Some canyon with water :)

[Tangled-Universe]

I've tried rendering it with GI surface detail enabled, no effect.
Also disabled populations, rendered then with and without surface detail enabled, also no effect.

I also rendered it with a white lambert shader covering everything and the crop became completely white. As supposed I guess? What's the thought behind that?

[JimB]

If you have a reflection shader somewhere in the sandstone's stack, I'll bet you the dark patches are a reflection of the dark water below, and the overall brighter canyon is reflecting the thicker haze.

[Tangled-Universe]

The same dark patches are in the first render at the top of the thread, just not nearly as pronounced due to a thinner atmosphere. Is it something to do with the gamma and atmosphere? Is it the reflection/specularity being more pronounced due to a thicker haze?

If you have a reflection shader somewhere in the sandstone's stack, I'll bet you the dark patches are a reflection of the dark water below, and the overall brighter canyon is reflecting the thicker haze.

Thank you Jim. The surfacing of the terrain contains no specularity or reflectivity shader.
Disabling the waterplane didn't help either. So it's not one of them or the combination of.
By the way, the shoreline is also the edge of the waterplane. I did this on purpose to save lots of rendertime on ineffecient backface-culling/clipping.

As for the gamma; I kept it untouched. I've disabled the cloud layer while testing Matt's suggestions and that also didn't make a difference.
Reducing the thickness of the haze (decreasing haze density value) didn't help either. The latest haze density was 1 @ 8000m heighth...nothing spectacular. Original density was 0.5 (first image in thread).

[Tangled-Universe]

A little bit closer to final realisation... I've reworked the surfaces and I still don't really like them, somehow they worked out differently from my test-renders, maybe because of the low cloud which I disable during testing. One of the things I've been playing with again is the subsurfacing of the water. I attached a warp shader to the density function of the subsurface shader to get the idea of streaming muddy water and I think the effect turned out pretty well
But still enough to do...

Rendered @ 960x720, detail 0.9, AA 14 and GI 1/1.
Rendertime 10,5 hours.

Still problems with the black patches.
Matt asked me to attach a white lambert shader and see if the problem persisted and it didn't.
So I thought I'd start with a brownish lambert shader as the very base layer of surfacing, but the black patches just appeared again

Matt I hope you have the time and willing to help me sort this out. What was the reason for attaching the lambert shader / what problem do you think it is?
I can send you the tgd if you'd like.
Thanks in advance.

Martin

[Oshyan]

The water looks too transparent without much apparent "tint" near the top-right of the image, but the "flowing" effect and sense of scale are quite nice.

I think Matt wanted to see if a completely white surface still showed the errors. If so they'd more likely be something other than surface-related - e.g. lighting (GI), atmosphere, etc. Since it *did* go away however it is more likely to be related to the surface mapping.

- Oshyan

[JimB]

Martin, have you tried inserting a surface shader between the Base Layer and the next shader on in the list, with zero displacement? If you made it bright blue, it might give more of a clue (for instance, perhaps there isn't 100% coverage in the other surfaces).

[Oshyan]

Good idea. Or just make the Base Layer a bright color.

- Oshyan

[Tangled-Universe]

I'm 99% sure I've tried that already, but to get that extra percent I'll try it again
The lambert shader I've inserted should already do the trick, isn't it?

Martin

[JimB]

I'd use a constant shader to eliminate any possibility of shadows and GI confusing the issue.

[Tangled-Universe]

I'll try that inmediately when I'm back home from work (9 am here so it will take a while).
Thanks for the good suggestion.

Martin

[Matt]

The reason for the Lambert is to see what the same scene looks like when there is only lighting in play, with no surface variations. As the problem went away, it's likely to be a surfacing problem somewhere, but it could still be something to do with low quality GI or something - it's hard to be sure. The next thing I would try is inserting a test shader (a bright red constant shader perhaps) at various places in the chain and you should be able to narrow down exactly where in the shader chain it's happening.

Matt

[Tangled-Universe]

The reason for the Lambert is to see what the same scene looks like when there is only lighting in play, with no surface variations. As the problem went away, it's likely to be a surfacing problem somewhere, but it could still be something to do with low quality GI or something - it's hard to be sure. The next thing I would try is inserting a test shader (a bright red constant shader perhaps) at various places in the chain and you should be able to narrow down exactly where in the shader chain it's happening.

Matt

Thanks for the explanation and suggestion Matt. I'll test them tonight and keep you posted
First it's time for the beach, 27 degrees celsius over here

Martin

[Tangled-Universe]

Hi Matt,

I've tried your suggestion but I can't pinpoint the problem as I get very different results when I move the constant shader along the chain.
If you have time, would you like to take a look at the tgd for me?
Thanks in advance.

Martin

[Matt]

You will get different results when you move the shader along the chain, but you need to do this to find out which shader introduces the black parts while the others merely introduce colours that you want.

Matt

[Tangled-Universe]

Yes Matt, I was a bit unconclusive I'm afraid.
As I said the results were very different, but in the end where-ever I put the node in, the black parts remain.
I'll send you the .tgd if you don't mind. I'm currently working on an other project so I'm not in a real hurry

Martin