Rendertime depends on Imagesize ?

hehe, what a question ? ;-), isn`t it ?

well, i have an image (my tentacle szene from “upload image galery”) which renders at 1000*xxxx quite fast. If i do switch resolution to 1200*xxxx (keeping proportions) it does render about 20 times longer. Its in the pre-pass period. I think, when it is through that period and it starts rendering the image, there wont be so much difference in rendertime per bucket…but, this seems to be a little bug…

Someone got an idea ?

Regards,

Thomaskl

what version are you using?

well, the latest build. 1.47.03. on Max7.5

that is kinda odd. unless at the 1000 resolution your using presaved IR maps??

ALso when you up the resolution of the render you can also lower the IR min max levels.
For instance if your render @ 640x480 with ir map of -3-1 you could render 1280x1023 with ir map of say -4-2 or even less perhaps and still get teh same details.

no, i dont use presaved IR-Maps…
Well, the GI Solution in this special image isnt so important for the overall look.
It`s a good tip to lower the IR Min/Max Levels…But. I am using preset Settings in LOW-Mode. I think it should resolve the resolution in quite the same time. The 1000 Render took 9h. and the 1200 Render hasnt finished Pre-Pass Period after 9h…This made me think…

Thank you so far,

Regards,
Thomas

yeah that is certainly odd.

I usually find rendertimes to be fairly “as expected” i.e. 15-20mins for a 640x480 render and around an hour 15mins or so. Which feels about right to me.. i did lower the IR min max values tho.

Well give whats in your scene maybe its hit a memory limit or something and is using dynamic memory now and hence the slowness (clutching at straws here) whats the memory usage on the larger render?

Maybe the slightly bigger frame size forces windows to swap a lot of data cause it can’t fit everything into ram. Swapping stuff makes things very slow. Have you watched your memory? At what level is it? Does your hdd works a lot when rendering the larger image?

If you are using 3d displacement, make sure that you increase your Edge Length as you increase the resolution. Since the Edge Length is tied to the pixel size, larger resolution causes many more smaller triangles to be generated, which may cause the dynamic memory cache to overflow.

Best regards,
Vlado

Thanks four ideas, guys,

well, Vlado was right. There was the problem. Oh Man ! Edge Length set to a higher value and now it`s working perfectly.

Best regards,

Thomaskl

Excellent!!!

Vlado, alot of these handy tips should be in the help index if that are not already :wink:

or a scale with image size checkbox next to edge length :wink:

that too :smile:

It’d have to be smart enough to remember what your render size was when you applied the displacement modifier in the first place though so it may cause problems elsewhere - maybe if there was a way to make max use tooltips that you could have a little reminder pop up in when mousing over it? No doubt all of this stuff will be taken care of by the full manual with 1.5 though.

Nah. Just switch it to be relative to render size. Percentage instead of pixels maybe?

Here’s a quick question guys - where does one find the edge length setting? :smile:

Cheers!
-- DJ

Arrr… “VRay:: Default Displacement / Edge length” :stuck_out_tongue:

What would an optimal setting be? Is there a min and max that I should use?

ta.
-- DJ

Either that, or the setting in the VRayDisplacementMod, depending on which you are using.

What would an optimal setting be? Is there a min and max that I should use?It all depends on the desired detail in the displacement mesh, so it’s a bit of trial and error. Don’t go below 1, for sure :slight_smile:

Best regards,
Vlado

Well its one of those things that depends on your displacement.
So yeah keep lowering it until it start to affect quality then raise it abit.

Or don’t use the ‘view-dependent’ option and set your edge length in whatever units you work in. If it’s 10km of grass you’re rendering, this will add to the rendertime tough… But for carpets and stuff like that I find the view-INdependant mode to work better and more reliable, especially in animations where you move closer and further away from the displaced object. At least your displacement will look the same at every resolution.

very true, good tip