Just don't get it.

Hi, this is probably the dumbest thread ever posted here, but I’m gonna post it anyway. I’ve been following a lot of tutorials on modeling hands, heads feet, and anything else there is a tutorial on. I can usually follow the tutorials pretty easily, and when I’m done, I feel like I get modeling. Then I try to model something else intuitively, and it turns out like crap. How did you guys learn how to inuitively know how to model stuff?
I also have a couple questions about “Rules of modeling”. If something that seems like a simple single object like a table, turns out to be easier to model out of 5 different cubes, do you have to try to make it a single object, or is it ok to have 5 different cubes? If that makes any sense.

There is no “Rule” To learn how to do stuff you just do a bunch of sucky stuff until you start to get stuff that looks better.

For the table I use 5 cubes, parented or joined meshes.

There is no “have to” when it comes to modeling. There are different styles and different techniques. Depending on what you’re going to do with the model, some techniques or tools work better than others, and we’ll be glad to help you out with suggestions if you’re in the middle of modeling something and get stuck.

The thing that helps intuition the best is a reference image in the background or a 2d sketch propped up alongside your monitor.

I’d do the table with subsurf on almost from the very beginning, because I wind up putting it on sooner or later anyway. If it was a picnic table, I’d probably use a single mesh. If the legs were fancy, I’d probably make them separately and then join the meshes into a single object.

Being an eternal semi-noob myself I’d recommend that you do a few more tuts before trying to fly by yourself. It will keep you from learning “bad habits”, which is something that happened to me. In the long run it will save you alot of time and energy.

Another thought: perhaps the tuts you’ve started with are too advanced for your level. The noob to pro tuts on the wiki are really good for building your confidence and I’d recommend that you start with those. If you’ve never modeled anything before and are trying to make a character’s head Blender can be mighty intimidating…

Um, here’s the link:


Different Objects (set of data represented by a pink dot, but is written into code that other datablocks can link to and interact with) are designed for their own sake (Meshes to hold renderable faces and Armatures to deform Meshes. How those two perform at rendertime is the object of the whole 3D CG exercise. How fast you model, or whether one method is faster, or in any other way better, than another doesn’t matter so much as modelling for the best render-time qualities.

The first work begins with a pencil and paper and you work out the proportions then decide an approach. If you’re modeling something mechanical Linked Data (Materials, Textures) and lamps can be used to effect and the task at rendertime can be improved if your models are designed (modelled) with that in mind. So Proportions are also layers for screws, rivets, little spiral meshes stashed away on layer 20 that you can steal an edgeloop from later for the landing gear, how many Scenes, not just the proportion of the model(s) but the proportion of the file(s) for that particular project.

Toon materials look good on smooth lines (edgeloops), so the proportions of the model itself should be done on paper, along with the proportions of the scene (Stage) that it fits into. Proportions are like musical notes played in scale; arms are not only long enough, they’re fat enough, and the fattening lines of the upper arm flow into those of the lower arm, then the hand, like Carlos runs from one bongo-beat to the next on three strings… And the character should fit the proportions of caharacers around it. Even characters for stills can be posed more realistically by using Modifiers so when you go back up the arm to the shoulder your pencil will have to show how you’re going to deal with the Armature or your musical note is going to sound like Sanjia on a bad day.

The only serious modelling tools in Blender are Mesh so you have to concentrate on either Poly modelling (start with a Face or Edge and Extrude and adjust, extrude and adjust) or Subdivision modelling (start with a hollow, rough mesh and subdivide with a knife). Subdivision requires a lot of deleting and rebuilding of faces to direct the flow of edgeloops, but in the end you have to do some of the same repair on Poly models too. Both methods can be simplified if you do the paperwork first.


A good word of advice someone wrote on modeling that I’ll never forget (went something like this):
Keep your mesh as simple as possible, otherwise it will be very difficult to edit later.
In other words, use as few verts as possible.

Also, keep a SubSurf modifier on during editing, and don’t apply it until you absolutely must. And always save multiple versions in case you need to revert to an older one.

Additionally, here are a few tools that are a great help:
Loop Cut: [Ctrl+R] This is great for putting creases into a SubSurf-modified mesh, among other things.
[W] > Subdivide Smooth …this one I sort of discovered by experimentation. It’s a SubDiv method similar to SubSurfing, but for specific edges of a mesh. Good for when normal subdivision is just too rigid.

Edge Specials Menu [Ctrl+E]: This menu contains some great tools like “Edge Slide”, which I now consider to be indispensible.

Change the rotation/scaling pivot according to your needs. This is an option I stumbled upon, which basically allows you to scale and rotate based on the position of, for example, the cursor, instead of the center of the selection. Now that I know it exists, I can’t imagine modeling without it.

The knife tool can be used in conjunction with the Ctrl key to snap to verts. This has huge implications because you can use any piece of geometry to cut any other piece of geometry. Just snap to the verts of the geometry to be used as a guide, cut the target geometry, and delete later (by selecting linked [L]) if no longer needed - OR keep re-using it like a cookie cutter to make the same cuts repeatedly to different parts of your mesh! I don’t know if this is the best way to explain, but contact me if you need clarification.

Oh, and one more thing I almost forgot to mention: The “Falloff Editing” option ([O] key) is not to be underestimated. This tool is a godsend, and whoever programmed it deserves an award, or something. If you don’t know about it already, look it up. It’s rad.

Hope this helps.

I’m fairly new to blender and what I’ve gathered is it’s just experience and some anatomy knowledge when doing humans, I have been using reference images, and I can tell the more and more I do it the more I get a feeling for it, I could probably do a fairly good human from scratch, but I guess it’s just practice practice practice. This is a great tutorial I did; http://www.blendernation.com/2007/08/14/video-tutorial-modeling-a-lowpoly-character/