[Radiance-general] Understanding rvu options

[Greg Ward]

The default parameters are set in rvu for speed over accuracy.  Decreasing the value of -aa will help a lot with your scene.

I recommend learning to use rad and/or trad to help set rendering parameters based on easier-to-understand variables.  Here is a paper on the subject:

	http://radsite.lbl.gov/radiance/papers/erw95.1/paper.pdf

And of course the man page is always a reasonable place to start:

	https://www.radiance-online.org/learning/documentation/manual-pages/pdfs/rad.pdf

Cheers,
-Greg

> From: Christopher Rush <Christopher.Rush at arup.com>
> Date: January 30, 2018 1:04:22 PM PST
> 
> What you're seeing makes sense to me, that the quality of rendering relies more on ambient calculation parameters when the calculation is 100% diffuse/ambient light, without a direct light source.
> 
> You also seem to be moving the parameters in the direction of higher accuracy and longer rendering time so shouldn't be detracting from validity.
> 
> As for reading materials, I'm sure lots of people could make suggestions. The book Rendering with Radiance is likely a top suggestion. Also, here's a presentation that I recently re-discovered from a past annual workshop:
> https://www.radiance-online.org/community/workshops/2011-berkeley-ca/presentations/day1/JM_AmbientCalculation.pdf
> 
> Actually you could browse all the past workshop presentations for relevant materials, although I'm not sure how easy it will be to judge them by the titles.
> https://www.radiance-online.org/community/workshops
> 
> 
> Regards,
> Chris
> 
> 
> -----Original Message-----
> From: Dion Moult [mailto:dion at thinkmoult.com] 
> 
> In order to get the leaves actually rendering and to get the shadows of the leaves less splotchy, I've had to increase -ps to 1 and -ar to 2048. My full options are -ab 2 -ad 4096 -ar 2048 -ps 1. Is this the right approach?
> 
> I feel guilty that I am adjusting these parameters to get a more "believable"
> image without fully understanding it -- I might compromise the validity of the simulation (although yes, the material in this test is nonsense, of course).
>