Path tracing an old scene Fields

Started by Hetzen, October 19, 2018, 06:55:12 AM

Previous topic - Next topic

Hetzen

As in the title really. Martin suggested I did a re-render to see the difference. Quite interesting results.

With path tracing, texture, displacement and objects seem to be smoother in this scene. I used a rock population to mimic hedge row trees, which look more like green rocks with PT. The field walls also crack over distance, which you get away with in the standard render.

The white texture on the fallen rocks was a mistake in the shader tree. I left it in to keep the comparison consistent.

The last image was about getting some v3 clouds in and trying to capture golden hour. I think the field textures need some tweeking in this light.

Tangled-Universe

Hey Jon!
Awesome to see you back here and do some rendering :)

The golden hour render is almost exactly as I imagined it when I made the suggestion :)
Where it "breaks" for this render is the tree model and perhaps you need a model for the rock wall as well.
Else it looks pretty damn nice. Right!?

I suppose this is one of the cases where the standard renderer is doing pretty well compared to the path tracer, but I will recheck tonight at home, since the monitors here at work are super-contrasty and shitty.

Keep rendering mate! :)

Martin

Oshyan

To get detail parity in any terrain displacement you need to increase Ray Detail Multiplier to 1. That accounts for some of the differences (primarily loss of detail in some areas) in the path traced version.

Cool to see this comparison in any case. And I love that you still had this scene laying around. :D

- Oshyan

WAS

Wow, what a difference that makes in lighting. it is interesting to see such a big difference in the distant terrain. As Oshyan said, the ray detail multiplier being increased to 1 may help with that. Though not sure I've seen such a obvious difference.

It is interesting how the field detail fades off, though is that actually more realistic to the human eye? My eyes are kinda bad, so this effect may just be that, but with the Tulip festival here in the valley, the fields cascading into the distance seem to blend smoothly.

Good to see you around again too! I just went back to your cobble function and did some fun with it.

Dune

Good to see your scene in PT now, Jon. I love the light in 40_03!

DocCharly65

I like the last version most, too  :)

I'm not sure yet what to think about PT from what I have seen in the forum.
Seems that I am a fan of softer and a bit more washed-out images. PT seems to increase contrasts.
Especially in the first render I am missing the surface details of the standard version.

Hetzen

Thanks guys.

Last render on this one, I switched up the Ray Detail modifier to 1 as Oshyan suggested, which made a big difference....

bobbystahr

Quote from: Hetzen on October 21, 2018, 07:01:04 AM
Thanks guys.

Last render on this one, I switched up the Ray Detail modifier to 1 as Oshyan suggested, which made a big difference....

Nicer for sure..sure wish I could afford to up my maintenance as PT looks really cool.
something borrowed,
something Blue.
Ring out the Old.
Bring in the New
Bobby Stahr, Paracosmologist

luvsmuzik

WOW!
Dig some more out.  Nice to see a great example like this. :)
Curious as to settings on close foliage, are you using translucency and specular at all? Very realistic lighting however done!
Fully appreciate the plotting of populations in this, nicely done.

Hetzen

Thank you.

Yes there's translucency on the leaves, but I think the main idea with the lighting was to get the hole in the wall to let light through, and get some dappled light on the tree bark. The golden hour version is all very nice, but this version is what I wanted to make, you're on a Sunday walk with the dogs in Gloucestershire. I even prefer the v2 clouds.

Hey WAS, forgot to comment earlier on the cobble function. Yes, I've enjoyed seeing you use it. You're right, it is two lanes of half sine waves offset. You'd have to sample the function several times to get the cobbles to overlap those. What you could do is use a Voronoi cell function to determine each grids grout thickness which will intern shrink the stone randomly, but I'm not sure that's a real solution.