Fake stone merge issues fixed?

[sboerner]

I've struggled for a long time to get fake stones to play nicely with distribution and merge shaders. Suddenly it seems with the latest release version that these issues may be fixed. Did I miss something?

Same file rendered in 4.5.71 and 4.6.31.

[Dune]

That's a weird difference. Can you post the tgd?

[sboerner]

Well, this is baffling. I just reopened the file to make a simplified version to post and . . . it now appears to be working fine in the old and new versions of TG. 

Earlier this week I spent an hour trying to get that network to function properly in the older version of TG. When I realized it was working fine in the latest version, I switched back to that. That's when I made the two test files, opening the same file in both versions and making the renderings I posted here. The older-version rendering shows what I had been dealing with the day before.

Maybe something I did, but I don't see how that would explain the difference between the two back-to-back renderings yesterday.

Gremlins? 

[sboerner]

Hold on. I just found the scene file I used yesterday and the issue is still there.

Can't share the entire scene (it's for a client). But I'll do a teardown later today to isolate the problem and will post a simplified version if I can.

Whew. Thought I was going crazy there . . .

[sboerner]

OK, it was partly operator error.

The network uses two merged fs shaders. I inadvertently set the density variation scale of the smaller size stones to zero. (Probably meant to set vary density to zero.) This resulted in some large gaps where the smaller stones were missing.

However, it appears that version 4.6.31 of Terragen handles this setting differently than earlier versions. So when I use it to render the scene, the smaller fake stone layer appeared to be fine even though it really wasn't. (This why I'm only taking part of the blame ).

When density variation scale is set to zero, there are serious mismatches between fake stone displacement and color in both 4.5.71 and 4.6.31. But the displacements are inverted.

Lesson learned: Pay very close attention to the density variation settings. (I'll be disabling this feature entirely for this scene.)

[Dune]

Interesting. I actually never touched those settings (never needed them), but now found some interesting behavior when changing the density variation and scale parameters. Good to know, so thanks for your 'problem'   Indeed unchecking vary density works the same as setting scale to zero.
But I don't see any differences between the renders in 4.6.07, 4.6.11 and 4.6.23 ( I don't have 4.6.31 installed). Only a slight change in shadow darkness (4.6.23 has a bit darker shadows). Easy to see when you see the previews after eachother in an explorer window.

[sboerner]

Yes, this was eye opening for me too. I wonder how many of my earlier problems with fake stones were caused by changing these settings (intentionally or not).

As you say, unticking the "Vary density" checkbox disables this feature. If you want to vary a fs distribution I'd consider doing that and using a pf mask instead.