Game Development Community

dev|Pro Game Development Curriculum

Learning 3D Modeling, mostly polygon modeling, mostly with Blender 3D Part 2.

by 'Sidikian' · 01/31/2009 (8:26 pm) · 14 comments

Well welcome back,

Welcome to the second part of my study Learning 3D Modeling, Mostly Polygon modeling, Mostly with Blender 3D. If you missed the first part of my study you can find it here.

Hope I haven't lost anyone yet and I hope I've opened some more doors in the great mansion of 3D Polygon modeling.

In some ways I feel like we could have just ended with Boyd's tut. and been said and done but as with anything with me I like to "dig deep" and really "keep moving foward" to get a total understanding across. So with that being said let's get back on the horse so to say.

I'd like to take the time now to point out 2 things which I believe can literally stop you from pulling out your hair and driving you up the wall (not that driving up a wall wouldn't be cool =p). Without knowing about these 2 things you might look at your model and go why does my model look like it was made out of Legos and everyone else's look like they were made out of top quality modeling clay.

So with out further ado I'd like to reiterate what Boyd talked about on pages 12 and 13 of his modeling ethic, which is Gouraud smoothing (Boyd called it display smoothing) and Subdivision Modeling.

Gouraud smoothing is an awesome little tool because it can really make your model look smoothed out with out adding extra geometry to the model. Basically as far as I understand it gouraud smoothing just recalculates how the lighting is distributed across the mesh when and where it's applied thus giving those areas a more smooth appearance.

Note: If you're new to 3D Modeling feel free to skip this next paragraph till later when you're ready to learn more about smoothing. For a further look into how Gouraud smoothing works you can watch Smooth Shading, another great video provided by the guys at The Guerrilla CG Project. It's important for me to point out that Blender has a work around for applying smoothing via angle and you may wish to look at the information under point number 6. found here. Also as far I know at this point in time it's not possible to control the vertex normals in Blender, it's only possible to see them unfortunately.

So just how much of difference can Gouraud smoothing (set smooth button in Blender) make? Check out this screen shot.
i280.photobucket.com/albums/kk168/scott_and_esther/HeadLeftSetSolid.jpgIt's the same exact head, it's just the head on the left has Gouraud smoothing, or as it's called in Blender Set Smooth, applied to it where as the one on the right does not.


Subdivision modeling is "A popular method of modelling using a "low resolution" polygon mesh which is then smoothed to produce a well rounded model. It's advantage is producing high polygon results while only creating a low polygon base mesh. This term is commonly shortened to SubDivs or SubD." as defined on the The Guerrilla CG Project. With that beind said let's watch Glen Moyes tutorial Subdivision Surfaces: Overview, but understand that Glen looks at topology a little differently than Boyd does, but both ways are right.

As it's says in it's definition Subdivision, or subsurface modeling, can make a high poly mesh from a low poly mesh.

So how does this relate to making our models for Torque? Well since we have poly limits for our .dts objects (models) chances are you won't be bringing in a highly subdivided model straight into your torque project (not to mention you'd have to convert the model, alt-c, or apply the subdivision modifier first, but that's another topic).

Subdivision modeling does have it's use though for our models for torque.

We can use subdivision modeling to make a low polygon model look like a high polygon model through a process where a normal map is produced from the high poly model. A normal map is simply a texture that is attached to the lower polygon mesh that tricks the engine into treating the lower poly models's faces as if they were the high poly model's faces when it comes to lighting.

In lame man's terms the normal map gets alot more detail with out the extra geometry. Just remember though it's still the low poly's faces that are really there so when it comes to producing shadows you'll have to make sure the low poly mesh has the basic shape of the high poly's or your shadows in game may not resemble your model's outline.

Now as I said Glen and alot of others look at topology as the mesh's wireframe pattern, which is nothing more than your model with the edges you've created showing. Here you see the wire frame of my model:
i280.photobucket.com/albums/kk168/scott_and_esther/Wireframe.jpgNote: Right side shows the wireframe and left side does not in order to show you what the wireframe is.


With that being said,
Glen pointed out near the end of his tutorial that we've just viewed, that poles can create problem areas for subdivided meshes as can other things so lets look into these issues and how to avoid/fix them with Greg Petchkovski's video Subdivision Topology:Artifacts.

Note: If you are new to 3D Modeling I'd highly recommend using caution when deciding to visit this study I'm about to list because it can be pretty confusing and overwhelming.
If you'd like to further look into the study of topology under the definition of it being a mesh's wireframe pattern a really in depth study of it is/was being done at Subdivionmodeling.com under their threads the The Pole and The Pole-revised. The Pole revised thread is pretty much a shortened form of all the information from the thread The Pole.


In an effort not to turn this blog into a book I think I'll end it here, but first I'll give you a quick look into what I'm going to try to go into in the next step of study.

In our next step we'll try to cover:
-a few modeling methods that I'm most familiar with/see potential in
-Introduction to Blender 3D's interface
-getting used to some of Blender's most frequent tools
-low poly topology and tips to help in key areas


As always please feel free to comment as I go so if you're interested in this study so that I'll know if there is enough interest for the effort it'll take. Also if this study gets enough positive feedback I'm going to make it into a resource to further help others learn. So again please feel free to comment.

