Procedural volumetric stars

Started by Denis Sirenko, April 24, 2018, 06:46:48 AM

Previous topic - Next topic

Matt

#15
Talking about inverse square falloff using functions - I used that technique here, but along a line instead of a point:

http://www.artofvfx.com/TRONLEGACY/TRONLEGACY_PF_VFX_06.jpg

(EDIT: for a line it would just be inverse rather than inverse-square)

Matt
Just because milk is white doesn't mean that clouds are made of milk.

WAS

#16
Quote from: Matt on April 25, 2018, 04:31:23 PM
Talking about inverse square falloff using functions - I used that technique here, but along a line instead of a point:

http://www.artofvfx.com/TRONLEGACY/TRONLEGACY_PF_VFX_06.jpg

Matt
l
Honestly haven't even ever thought of that approach. So far I've been using the modulator just to give the stars their individual colours via a extreme contrast mask. I'm curious where one would start?

I have done luminous background stars, they work pretty decent, though I've had numerous issues with camera positions and stars not appearing right. I don't know if that's the curve/angles of the background object or what. This method also is of no help for nebulas/galaxies.

I'm still tinkering with an object population and some PFs to see if I can get an actual volumentric looking population with some decent lighting, but so far it's all just a headache and not even remotely as simple as a cloud setup. Which works much better in final result. I know you say the light is working the same, but somehow clouds (not being a solid object or what, my cloud at least not not totally solid, meaning light is being produced by semi transparency) and clouds capturing this works differently than an object. It's plainly visible.

WAS

#17
I'm curious what the actual luminosity of the Ambient slider is. For example. What is "1" on the ambient slider compared to "1" on the Luminous slider? Cause They also yield different results, adding to the concept that what TG is interpreting the two instances differently.

On a side note I'm finding it crazy how my assumptions on films are starting to pan out with TG use. I also thought Tron was using TG clouds. There's a Star Trek game in development where the guy has used TG to create his sky boxes instead of using TrueSky. Right off the bat I knew it was TG. I'm not sure if that means I'm recognizing true realism or some sort of surrealism, but it's really interesting to discover.

Matt

Quote from: WASasquatch on April 25, 2018, 04:38:19 PM
l
Honestly haven't even ever thought of that approach. So far I've been using the modulator just to give the stars their individual colours via a extreme contrast mask. I'm curious where one would start?

For an individual point you calculate the distance from the point (using either function nodes or a distance shader) and then divide by that distance (or by the square of that distance). For many procedurally generated points you might need to use a technique similar to how people have rendered craters using a Voronoi function. I haven't tried it though.

Quote
I'm still tinkering with an object population and some PFs to see if I can get an actual volumentric looking population with some decent lighting, but so far it's all just a headache and not even remotely as simple as a cloud setup.

Quote
Which works much better in final result. I know you say the light is working the same, but somehow clouds (not being a solid object or what) and clouds capturing this works differently than an object. It's plainly visible.

I have not seen that difference demonstrated. I want to understand what difference you're talking about, but I understand that you have a technique that works for you and that's really what counts.

Matt
Just because milk is white doesn't mean that clouds are made of milk.

Matt

Quote from: WASasquatch on April 25, 2018, 04:44:34 PM
I'm curious what the actual luminosity of the Ambient slider is. For example. What is "1" on the ambient slider compared to "1" on the Luminous slider? Cause They also yield different results, adding to the concept that what TG is interpreting the two instances differently.

Ambient is multiplied by the colour of the cloud (and also by density at each point). If the cloud colour is black, ambient will not make it any brighter. If the cloud is red, then white ambient will make the cloud a brighter red. It's like shining light onto the cloud, except it has no direction, it's just "there". The amount of ambient light you see depends on how much the cloud can reflect back to you according to its colour and density.

Luminosity (also known as emission) is different in that it works even if a surface or cloud is black. It simply emits light regardless of any other reflection characteristics of the surface. If the cloud layer had a luminosity control (it doesn't yet because most needs are covered by the ambient control), then it would be completely independent of the cloud colour. However, just like with ambient, the amount of light you see would still be affected by cloud density.

Matt
Just because milk is white doesn't mean that clouds are made of milk.

WAS

#20
Quote from: Matt on April 25, 2018, 04:52:50 PM
For many procedurally generated points you might need to use a technique similar to how people have rendered craters using a Voronoi function. I haven't tried it though.

I actually was thinking of using that very voronoi setup in general for the original mask, as it creates more varied (further apart) points, which are also perfectly round, and can be clamped in to desired points.

Quote from: Matt on April 25, 2018, 04:52:50 PM
I have not seen that difference demonstrated. I want to understand what difference you're talking about, but I understand that you have a technique that works for you and that's really what counts.

Matt

Basically, I'm asking what the calculations are on the sliders for luminosity. When I setup two instances, one a pop, and one clouds, at the same "supposed" luminosity, they act different whether their emitted light.

I'm wondering if this is because the Ambient Slider is not the same as the Luminous slider, or maybe because our clouds are not solid objects. They're transparent objects with opacity, and their own internal scattering before emission?

WAS

Quote from: Matt on April 25, 2018, 05:02:40 PM
Quote from: WASasquatch on April 25, 2018, 04:44:34 PM
I'm curious what the actual luminosity of the Ambient slider is. For example. What is "1" on the ambient slider compared to "1" on the Luminous slider? Cause They also yield different results, adding to the concept that what TG is interpreting the two instances differently.

Ambient is multiplied by the colour of the cloud (and also by density at each point). If the cloud colour is black, ambient will not make it any brighter. If the cloud is red, then white ambient will make the cloud a brighter red. It's like shining light onto the cloud, except it has no direction, it's just "there". The amount of ambient light you see depends on how much the cloud can reflect back to you according to its colour and density.

Luminosity (also known as emission) is different in that it works even if a surface or cloud is black. It simply emits light regardless of any other reflection characteristics of the surface. If the cloud layer had a luminosity control (it doesn't yet because most needs are covered by the ambient control), then it would be completely independent of the cloud colour. However, just like with ambient, the amount of light you see would still be affected by cloud density.

Matt

This seems to shed some real light on why interactions are different with the two setups.

Matt

Just because milk is white doesn't mean that clouds are made of milk.

pclavett

Love your experimenting with the nebulas and stars !
I just love space scenes but using pictures for background has its issues.
Love the  fact that some of you are experimenting into procedural generation of these items.....will follow closely.....not sure i will understand though !
Thanks for sharing !
Paul