Cutting a hole into a model: Discussion

[ub3rman123]

I've found that if you're working on a flat surface and need a hole cut in it more rapidly, then you can instead delete the flat face, then make the outline of the hole you need, then make new faces around the hole.

You get pretty much the same thing, but not nearly as much merging to get rid of unnecessary verts.

Also: If your wire section of the hole (The circle bit sitting in the middle) isn't aligned quite right to the face you want the hole in, you can select the circle AND the corners of the surface and shrink them along the Z axis (S and hold down Z) to have them all at the same height, to avoid any uneven faces.

Edited June 25, 2011 by ub3rman123

[InsanitySorrow]

Only problem with that way is that it's not a very clean method, it'll work but might cause problems if you decided to create a high poly model later on

[ub3rman123]

Only problem with that way is that it's not a very clean method, it'll work but might cause problems if you decided to create a high poly model later on

You mean when running Subsurf on it?

[InsanitySorrow]

Using subsurf on that one will most likely ruin the edges and create the look you aren't looking for

[Artisanix]

There are two ways we can do this, the first is to use the subdivide option, depending on the edges you choose this could potentially add vertices all around the model which isn't very clean and will bump up the poly count quite a bit.

So what we're going to do is use loop cuts instead, these are cleaner and will allow us more control on where the new vertices will go.

Your tutorial is quite informative and detailed but... there's a third way to do such things :] And I mean boolean operations here...

And here's how to do this (all actions in object mode):

1) create a cube, rescale it up (for instance) two times

2) create a cylinder with 24 vertices in the center of that cube and rescale it up along Z axis (for instance) three times

3) select the cube, and with SHIFT pressed down select that cylinder

4) menu: Object -> Boolean operation... -> Difference

5) select your cube and cylinder and hide or delete them and you will be left with cube that has a hole in it.

two objects - cube and cylinder before boolean operation:

cube with a hole after boolean operation (difference) - original cube and cylinder had been deleted:

Well, your way of doing such things is interesting and give a chance to learn a few techniques that may be useful in some future for someone,

but boolean operations in this case are much simpler, much faster, and more flexible (by using extrude for that other object you can easily create holes with various shapes).

Besides the cube with hole produced by boolean has 56/56 vertices/faces after conversion to triangles, while the cube made with instructions in your tutorial has 112/112 faces/verticles (even before conversion?) from what I can see in your pictures - and that's because all sides of the cube have unnecessary faces me thinks ;]

[InsanitySorrow]

Yeah I'm aware of that method and have used it myself before, it's something I refuse to teach others as it's not a very good method to use.

Have you tried adding more edge loops?, Bevelling the edges? or adding a subsurf modifier to it?, You'll most likely encounter the same problems as ub3rman123 above. There is no doubt it's a quicker method that has a slightly less poly count, but it's not a clean method (Which is what this tutorial is about) and will likely cause a lot of problems if you try to work with the model more.

My cube does have quite a few extra faces, but then it's a tutorial and I did warn about that .

[Artisanix]

Yeah I'm aware of that method and have used it myself before, it's something I refuse to teach others as it's not a very good method to use.

Let's explain something here ;] There's no one and only and ultimate method to do certain things in modeling and which is flexible enough to cover all the variants from simplest to the most complex. So it's usually good and useful to know more than one way to do things. Because one method may be far better suited for simple modeling while another for more advanced objects. And while that another method is supposed to be good for complex meshes and in the same time it is quite complex by itself then it may be not the best solution for simpler things ;] So do not deny existence of boolean method as a teacher ;] because for particular things it is far better than edge loops, etc. while for more advanced stuff your method may be the only reliable and decent solution...

[InsanitySorrow]

I'm not going to start arguing points here, that'll get us nowhere and is completely pointless. But I will say this:

I do agree that there are multiple ways of doing things, that's one of the great things about modeling. I don't agree with the rest though, yes booleans are quick and easy to use, but they are not very good and are best left alone unless you don't plan on altering your mesh after using it.

I'd rather teach a good, clean and effective way of doing things that won't cause problems later on if you decided to do additional work on your model, even if it means it takes a little longer to do and the method is a little more complex.

If you don't agree that's fine, you're more than welcome to share your own preferred methods of modeling, I'm sure others would find it useful , But as I said this is my chosen way of doing things and this is what I want to teach others

[auryga]

I've used the boolean method first, and then something like ub3rman123 method, but never something like IS way. And I had several of the problems already mentioned by IS, beveling especially, and also see some advantages for my way of texturing here.

Another great example why I always look into all tutorials, even if they seem to cover such an "easy" task.

Thanks a lot, IS, and TrollF and ub3rman123 too, for the additional aspects added. IS comments explained the "backgrounds" much better to me, hadn't seen some of the benefits before

[InsanitySorrow]

You're welcome

[Nekronom]

If I may pitch in my twenty five dollars, I would like to say, That while there is nothing wrong with IS's way of modeling this hole, it does indeed waste precious verts that if left un-optimized could lay siege to the CPU; However, given that it is placed in a cube, It would in fact be optimal to do it his way if you are using sub-D modeling methods. The optimized tri corners are a completely acceptable method however, given that correctly layed smoothing groups would make them imperceptible to the player. So in conclusion, both methods are acceptable, and both methods are correct, and while both methods are preferred in a specialized area, they suck in the other, tris don't sub-d well, and lots of wasted quads are well...wasted quads. However, I feel I'm ranting to completely competent people who already know what i'm sayings, so I will be quiet now.

[InsanitySorrow]

You can all discuss as much as you like here now, it's no longer going to distract those using the tutorial.

It's amazing how people seemed to miss my mention of it being only one way of doing this

[ub3rman123]

What method do you use to make it more high-poly? It might prove useful for a few meshes I have.

[InsanitySorrow]

What method do you use to make it more high-poly? It might prove useful for a few meshes I have.

I apply a subsurf modifer to the model, usually a 2 level subsurf. I then use edge loops to define the edges and overall shape, without them the model would collapse since it wouldn't have the support. To add additional detail I'll use more edge loops and carefully subdivide parts too and model from there.