Help - Search - Members - Calendar
Full Version: Radiosity Render Timing
Hash, Inc. Forums > Technical Direction and Development (Learning Animation:Master) > A:M Rendering, Compositing and Special Effects > Texturing, Lighting and Effects > Lighting Effects
MattWBradbury
I wanted to test the rendering time of radiosity before post processing functions such as Jittering, Sampling, and Irradiance calculations. The tests were just to see how long it would take to render out a simple hallway when increasing the number of Photons by a multiple of 10 starting at 1000. The first three tests did not seem to differe from each other. They where between one minute and thirty-three seconds and one minute and thirty-seven seconds; however, the final render took three minutes and twenty-one seconds to render. Here's the examples:

1,000 photons (1:34):
[attachmentid=19658]

10,000 photons (1:33):
[attachmentid=19659]

100,000 photons (1:37):
[attachmentid=19660]

1,000,000 photons (3:21):
[attachmentid=19662]

I find it odd that the render time would decrease when the number of photons were increased. I also find it odd that the number did not differe much when increased by a multiple of 100, but on the last test, gave an output more than twice that of the three prior times. I'll do some more tests to see what kind of curve this is.

As of right now, the sample is very small, and only renders a third degree function. I should also try this without adjusting the sample area.

[attachmentid=19663]
MattWBradbury
I did a second test. This time around, I did not change the Sample Area (Constant: 100,000), and I've added three more points to collect timming data from.

1,000 photons (2:00):
[attachmentid=19668]

5,000 photons (2:46):
[attachmentid=19669]

10,000 photons (2:52):
[attachmentid=19670]

50,000 photons (2:38):
[attachmentid=19671]

100,000 photons (2:44):
[attachmentid=19672]

500,000 photons (3:36):
[attachmentid=19673]

1,000,000 photons (4:40):
[attachmentid=19674]

Strangly, the coordinates seem to point towards a solution set of a third degree polynomial. I'd say that the best ratio for quality and time is around 100,000 photons.

[attachmentid=19675]

Though as you can see, there is no visible amount of light coming through geometry at one million photons, so if I were going to render a Still, I'd most likely wait for the extra time.
Mr. Jaqe
Interesting...
the_black_mage
hey i could really use this thx!
Nunsofamerica
I know this is a little off topic, but how did you model that house? please send it to me in a personal message, thanks
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2012 Invision Power Services, Inc.