Messin' with Mhaze's Rock Study

Started by RichTwo, August 21, 2019, 11:49:18 am

Previous topic - Next topic

RichTwo

I have been tinkering mhaze's Rock study .tgd, not really straying too far from the original, other than color and fake stones.  Also used Jordan's Basic Stars background (doubled the luminosity factor), lowered the blue atmo density and tossed a planet out there.  I first rendered with Path Tracing off it whipped out in just over 30 minutes.  This was done with PT on and it was over 11 hours in the render.  That's over 20X the time.  Wow.  But is it worth it?  Yes, but I'd sure like to see become more reasonable, like 2 - 3 hours instead.  Oh - almost forgot: I used Matt's render group with no changes other than PT on.
You don't need a weatherman to know which way the wind blows...

WAS

Nice Rich, I like the colour tones of the scene. Looks arid, Rocky, and perhaps a bit dusty. .

That PT time does seem a bit unreasonable. Is there any reflections?
Check out my Terragen Discord: https://discord.gg/Vy5FRTE

RichTwo

Quote from: WAS on August 21, 2019, 12:28:56 pmThat PT time does seem a bit unreasonable. Is there any reflections?
None at all. Anti-aliasing and micropoly set at "reasonable" levels, too.
You don't need a weatherman to know which way the wind blows...

WAS

Quote from: Rich2 on August 21, 2019, 01:21:26 pm
Quote from: WAS on August 21, 2019, 12:28:56 pmThat PT time does seem a bit unreasonable. Is there any reflections?
None at all. Anti-aliasing and micropoly set at "reasonable" levels, too.

Hmm. I wonder if because the background node is using shaders, and in a whole lot of darkness, with luminosity, it's tripping off a lot of extra computation with PT.

This definitely doesn't seem like normal behavior, and I don't think this has been tested with PT yet (background node textured with lots of darkness and light).
Check out my Terragen Discord: https://discord.gg/Vy5FRTE

RichTwo

Hmm. I wonder if because the background node is using shaders, and in a whole lot of darkness, with luminosity, it's tripping off a lot of extra computation with PT.This definitely doesn't seem like normal behavior, and I don't think this has been tested with PT yet (background node textured with lots of darkness and light).

Actually I started out doing full-size test render without that background node and was getting 8 - 10 hours every time.  Maybe I'll try PT on the start up as is and see what happens.  Thanks for your concern (or curiosity at least)!
You don't need a weatherman to know which way the wind blows...

Dune

PT is really only very good for landscapes with lots of foliage, so I wonder if it's wise to do a barren (nice) scene like this in PT. Why not do a standard render and compare?

mhaze

Nice, excellent moody, barren terrain.

WAS

Quote from: Dune on August 22, 2019, 01:43:05 amPT is really only very good for landscapes with lots of foliage, so I wonder if it's wise to do a barren (nice) scene like this in PT. Why not do a standard render and compare?
PT is good for everything, and this needs to stop being said.

It effects both highlights of surfaces, and shadows of surfaces, and additionally it's exaggeration feature is amazing for surfaces, while not maybe as realistic unless applied for surfaces only to help bring out shader work. Great for compositing.

Yes it makes vegetation look great with layering, and soft shadows, but it works to a more realistic scene. I can see what PT is doing in his scene plain as day. Just wish it was faster.
Check out my Terragen Discord: https://discord.gg/Vy5FRTE

RichTwo

Thanks for the input, all!  I did a standard render and there was quite a notable difference.  See for yourself...

This was a 30 minute render.
You don't need a weatherman to know which way the wind blows...

luvsmuzik

I love them both.....what was your multiplier on PT version?

RichTwo

Quote from: luvsmuzik on August 22, 2019, 03:55:21 pmI love them both.....what was your multiplier on PT version?
Thanks - 

Multiplier? Respectfully, I don't know to what you are referring.  If it's micropoly 0.6 /AA 6.  Pretty much middle-of-the-road.
You don't need a weatherman to know which way the wind blows...

DocCharly65

To be honest, I must say that I like the standard render a little bit more... I like that glowing brighter look. But al versions have their Charme. Nice idea and good work!

Dune

The shadows are deeper in the PT version, but the reason why I said about PT versus standard is that not everybody has a fast machine, and sometimes you have to make a time/quality decision. Of course PT is more natural, but using it just because it's possible, may put certain users off due to the lengthy renders. And we don't want that!

Btw, with less GI the shadowed rockwall would probably also darken a bit more in a standard render.

DocCharly65

I absolutely agree, Ulco. The speed is of course also depending on the scene. Many reflections can be the death of the render speed.
These days I gave up giving PT my first try... I first rendered a helicopter scene in the hangar (of course with all those reflective objects) within 20 minutes with standard render. I thought that must be an ideal test frame for PT... after 1 h and only the first row of rendered buckets I stopped that experiment ;D

(I tried both: robust sampler and defer all shading - without these two settings there wasn't even 20% of one bucket rendered ;D)

WAS

Quote from: Dune on August 23, 2019, 01:56:55 amThe shadows are deeper in the PT version, but the reason why I said about PT versus standard is that not everybody has a fast machine, and sometimes you have to make a time/quality decision. Of course PT is more natural, but using it just because it's possible, may put certain users off due to the lengthy renders. And we don't want that!

Btw, with less GI the shadowed rockwall would probably also darken a bit more in a standard render.

Longer render times are inherent to PT though compared to Raytracing for very obvious reasons though. I do feel there shouldn't be confusion there and a user informed on renderering engines. We shouldn't just assume people are completely incompetent. PT provides these more accurate results because of brute Force path tracing. Average "smooth" or noiseless PT render has upwards 100,000+ samples per pixel.

There does seem to be some contrasting differences between renderers too. This could be because the path tracer is taking into consideration subtle shadows more precisely, natural reflection of surfaces can reflect darker shadowy areas I'd imagine too when the terrain starts directing away from the light and towards shadows.
Check out my Terragen Discord: https://discord.gg/Vy5FRTE