Can someone explain the process of modeling?

I am having a really hard time understanding what i am actually doing. First of people say that you should make Normal Maps and ect but i really don’t get how. Do i model the high poly version first and then the low poly? How then do i get the right shape? And also i see alot of renders which look really good with simple meshes but how really do i make the textures of my mesh nice what i am currently doing is basically Just uvmap export the uv’s add texture and bang i have my not so good looking texture.
Please answers this in a somewhat easy to understand and informative manner cause I try to read it elsewhere but really don’t get it seems like it differentiates from person to person. As an example if you check my portfolio you can see a SciFi gun i made which does not look very nice and i don’t really get how i can make it better.

Thanks for you answers whatever it may be.

IMO, I would not worry about UVs and Normal maps until you understand the basics of modeling. I did look at your portfolio piece and it looks like there are some basic concepts that you might not be understanding but don’t be discouraged. Modeling can be kinda hard until you wrap your head around it. Maybe you should check out BlenderCookie and find some basic tuts to start off with. Go slow and do some simple models (that gun is quite complicated to be honest… probably not the best place to start).

Good luck and keep working at it. You’ll get there quicker than you think. :slight_smile:

its a very deep question. and impossible to explain in a forum post.

Instead let me ask how you have been learning. do you watch tutorial videos? or do you just google how to do things? There are many great tutorial assets on the web. worry about learning the basics first, box modeling, etc. before you even think about moving on to normal baking etc.

as for texturing, that is an art all on its own in the 2d realm. Good textures start with good unwraps. if your using smart unwrap, stop! you have to think like a painter, assign your own seams to the model in places that are convenient for you when your painting the texture, and use the normal unwrap. it can take a long time with some models to get a clean set of UVs for a texture. but you only have to do it once if you do it right, so take the time and do it right.

as for the normal baking question. its a little of column A, little of column B. what i usually do is, at some point in the modeling process when i have the sculpting or detailing “basemesh” completed, i save a copy of it off to another layer, then i proceed to detail the model as i see fit, then i have bring the copy back and use it for baking. of course this is a trick i’ve only used a few times as most of my work as been with low poly modeling. so i might not be the best source of info on that.

why don’t you try to narrow down your questions to what seems to be the most important to you first, and we can go over it in more detail.

A question what do you think i could have done better?

I have been using mainly tutorials for help and sometimes i just google how i can do certain things. And it is hard for me to say what basics i don’t know. And what exactly is box modeling? was that not what is used?

There is so much to modeling, that’s why people said it’s not possible in a thread. It takes time to learn too and once you’ve learned, you keep learning more and never stop.

Box modeling is a modeling style where you start with a simple form and modify that form until you have the structure for your final model. That primitive is often a box/cube, where the name comes from. To compare, another modeling style is extrusion modeling (also known as poly-by-poly or edge-by-edge) where you outline main flows in the model structure by extruding continuous set of edges or polygons and then fill the gaps between those.

Which one to use? Whichever you like. With experience you will typically use both.

box modelling is only a technique for when you start with a default object (often a simple primitive like a cube) , and with adding loop cuts, subdividing, extruding, etc, you model something out of it

edit : oops JA12 was faster

[Demew;2732130]

I won’t say this is the right way or best way to do it, But it is what works for me as a self learning retired hobbiest. Now that the caveat is is done with. Here is my workflow such as it is.

1- Save and make duplicates often. It will number them up for you and you can always mix and match between versions of your work.
2- Make the form of what you are doing first and keep a copy of the lowest poly version of your model as possible. This is useful for baking the detail from the higher detail level mesh to your lower detail level mesh.
3- spend some time using nodes and import as images. Get a feel for it.Play with the node groups and spend some time looking up nodegroups. Both as google search and from blend swap. You can do a fair job of making your own bump,normal and displacement maps just using nodes. (BTW Baking from surface to surface might be somewhat broken in blender atm But you can manually set up a surface to surface bake. with two image planes. Just putting that out there. Make one a light source works well, And you can output normal to color for aid in that baking.)3A- If you make something cool in your play…Save It for future reference.
3B- https://www.youtube.com/watch?v=W07H7xeUnGE

4-Don’t fall in-love with any one piece of work. Or with any one tool. Don’t be afraid to scrap and move on.

Oh I think I have been using both (mainly extrusion) but why should i use just box modeling? it is basically the same thing is it not? Cause if that is the basics i have known that for a while and I am well above that in the learning curve (i think)

Simply put, those are just words used to describe the process. I.e. when asked “how did you do this?” you would say “Oh, I box-modeled it” instead of saying “Well, I took a cube and then added loopcuts and shaped it and then extruded some…”. What I mean is, don’t get fixated on the idea that you might have “missed” one technique or the other. Anyone building something with polygons uses those techniques whether they know how they’re called or not.

Demew: after closer inspection of your scifi gun, 2 issues that might give you immediately better results would be to 1: add an edge split modifier under the subdivision surface mod, what this will do is take some of those soft edges and make them hard, while leaving the majority of the surface smooth and curvy, at least i think thats what your trying to achieve with that “indentation” on the side of the gun.

2: the red shroud needs some thickness, it looks like a piece of paper wrapped around it in those cutouts. you can accomplish this 2 ways. one is to apply a solidify modifier in the stack, preferably above the subdivision, the other is to simply select the shroud and extrude it along face normals, then scale to taste.

if you would like, you can post the .blend for it and i’ll take a deeper look at your topology and see if there is anything else that i can point out that will improve your modeling technique.

A problem i found is that i had for some reason applied the Subdivision layer :confused:
I guess i could do a quick remodel doing the same thing?

yeah, thats a problem! haha.

general rule of thumb to go by is NEVER apply modifiers unless you have to. for instance if your modeling a character with a mirror modifier, there comes a point where you have to apply the mirror to continue or other parts of the process wont work correctly. but things like subdivision should stay unapplied in case you need to make changes later.

of course if you are exporting a model for a game engine or CNC machining/3d printing, you will need to apply all the modifiers first, but i usually make a backup copy of the model and stick it on a different layer before doing so. or if your really paranoid, save out the blend file to a different filename. haha

:slight_smile: Well it is not a problem only took about maybe 2 hours to model and i can probably do it quicker this time and make some changes i was not happy with. And how should i send you the blend file?

Do you wan’t me to texture it as well?

sorry for the delayed reply, you can post it to dropbox or some other file sharing service and just drop a link here. textures not necessary, however if you want some pointers on your UV mapping you can go ahead and unwrap it so i can see how you go about it.


There you go not textured though and used some of your advice

ok, taking a look at it now. let me just ask first, what is your intended end use for the model. i notice you had it set to 3 levels of subdivision, which is usually a crutch to cover a mistake in modeling! hehe. if you were going to offer it as a game model or something like that, 3 levels of subdivs would totally break the tris budget of any current game. maybe in 10 years! haha.

let me give it a good inspection and see if i can find a way for you to get the same effect without all the subdivision.

Did not really plan to use it in anything

First issue im seeing is overall rough topology,


Mesh density should be relatively consistent, with extra loops to focus on edges you want to be more defined when you subdevide.

i took the liberty to fix up a portion of the model, starting here:


After several merges and some vertex alignment, i get this:

which is not only more visually pleasing in mesh form, it will give you better results when you subdivide, as well as give you good edge flow for loop operations. take the subdivision modifier off, and just go through the whole model, see how consistent you can get the topology. once everthing is nice and neat, you might be able to get the effect you want with smooth shading and some corner bevels and just leave the subdivision modifier off.

let me know if you have any questions so far!