Here’s an initial hi-rez test of a nebula i’m working on. Full image size is 2048x1536, so yeah it’s big. And this is just a ‘test’ image to see how certain details are working out, and if/how noise settles down. (needs more noise correction)
I really wanna get nebulae down. I’m HOPING I can somehow ‘break in to’ astronomical illustrations for magazines, or even sci-fi book covers, , , etc …
I’ve previously tried using point-clouds as the volumetric density input; but that results in spheres of volume around each point. Perhaps workable using post methods, but not ideal.
This image I sculpted a sphere into a somewhat nebular shape, and used the volume of THAT as the volume’s ‘shell’. Plus a couple of 3D procedural textures (actually the same texture scaled differently) to ‘break up’ and further structure the gas clouds.
Clearly I’m getting some undesired results in certain areas. Some parts are just WAAAYY too hard-edged, and I don’t know how to soften the volume there. Also, there is still too much noise in the thinnest bits, and I don’t know which setting will help that the most with the lowest render-overhead. Light shadow samples? Volumetric resolution? Step size? Threshold? … I’m thinking upping the light shadow samples from 6 to 8 or 10 will help, but not sure.
Full rez image was about 3.5 hrs render time. I know it will go up from here.
I still want to add in more stars. Currently this scene is lit by one point light inside the sculpted cavity. A star-cloud would of course have more. But they will ALL be much dimmer. The current, primary, star is a strength 3 point light.
However, the current ‘star’ is inside a cavity in the nebula object. Many of the others will be outside the nebula itself, and thus will need to be visible as points of light, yet still illuminate the mesh.
The only light I know of that can do both is the spotlight, and that increases render times DRASTICALLY!!! I’m wondering if maybe I can use small spheres as mesh-emitter lights? then I could give them proper flare/etc in post, and hopefully they’s still illuminate the nebula without too much render overhead.
Of course, I haven’t yet tested mesh-lights with volumetrics. I have NO IDEA if it works, or how long it takes. OR the light quality.
Do I have any other options for adding VISIBLE stars that cause flares but also illuminate the mesh?
Also, I need to add in dark dust clouds. Dark bits stretching through this volume.
Also, I see I need to add in another cloud texture (or something) to make it all look less ‘smooth’.
Also, I’m not sure about the colors right now. I’m basing the colors off of both visual and filtered image ideas (NASA, APOD, ESA, etc …) … Thus blue is reflection nebula – light reflected off the dust cloud from any stars. All other colors are light generated by the atoms of the cloud itself by excitation from the radiation from any stars. Red would be hydrogen alpha (HA) emissions. I also need to add in green – singly ionized oxygen (OII), which is a visible wavelength (unlike the red HA emission, except under the darkest skies and the clearest views). Also in many space photos is an orange color representing doubly ionized sulfer, I think, (SIII). PLUS the dark clouds I mentioned previously.
I think I just need to duplicate the object, and give that dup a volume material for the HA and SIII bits, and change my current Blue-Red scheme to Blue-Green to account for the OII emissions.
And of course add more stars.
And a 3rd volume OR particle system for the dark bits of molecular dust.
I can put up a .blend, but I’m not sure where. I’m a little sketchy on FileFront, only one I know for files as such … Any other ideas where I can post my .blend? thx
Full rez image. Unfortunately, I don’t know how to ‘inline’ thumbnail images on this forum ::
http://img853.imageshack.us/img853/772/nebulasculpthirez01.png
does this work? http://img3.imageshack.us/img3/4373/nebulasculpthirezsmall0.png
Perhaps have anuther look thru the manual too, may have been updated …
