposed to the idea(A part of my initial concern was that Blender makes it kinda hard to use multiple maps, but everyone so far seems to be on Autodesk software). If we went ahead with your model, it could be a nice setup do make some educational posts that I wasn't anticipating to write on our very first 3D project.

Getting back on track, your model is dangerously close to what I've wanted to see in terms of topology. I may have to see it without the wireframe on, but the head doesn't look quite as defined from the neck/body as it should be. I'm also pleased to see your ATD on the tail rings, where you have the really smooth transition leading up to the immediate change in direction as the lines in your geometry form the rings. Kudos.
 
Last edited:
Thank you, and I would love to see the work you've been a part of.

Also I just realized that I may have been unclear about having multiple objects. When I said multiple objects, just the same way that Granbull is set up, I was referring to multiple sub objects along the same mesh. I didn't mean a separate object as a whole. For example how I did the spikes for the CAP pokemon, it would still be one object, but the spikes would be their own sub object, or element. So you would still just be using a single texture on the whole thing. Although just as you wouldn't want multiple objects referencing all different materials in a game, it is also possible to have few materials and a bunch of objects referencing those materials in different combinations to maximize performance and efficiency. I'd imagine you would do something like that to get the most out of civilians/pedestrians in a game.

And yeah, some engines have bone restrictions. If I recall, Unity3d used to have an issue with this a while ago. Then again, it's possible that I screwed up something with my bones and didn't know how to fix it at the time.

With the head, I didn't add too much definition mainly because I was trying to imagine whether or not during his animation if his head would be turning all too much. If so, I can go in and adjust it.
 
Also I just realized that I may have been unclear about having multiple objects. When I said multiple objects, just the same way that Granbull is set up, I was referring to multiple sub objects along the same mesh. I didn't mean a separate object as a whole. For example how I did the spikes for the CAP pokemon, it would still be one object, but the spikes would be their own sub object, or element. So you would still just be using a single texture on the whole thing. Although just as you wouldn't want multiple objects referencing all different materials in a game, it is also possible to have few materials and a bunch of objects referencing those materials in different combinations to maximize performance and efficiency. I'd imagine you would do something like that to get the most out of civilians/pedestrians in a game.
.
I actually had to reread the conversation before it clicked, Because I've done multi-objects and sub objects that share a single texture image before.
(Sorry, I'm a dingus at the worst of times)
Doing those, at least for making a UV map, makes it a little easier to lay everything out because there's less sewing to worry about.
The more I think about it, the more I like the idea of making the wing claws (and maybe even the tail rings) separate objects. Those simple shapes would make flatmapping a lot easier to do.
 

paintseagull

pink wingull
is a Top Artistis a Forum Moderator Alumnus
Yveltal G0dsl4yer koldrage viiragon

Hey everyone, so the sprite poll will be going up this weekend. Once that is done, we are at a point where we could feasibly vote on wireframes. That said, we don't have to. I'm just wondering what kind of time frame you guys think is reasonable to have your wireframes ready? There is no real need for a deadline, since we will use the sprite in the playtest, but in the interest of keeping things moving, it would be good to try and get a ballpark idea. And just a reminder, wireframes don't need to be perfect at this stage. We can finalize details and minor tweaks after the poll.

And another reminder, if you have any suggestions for how this process should be run, please don't hesitate to tell them to me/Yveltal or just post them in this thread!
 
The wireframe I have submitted will probably look like what I submitted or very close to it. It's not by any means perfect, but as you said and I agree, that changes and tweaks will be made once the poll has gone through.
 
FINAL SUBMISSION

Also, sorry I took this long for this. Haven't had a chance to take a break from college, and now that a hurricane is coming I had to do it sooner before we lose power for a week. x.x

If any links are wonky let me know.
Koldrage, the important part of this part of the poll is to see the wireframe. Have you read the rules of the segment for this entry? Don't get me wrong, it looks good, but we need to see a low poly version/topological make up to know if it'll work for animation.
 
Koldrage, the important part of this part of the poll is to see the wireframe. Have you read the rules of the segment for this entry? Don't get me wrong, it looks good, but we need to see a low poly version/topological make up to know if it'll work for animation.
Oh right right right. Sorry, sorry x.x I messed that up. Do I just update the post or delete it and add a new one?
 

Users Who Are Viewing This Thread (Users: 1, Guests: 0)

Top