NOTE:If in any way I've linked to something that the authors of the works wouldn't want linked to and if you are one of those authors just ask and I'll remove the link.




#1
01/31/2009 (8:29 pm)
Very good read. Thanks Scott!

Do you have any plans of creating a collection at some point? I don't know if you have an ending planned or not. For instance, Lara Engebretson is currently doing a video tutorial series on creating a 3D model and exporting it for Torque use. She usually plans out a series to have a certain amount of parts.

Of course, your lessons are slightly less tutorial and more information based, but still, any organization is good organization =)

I look forward to the next one
#2
01/31/2009 (8:45 pm)
Thanks Mr. Perry I appreciate your comment =). As of now I'm not certain how many parts it'll take to cover the information that I'd love to cover. Basically I think I'm heading the direction of understanding then moving towards hands on, then more understanding, then more hands on. My main goal that I would love to reach is to teach how to create a low poly character, a high poly character, make a normal map from the 2, how to unwrap the mesh to create a texture for it, teach how to rig the low poly character, how to set up the needed nodes for the character, how to weight paint the low poly character, create basic animations for the lowpoly character in .dsq format, and then how to get it exported and into Torque all from Blender 3D. I don't know how far I'll make it lol, but we'll see =p.
#3
01/31/2009 (10:10 pm)
walkercreations.org/blender.html
scroll down 25% of the page.

Neal Hirsig's blender video-tutorials.
#4
01/31/2009 (10:45 pm)
Well I guess my work here is done thanks eb! lol! Hmm perhaps I'll still continue though depending on feedback.

Also eb I encourage you to take a look at some of the links I've posted and then give me a more explained reply.

Take for instance the link's from Subdivision Modeling.com. I doubt that,as nice as those tutorials that Neal has made are, he goes into as much depth as the guys do in those threads about topology, poles, edge loops, etc. Or go back to my first blog on the topic and click on the link on Boyd's tutorial, read it, and then tell me his perspective wasn't worth the read. No hard feelings, just simply stating there is a lot of info to be had out there and just dead ending it to one person's perspective might not give us the full range of potential of learning.
#5
01/31/2009 (11:26 pm)
Holy crap. No one has called me Mr. Perry since I used to teach children's classes. Mich is fine =)

Something I've learned in my short time of teaching and writing docs is that no one has the same teaching style. Video tutorials are different from text, people are different from people. This is why panels and group discussions on the same subject can be so productive. You can never have enough instruction or insight on a single subject.

So, you do have a goal. It doesn't have to be an *end*. Just like I do not have an end for my documentation work for GG. I just have goals and milestones.
#6
02/01/2009 (3:27 am)
Curiosity has the better of me here:
How/Why would I be 'dead ending' something by posting an additional link as a resource for someone to use ? & How would you gather the thought; that I was intending such a purpose to my link ? Quite odd indeed. Laughable really. ;)

- Boyd's perspective on the basic rules & science concerning the use of geometry in cad, does not intrigue me. If I did not already have a grasp of the fundamentals, I may have perused...but the point remains.

I am glad to see that you're enjoying yourself while learning..and that you want to share your experience. Just hold back on the defensiveness a bit, ok ?
#7
02/01/2009 (8:46 am)
Thanks for your additional link eb, after all my main purpose is to have a collection of some decent sites for others to refer to in order for a better learning.

Perhaps I am more defensive than I should be, but for whatever reasons it is how I am, sorry if it was misplaced.

Edit:
Boyd's tutorial is not limited to Cad, if that was what you meant by that statement I do not agree.
#8
02/01/2009 (4:40 pm)
I was 'generalizing' but sure..nothing is ever so simple when based on ideas that are complex.

Once again, Good luck with this...I am sure you will help loads of people with these blogs archived.
#9
02/01/2009 (11:24 pm)
That head is looking pretty good. You make good head.

The polygon flow could be cleaner. Too many "poles," and there is a bit too much variation in the size of the faces on the more open parts of the model.

I think it's great that you write up these blogs. They'll be useful for others, and your last blog got me to check out Blender (I'm a Max/Maya user).
#10
02/02/2009 (10:00 am)
I agree Joe my topology for that head that shows the wire frame could defiantly be better, but it was/is my first main try at a character model.

I was/am trying to find that balance of good structure, without high geometry because of the triangle limits for our characters for torque. The main reason why I posted that screen shot wasn't to show my topology anyway lol, it was just to show what a wire frame of a mesh was and I had that model handy since it was mine lol.

I should perhaps find a different picture with better topology though so I don't leading anyone to thinking that is the way to go topology wise.

Edit: Just threw a quick note in under that picture Joe so people understand better.
#11
02/02/2009 (10:22 am)
Looking good Scott. I'll be following your lead on this one and see if I can come up with something as half as good. So keep up with the tutorials mate, very useful for me and others I am sure.
#12
02/20/2009 (9:30 am)
Changed image to show wireframe. I've been retoping my model to get better edge flow and thought I'd throw that pic. in instead of the other one I had.
#13
02/24/2009 (7:10 pm)
I've recently bloged a part 3 to this study and you can find it here.
#14
02/25/2009 (12:15 pm)
This is turning out to be a fantastic collection of info.
Thank you.