Indigo Renders...
-
Basically unbiased means no approximations, so slower but more accurate. The hardest part of writing an unbiased renderer is making it happen fast.
-
@charly2008 said:
Hi,
Last night, I have downloaded and tried the INDIGO renderer. I used the standard settings for the scene below.The Time required for this scene was amazing, about 1.5 hours.
Then I rendered the same scene with Kerkythea with the preset no.3 (PhotonMap-Quick). The Time required for this scene was about 5 Minutes. Below you can compare the results.
Karlheinz
These kind of scenes don't really show the speed of Indigo. You want to compare two interior scenes that are purely lit by sunlight. And include the set-up time of each renderer. Indigo might win that race. ;D
-
I was a bit out of the loop on the renderers, what is a difference between biased and not biased render?
-
Unbiased means the renderer models light in a realistic fashion, so you render is made by shooting rays around your scene and modelling precisely how they interact with the materials in your scene.
Biased renderers dont necessarily stick religiously to an accurate model of light, so they can make good guesses at how light behaves to speed up the rendering process as much as possible. The downside to this is that you dont always get an accurate render out the end, and some effects (such as caustics) can be more difficult to reproduce.
-
Within my limited knowledge, this is what I understand:
All currently available render methods take shortcuts, as they don't simulate light's interaction with every single subatomic particle and physical force present in your scene. They are all physical/mathematical models, or approximations to reality. The difference between "biased" and "unbiased" methods is that the "unbiased" ones use a more-or-less brute-force approach, meaning that they interpolate nothing, while the "biased" ones do interpolate some values to arrive to a quicker, but less accurate, solution.
Biased methods could be compared to a map of voting trends by cities based on an "Exit-poll", while Unbiased ones could be compared to another map based on actually counting vote by vote in successive random passes over the whole country.
And of course, I want both methods for different situations
-
@notareal said:
If you are interested about bias in rendering, I suggest reading this http://www.cgafaq.info/wiki/Monte_Carlo_Bias
^ explains biased and unbiased. Did you know, even unbiased renderer can fail - make a tiny hole in a wall and see if light comes trough.
-
Technically it wouldnt fail, it would just take a very long time to complete.
-
Not that it really matters, but it will fail if hole is small enough. Stopping criterion will come to play. Just try it out with your favorite renderer. Two rooms and a microscopic hole in a wall between them. Lightsource in the other room and camera at the other.
-
thank you so much people, an awesome lecture
-
I've been doing some renderings of the house that we are currently building. I'm pretty happy with this one except that I wish I had taken the time to model the seams of the furniture.
-
Dale that is simply wonderful imho - I really like the whole ambiance of the image and would not fret too much about the seams on the furniture. Really good.
Advertisement