Page 2 of 24

Re: A natural situation

Posted: Mon Dec 02, 2013 8:17 am
by CTZn
Nice, big set ! There's this feeling of openness.

On a side note, glass acceleration is expected to have a very small general performance hit with a great benefice for really flat surfaces. It is usefull if your cloud banks are to be bound within a box.

Re: A natural situation

Posted: Mon Dec 02, 2013 9:08 am
by pixie
Awesome stuff, congrats! :mrgreen:

Re: A natural situation

Posted: Mon Dec 02, 2013 9:11 am
by OnoSendai
Some wicked fireflies on the second image.
I think they're due to eye->participating medium ->specular sea -> sun.
So actually this is a type of path that bidir would sample better (in theory).

Re: A natural situation

Posted: Mon Dec 02, 2013 2:32 pm
by yonosoy
This time path tracing seems a too hard option.
test_012(path_tracing)21h.jpg
But metro is near...
test_012(metro)20h.jpg
The meaning of use glass acc. on is a litlle bit strange. Acomplished the condition of very low curved and produces this with bidir. metro in 3 hours.
test_012(bidir_metro_glass_acc_on).jpg
This use has been intensily tested with general good result (test_011). Actually the two last test running are glass acc. on for metro and bidir metro. The try over simple and bidir path tracing with glass acc. on seems extremely hard too. Metropolis is more or less able to handle a complete definition with an ocean that extens over the 98% of the planet (this is at the moment the North Pole!! All the rest is clear natural water with a depth of hundreds of meters and a subtle SSS).
The concept is easealy developed at the moment. Is intention to share in a weeks the procedural file (232 Kb) of houdini.
At the moment half load of polygons 23 mill.

CTZn, now is possible connect Maya with Houdini. This project starts with Mti and is pretty nice for us. So maybe the workflow (that actually tends to blender and cinema) could move to this connection.

Re: A natural situation

Posted: Tue Dec 03, 2013 5:20 am
by CTZn
Nice tests serie yonosoy, that is usefull.

I'm working on a radius based clouds shader, I'm rendering a dummy test scene now. The issue is that atmosphere and cloud volumes are exclusive, or, they are defined together. That implies to model the atmospheric scattering and drop the specialised Indigo medium at once.

It would be great in this regard to have a shader input meant for clouds and haze for the atmosphere medium !

You are welcome to use MtI of course. Even though I will not be available soon for some time (I'm moving place) I'm curious to know what would make the connection easier.

Re: A natural situation

Posted: Tue Dec 03, 2013 5:24 am
by CTZn
I think they're due to eye->participating medium ->specular sea -> sun.
Most certainly, they don't seem to happen on my side where there is no specular material involved beside clouds.

Re: A natural situation

Posted: Tue Dec 03, 2013 7:41 am
by yonosoy
Results for glass acc. on ( the transparency of bidir!!!).
test_012(metro_glass_acc. on)02.jpg
17h. I3 (4 slowly threads)
test_012(bidir_metro_glass_acc_on)_02.jpg
22h. amd (4 threads)
...radius based clouds shader... (the point.). Thanks CTZn.
The "wicked fireflies" only are present when an ocean is involved (more in the way that water surfaces increase).

A more suitable solution of the test_011 raw. Turbidity 3 in all the serie.
test_011(metro_glass_acc_on)02.jpg
Note: Do you see the wide light beam of aperture difraction in bidir ? Is present in path but the phenomenon loose power with the convergence. In the meeting point we almost saw this well described: paralell bands of difraction (couloured) covers this "captured beams" (sorry, working in the fly since years and no documents about this was saved!). All this about physics considerations in the cam model. Difficult to complete with athmosphere but usually present.

Re: A natural situation

Posted: Tue Dec 03, 2013 9:39 am
by yonosoy
Some post pro A. D. at the meeting. Simples test out of convention (and a sample of how in the ocean could move).
Exploring what (maybe) seven parameters could do. No obstacle map was used.

Almost some documents was saved... but the scenes, param, etc... I don't know.

Re: A natural situation

Posted: Tue Dec 03, 2013 10:24 am
by CTZn
yonosoy wrote:...radius based clouds shader... (the point.). Thanks CTZn.
Hold your horse, so far I've got the envelope wrong ! I will post in my wip thread when I've got it right. I keep tweaking the code and there's no way I'm getting a clear picture ;P

Re: A natural situation

Posted: Sat Dec 28, 2013 10:21 pm
by yonosoy
Success only mixing the cloud layer with the atmosphere medium. Same conditions as tried to explain before (Fakeshamus tricky scene). All info the pigs, focused over CTZn and OnoSendai, but for general test too.
simple_clouds(breaks_down).png.jpg
breaks down at 45 samples pixel
clouds_in_atmosphere.jpg
an hour of process 165 samples pixel (still running stable)
clouds_in_atmosphere(working_5).pigs
(16.85 KiB) Downloaded 155 times
If there are some incoveniences (finally breaks, etc...) sorry. Working in faith.

Edit: Still a problem from under the volume. Maybe yet solved (IOR connections (?)) ... The run don't breaks.

Re: A natural situation

Posted: Sat Dec 28, 2013 10:42 pm
by Oscar J
That first pic is really good!

Re: A natural situation

Posted: Sat Dec 28, 2013 10:47 pm
by yonosoy
File was loss again, working around. In this piece of CTZn work there are thousands of awesome situations. Bad hands to express... (shaded, instancied, ... at the moment this "dry" documents).

Re: A natural situation

Posted: Sun Dec 29, 2013 5:13 am
by CTZn
Awesome yonosoy, you achieved far better results than I did !

Re: A natural situation

Posted: Sun Dec 29, 2013 12:04 pm
by yonosoy
An approach for the next serie. The datum, 13300 millions of polygons (indigo stops the count at 9999 millions) with a consumption of RAM of 737 MB. Speed: 34K samples/sec. Time: 43h. (... only the cathedral, there will be a city around ...).
cathedral_approach.jpg
Explanations soon.
Thanks Oscar, now it's feeled that the answer was as dry as the documents... Not a cold thread but so focused in the work. Feel free to participate ...

Re: A natural situation

Posted: Sun Dec 29, 2013 2:11 pm
by yonosoy
Second positive (maybe now it's better to wait for the new medium and conditions). Speed: 23k samples/s Time:4h30min. Since there are no details and a propper material, till here (only 600 samples per pixel).
clouds_in atmosphere_2.jpg