Smooth patchcap and brush transitions - how?

[Clueless Newbie]

When several brushes lie on the same surface, I get a smooth surface where I can't tell one brush from the other. But when when some of those brushes are actually patch caps, I get some very visible borders. The shadings just don't look right.
For example, I have a circular hole in a floor. The brushes create a square hole, and the patch caps "round out the corners". But the patch caps are very visible. WHat should have been a circular hole in the floor is instead a square plate with a circular hole in it.

Is there a way to fix this so the transitions are smooth and non-obvious?

[TVR]

It would be of benefit if the compile switches used were listed; the situation as it stands seems remarkably similar to a standard vis/lighting crack common with patch mesh.

[Clueless Newbie]

It would be of benefit if the compile switches used were listed; the situation as it stands seems remarkably similar to a standard vis/lighting crack common with patch mesh.

VIS:
-fast

LIGHT:
-deluxemap
-fast
-patchshadows
-bounce 8
-samples 8

[morfar]

-patchmeta fixes patch holes.
But you'll get z-fighting on patches when realtime lights is nearby.

[Clueless Newbie]

-patchmeta fixes patch holes.
But you'll get z-fighting on patches when realtime lights is nearby.

I'm using -patchmeta in the BSP phase. The problem isn't nonsolid or unshaded patches, but patches with wrong shading at the edges.

[morfar]

ok, I didn't read the first post.

[torus]

Wrong shading as in, an incorrect lightmap? Or are you unable to correctly position the texture of the patch so it is a smooth transition?

[tZork]

i generaly avoid trying to match patches with brushes. in thise case i would proly try to make the walls themselfs from patches too, and just have a caulkd visblocker inside.

[Clueless Newbie]

i generaly avoid trying to match patches with brushes. in thise case i would proly try to make the walls themselfs from patches too, and just have a caulkd visblocker inside.

Well, the patches don't match with each other either. This would just give me a different kind of uneven transition. B/

[TVR]

The visibility -fast switch should be relegated to rough, test compilations solely to satisfy the prerequisite of the light calculation phase.

Removal of said switch is probable to resolve such issue.