What thing(s) have you found effected displacements in your map that you didn't expect?

[Werewolf]

If you had seen my previous thread, you'll know that I have been having trouble with displacements not being rendered when they should. Eventually I gave up and went to bed. Now having slept on it, I began thinking.

While it is capable of a great many things, we know that the hammer editor can also be buggy pile of crap at times. So I am wondering if the troubles I am having with displacements are due to something not mentioned in the various instructional guides we have here and elsewhere on the internet. Either something no one has noticed cause a problem before, or just one of hammer bugs that happens from time to time.

For example a few of the displacements I am having trouble with are in the skybox, which has a func_viscluster in it. Another area of the map with displacements has a lot of hint/skip brushes around it. There is no reason to assume either of those things would cause a displacement not to render, but through some hammer magic-fuck-up-ery they still cause a problem.

QUESTION: What thing(s) have you found effected displacements in your map that you didn't expect?

 

[killohurtz]

1. Displacements may not retain their correct shape in instances. I don't know the specifics, but my issue involved tunnels at 45 degree angles. Collapsing the instances before compiling fixed it.

2. IIRC a map called pl_displacement was named after displacement rendering issues. I can't get a post number on mobile so look through this thread. Maybe its author has information that could help you.

 

[Sergis]

1. Displacements may not retain their correct shape in instances. I don't know the specifics, but my issue involved tunnels at 45 degree angles. Collapsing the instances before compiling fixed it.

as far as i understood during my instance displacement troubles the displacement direction ignores the instance rotation

so if you displace a vertex 10 units along x axis the vertex will be displaced 10 units along main map x axis no matter the instance rotation

very unhappifying

displacing them along the instance rotation axis should work fine i guess

 

[Fruity Snacks]

Displacements can work in instances, if you don't vertex manipulate them too much. (Most of snowplow early on, including the entirety of the final point was an instance, displacements included).

Displacements must always be a quadrilateral. Not too many people know this, I think, but they must always have 4 sides. Anything more or less and they break or vanish.

Weird VVIS cuts can cause displacements to go invisible.

 

[Werewolf]

Weird VVIS cuts can cause displacements to go invisible.

What sort of weird cuts? Like around where a something that should be func_detail'd but hasen't? Of course I don't mean the displacement itself, just some geometry around it that should ideally be a func_detail.

 

[Fruity Snacks]

I don't know of the specific cases, but anything that causes a lot of small cuts in a wide area. Usually a cluster of microcuts or something can muck with things.

bad areaportaling too, forgot about that.

 

[Idolon]

I don't know of the specific cases, but anything that causes a lot of small cuts in a wide area. Usually a cluster of microcuts or something can muck with things.

I've had a few times where a bunch of small cuts in one area could actually make small portions of world brushes appear invisible - I have a hunch that, if this happens anywhere in a displacement, the displacement goes invisible.

 

[I dinne ken]

Sewing a displacement with a blend texture's usual beta texture as the alpha to a normal brush will make the original alpha texture on the displacement have an outline on the border between the brush and the displacement

1.

2.

 

[UKCS-Alias]

1. Displacements may not retain their correct shape in instances. I don't know the specifics, but my issue involved tunnels at 45 degree angles. Collapsing the instances before compiling fixed it.

Any XYZ modifications you made are applied as if the displacement was not in an instance. So if you moved something on the Y axis and your map was rotated 90 degrees on the Z axis, your adjustment moves into the X axis in the instance itself.
As image (the red line is the adjustment you made on the displacement and the dot its original position):

(the A and B are reference points)

And on collapsing it handles the displacement correctly by giving it the correct angle. This goes wrong in vbsp.

 

[Crowbar]

Sewing disps also sews alpha well (@normanconqueror, that's what causes what you showed in your post).

 

[Gale]

If you subdivide the sides of a cube to make a cylinder, and change the scale to 2 it will make a nice looking pillar! I've deleted the tops and the bottoms so you can see the shape better on the grid:

 

[Pocket]

"Affected." And... ooh! I got one! It's only sort of a displacement thing, but blend textures don't work underwater. Check it:

It's really weird and I don't know if anyone ever talks about it.

 

[Blade x64]

Dislacement blends still work, the issue is that blendmodulates do not when viewed through the water shader/vis system.

 

[Pocket]

That's what I meant by "blend textures"; I keep forgetting the proper name.

 

[henke37]

That's because you mix up textures with materials.

 

[Pocket]

No, it's a texture. It's a red-and-green texture that determines how the other textures blend. If I had said that blend materials don't work, that would have been wrong, as Blade pointed out.

 

[Skittelz]

Off-grid displacements can't sew with other off-grid displacements. BUT if the displacement is half sized (and on one edge of the other displacement) it will sew together. For example:

Before and After (off-grid)

Before (equivalent)

After (equivalent)

Works with halves

 

[Idolon]

snip

That's just how displacements work. They have to share an edge (or exactly half an edge) to able to sew to each other. The grid has nothing to do with it.