questions about TG GI system and AA

[credemon]

good morning all 
i m kinda stuck about cloud gi quality
whats the difference? i mean in usage.. with gi caching enabled i tried to render 100 frame 
i used still /high since there is no changing on clouds shape or animated surfaces or etc.. cached at 5 frames for 100 frames with a total of 20 gi cache file...
but when i used sequence quality.. well caching seems like forever... 

i tried to figure out whats the difference between there 2 from help files.. but there was no info actually.. or i couldnt  find in the wiki either 
and another question ... 
did you ever tried to render a close up clouds with animated cam? 
if so what was your render times and AA & PNT was? 
btw for the scene i changed the clouds voxels were 1300 gi caching was extremly slow with a amd ryzen 5950x.. 

[credemon]

and another thing... pls correct me if i m wrong...
the higher radius you have the more millions of voxels you gonna need...
also when i rendered a crop there is a little bit difference between what i see in the render windows VS output file.. is this becouse of "soft clipping"? whenever i disabled it.. what i see and what it outputs become the same ... (btw OpenExr 32 bit )
and will the atmos sample and light (sun) samples affect over the cloud ?

[Kevin Kipper]

Hi credemon,

Thanks for your questions about the cloud GI.

The online documentation about this particular feature is located here:
https://planetside.co.uk/wiki/index.php?title=Terragen_4_Global_Illumination#GI_in_Clouds

I would also recommend watching Part 23: Render Settings Global Illumination of the Terragen for VFX tutorials.  This feature is explained at about 02:48:15 in the compilation video.
https://www.youtube.com/watch?v=g0dtt4IusJM

The "Cloud GI quality" setting determines the accuracy of the global illumination for Cloud Layer V3 type clouds only.  Each entry in the drop down menu doubles the number of light samples from the previous entry. 

By "...close up clouds with animated cam?" do you mean rendering an image from the point of view of a camera within a cloud?  I can't recommend Antialiasing and Pixel Noise Threshold settings as they might be different for every project or even within a given cloud layer due to the density of that part of the cloud and movement of the camera, but I can confirm that rendering volumetric clouds from within a cloud is time consuming.  Rendering smaller resolution images as a test to determine how long render times will be on a larger resolution might be worthwhile.  If your final image is 1920 x 1080 and you render a test image at 960 x 540 which takes 1 minute to render, then it's a pretty good guess that the larger image would take about 4 minutes to render as it's twice the width and twice the height.

You're correct that the higher the cloud layer radius, the more millions of voxels you'll need to cover that area.

By default, the Render View window applies tonemapping to the image displayed in it.  You can actually adjust the settings found under the Renderers > Tonemap tab and see their effect in the Render View.  For example, you could disable the Soft clip effect checkbox and the Render View would immediately reflect that.  The EXR image never has tonemapping applied to it.  If you want your final rendered image to include the tonemapping, try saving it as a TIF file instead.

I'm not quite sure what you mean by "..atmos sample and light (sun) samples affect over the cloud?".  The Atmosphere samples aren't necessarily used to calculate the cloud, but their effect in calculating the atmosphere has an effect on the cloud.  In other words, really low samples in the atmosphere would change the look of the atmosphere, which might also change the look of the cloud.