Working on a canyon scene I decided to ad some fake stones and masked water. Did some cropped renders to try different setting/values and all of those renders took little over 5 minutes to complete. After I was satisfied with the result of the cropped render I started a render the whole scene. Before the water and fake stones the whole scene took 38 minutes, but now after 1h 45 minutes the render still was not finished so I did abort the render. Thinking the fact I did crop the partial render could be of some influence I now did crop the whole scene and started a new render. Alas, after 1h 45 minutes the scene was still rendering. It took 1h 58 m to complete. Now I wonder what's causing this?
If you have water material applied to surfaces which will case a lot of interreflection then this can slow down renders a lot. In that case you may be able to get a good enough effect without "true reflections", so you could replace the Water Shader with a Reflective Shader and make sure Raytraced Reflections was disabled.
If that's not the issue then I'd probably have to see the TGD to have a better idea of what could be causing it.
- Oshyan
I get that... But why is the same part with water and all rendering in 5 minutes on one occasion and does it take over an hour to render that particular part on the other occasion? Why do the same square inches with the same materials and the same settings have such different times to complete? It's not the fact the scene with and without water have different render times that I'm wondering about. Water has always been a time consuming addition to a scene, already in the days of "classic TG".
Oh, I see. I guess I didn't quite understand your original question.
Well, it's possible that in the crop it is not accounting for some of the information/details outside of the crop area, and those "missed" areas are accounted for in the full render and are for some reason demanding to render in reflections. One question then would be whether the water looks different in the crop vs. full render.
- Oshyan
Ahhhahhh, so the fact terrain information from outside the cropped part is not taken into account on a cropped render is causing this seemingly discrepancy.... Hmmm, now I will have to do a cropped and a full render with higher quality and size settings to check if there is a blatant difference... If not I will edit the partial into the full but not finished render, quick and dirty. ;D
Well, keep in mind that's just a theory on my part. It may not be correct or accurate, but it's worth checking to see if there is a visual difference, if possible.
- Oshyan
Sure Í understand that this all is based on assumption. But there should always be room for testing, especially if one thinks the results justify the time spend/spent?
I will do a full sized render without water and a cropped render with water and combine them. Reflections will be the key: Reflections of the surrounding environment in the water and reflections of the light bounced by the water, if you know what I mean.
That thumbnail is great looking!! When I click on it of course it is a lower quality test, but I can clearly see the awesomeness that awaits a finished render!
Well, that awesomeness is not entirely my doing, in fact most of the original TGD is beyond my comprehension, even after dissecting it for almost a week now. :D The original TGD is created by a guy called René van Megen and he did share it through the FB Terragen group. I simply pushed some sliders, manipulated values and made some additions.
So why is René van Megen not here, but on #*~^ facebook?
René thinks he has a too bad english for this forum.
And it is a pity. We could have to learn so much from him.
David
I think an intervention is called for heh heh...we need his added imagination here...
Well, we also have Bastien, don't we ;) There's always Mr. Google.
One of my reasons to participate here was to learn to understand and write better in English.
Not sure how much it worked but yeah that should not stop anyone :)
True Ulco ;D
Hopefully for me, Bastien is french ::)
David
Quote from: Kadri on September 20, 2015, 03:46:25 AM
One of my reasons to participate here was to learn to understand and write better in English.
Not sure how much it worked but yeah that should not stop anyone :)
You are completely intelligible to me Kadri...well done.
Quote from: bobbystahr on September 20, 2015, 09:26:19 AM
Quote from: Kadri on September 20, 2015, 03:46:25 AM
One of my reasons to participate here was to learn to understand and write better in English.
Not sure how much it worked but yeah that should not stop anyone :)
You are completely intelligible to me Kadri...well done.
:)
The fact that you are a non-native English speaker seldom - if ever - crosses my mind Kadri. :)
- Oshyan
Quote from: Oshyan on September 20, 2015, 09:05:22 PM
The fact that you are a non-native English speaker seldom - if ever - crosses my mind Kadri. :)
- Oshyan
Thanks Oshyan :)
Yeah, your English is pretty good. I've heard native speakers that are less intelligible ???
On the main topic, is there water being viewed at grazing angles (as in looking towards the horizon)? This has always caused HUGE slowdowns for me, so much so that I either avoid that kind of shot or I use 'defer atmosphere' (which seems to help).
Not sure what is meant by "grazing angles" so here's the merge of the full render (without water) and the partial render with water so you can see for yourself. And judging from some of the other versions of this project the extra amount of bouncing light introduced by the water is indeed causing the extreme long render time. In this case the corner with the water in it took almost 6 hour to render and though the .GIC used did NOT contain any water there is still a huge difference in the way the vegetation is lit in that partial render. Oh, the full render without water was completed in little over 5 hours.
A render like this shouldn't take long at all! Maybe 30mins on an i7. Did you mask out the area where water is not visible? That would save a little render time. BTW, terrific rocks.
QuoteDid you mask out the area where water is not visible?
Yup, the water is masked the dandelO way :) And the CPU is a i5 :(
Yeah, but did that work alright? Have you checked? In this case a simple shape or painted area where the water is will also do, painted from top down. Also sorting bias of the lake at -1000000 will help.
The mask for the water is painted; the plane used is 110x280, just enough to cover the area. A version with the traditional lake at the smallest required size didn't change a thing , abandoned that method after 25 minutes and even less water rendered.
So that's probably not why it took so long. Just curious what did that.
Just wondering,
Looking at the rock structures, can these shapes be converted to geometry and re-textured? or is it possible to "bake" to vector displacement?
With all the undercuts baking to a heightfield isn't an option but I think if these shapes didn't require calculating a procedural shader the whole scene would render a lot faster.