A round hole

(Dark Killer) #1

How would you make a round hole in a wall, the cylinder tool is a lil weird and I still can’t get it to work properly, so how would I go about making a nice round hole in a wall?

(Ubiquitous) #2

Two square brushes, then carve out crescents in each, should make a cylinder.

(Thej) #3

or just make a square brush, keep it selected, go to brush at top of the editor . choose arbitrary sides and choose a number of sides that suits u.

(weasel) #4

Cut a square hole, create an endcap, go to curve->cap selection, choose inverted endcap, then copy it and rotate it to fill the other side.

(damocles) #5

What weasel said. Or if you don’t want to use patches (because sometimes texturing patches can be a right pain in the arse) then your next best bet is to make a square hole. Now fill it with a squre detail brush. Duplicate the brush. Now select the filler brush and it’s copy and select region->set selected brushes. Now select only one and use the brush->cylinder->arbritary sided to create a sufficiently curvy brush. Now do a CSG subtract.

SHOCK HORROR! A CSG SUBTRACT!!! Despite what some people say, CSG is not the be all and end all of evil. If you use it wisely it can bve a good time saver. Just don’t use it for simple operations as it will often result in poorly constructed walls that do little for performance. But for something like this, as long as you’re careful, it’s a bloody useful too.

note: I suggested doing a region->set selected brushes because it ensures you never accidentally cut through another brush using CSG subtract. A lot of the time there are brushes that you hide or filter and forget they are there until it’s too late. ALWAYS set the region to the brushes you want to subtract with before subtracting or you’ll be sorry. I know I was…

note 2: Make sure the hole filling brush is square, because the brush->cylinder tool will only make perfectly square brushes into the correct size cylinder. Any non-squaqre brushes result in cylinders that are too big or too small for your hole (oooh matron)

(SCDS_reyalP) #6

Here is why using subtract as Damocles suggests is a bad idea:
http://www.csis.gvsu.edu/~bickelj/quakin/html/csgsubtract.htm
If you don’t care about z-fighting, overdraw, and generally sloppy brushwork, go ahead and subtract cylinders from squares.

The right way to make a round hole is to make square hole, and then make the round part out of patch meshes. If you insist on making it out of brushes, you can get better results in not much more time by correct use of clip, clone and 90 degree rotations.

(damocles) #7

Overdraw sure, but I’d be very interested to see any noticeable z-fighting from the suggestion I posted.

And the overdraw should be all but eliminated by the compile in q3map2.

And the brushwork would only be sloppy if you don’t first refine the hole so that only a square portion is being filled and nothing more (to reduce the lengths of the wedge sections that will be made).

You can tell that the person who wrote that anti-CSG article is biased against it by the last example on the page. A wise CSG user would have made the area to be subtracted from as tight a square around the hole as possible. Then CSG would have simply left four small triangles in the corners.

I don’t think I explained that too well, so take a look see - hole made using CSG:

(ChumChum) #8

Hi. My name is ChumChum… and I use CSG subtract.

(weasel) #9

damocles: You still have overdraw in your example. Not as much, but it’s still there. The way it’s done in the article is they best way (if you want an octagonal hole). If you want it to look like a cylindrical hole, patches work best. I hope the way radiant draws it isn’t the way it’s cut up in game, though. Ideally, it should use triangles for the caps, those trapezoids are wasteful.

(Dark Killer) #10

thx a lot i found that CSG stuff a lil complex and time consuming :cool:
so i took weasels idea i had a whole in my wall in about a minute :clap:

(SCDS_reyalP) #11

weasel, your caps will result in sparklies (because the verticies don’t match up with the brush verticies). It should look like this:

The brushwork around it also need to be cut so its verticies line up with the verticies of the patches.

Damo, here’s some z-fighting from the subtract method:

Obviously, q3map2 does not eliminate the overdraw in that situation. That method also uses significantly more triangles than the version shown by Hr. O

(Chis) #12

signifgantly more triangles? damocles’ and Hr.O’s method both use 12 tiangles… Q3map2 deals with overdraw? in what situations if not this?

(SCDS_reyalP) #13

You are not counting T-junctions.

Q3map2 deals with overdraw? in what situations if not this?

If I under stand right, it will eliminate faces which are completely covered. In the above damocles’ example, they are not.

(damocles) #14

I may be wrong about this, but I alwaqys thought that q3map2 did T-Junction checking along the edges of brushes where they meet with others. And in the CSG subtract methos for the hole, the edges of the “flat” end of each wedge shaped brush is against the slope edge of another, meaning that the T-Junctions should split the faces that are visible. And then the ones that are not visible (because they are pressed againt another should have been removed).

Like I said, I could be wrong about that, and it may be that q3map2 cannot identify all the situations like that, which would result int he overdraw seen in Reyalp’s image.

(SCDS_reyalP) #15

Those brushes were created using your method and compiled with q3map2. Z-fighting is clearly visable.

The T-junction comment was about the number of triangles, and has nothing to do with the z-fighting and overdraw, it is just another disadvantage of that approach.

All this may seem nitpicking and pedantic. Everyone is free to map however they want, and excellent maps can be made with less than optimal construction. I hope that pointing out these issues at least gives people the option of avioding them.