Radiance Parameters for Daylight Analysis


#1

HI!

I am setting the radiance parameters for a daylight analysis of a simple shoe box with an exterior ground.

While setting the radiance parameters for the simulation I noticed that some parameters are somehow offset for the notion I had of theirs maximum. But since I guess that @sarith set it, I think that somehow I am missing something.

Since my scene has a maximum dimension of 30 meters and I set the -aa 0.1, the -ar 128 is more than enough for this case. Because I need accuracy but I do not have a complex shape I set an -ab 4 .
When looking at the -ad and -as I was surprise to see that were set at 25000 and 4096 respectively, when I knew that 4096 and 1024 were their maximum.
So I checked this post and long as this page and probably I am missing something.

A part of this clarification I would like to have a better understanding of the parameters -c and the difference with -as and how should I use them together.

Also, I have noticed that there is a strange definition of the limit weight parameter, that in my knowledge should be a number, but it is set as: -lw 4e-07 (I enclose screenshot).

Thanks for the help.
Best regards,

Federico


#2

@Federico587 This questions warrants a detailed reply, however, I am short on time…so very quickly… Rcontrib does not rely on ambient caching so -ar, -as etc don’t have any relevance. For convergence of values, the only three parameters that really matter are -ad, -lw and -ab (-c matters in case of imaging but for illuminance it is set to 1).
The ambient settings you referred to on the rpict page are not relevant for Rcontrib (annual daylighting simulations in HB+). This part of the API needs some cleaning up. As far as performance is concerned it does not matter as Rcontrib converts all -aa values to 0 internally.

Read more here and here.

By the way, 4e-07 is a number … 0.0000004.


#3

Thanks very much for the reply @sarith. I will check both links you posted here, and I am still laughing for the -lw!
Best,

Federico


#4

Hi @sarith,

Thanks again for providing the links to understand more about Rcontrib.
In the first link you have provided indeed there is some info about the -ad and -lw but I could not find any “general rule” to use them. With the rpict, for example, knowing the max size of your model you can adjust the parameters -aa and -ar.
Do you have any recommendation for the Rcontrib?

the model I am working on has a max distance of 30 meters, no shadings, and a grid distance of 0.5 meters. For a grid based simulation with high accuracy, which parameters values should I have to use? If I set the complexity of the Radiance Parameters component to 2, the results will have an high reliability?

I know that you are busy, so any suggestion from your end will be very welcomed.

Bests,

Federico


#5

I leave the hard questions for @sarith but as a reference which is not a comprehensive study but a good starting point check this page to see the effect of changing -ab and -ad on the results. Change the room to 2 which has some level of complexity.


#6

Hi @mostapha,

Indeed that page you say it is very useful and it is the first resource I checked before asking.
I am waiting to see if @sarith can give some good tips.

Thanks again,

Federico


#7

Hi @Federico587, sorry for the late reply. The only ambient parameters that are relevant for rcontrib/rfluxmtx (i.e. annual simulations in HB[+] ) are -ad and -ab.

You can set -ar, -aa etc but they won’t have any effect as rcontrib works through pure Monte Carlo raytracing.


#8

Here is a little bit more explanation for future reference and for beginner users:

If you set -aa to 0, it tells radiance to turn off the ambient cache and get a straight Monte Carlo path tracing method. This is an option when you use the standard raytracing with radiance (e.g. rtrace). In daylight coefficient based simulations (e.g. 3phase, 5phase, annual daylight with honeybee[+]) which uses rcontrib or rcontrib based commands (e.g rfluxmtx), rcontrib sets the -aa value to 0 during the calculation regardless of user inputs, for the reason that @sarith mentioned above. With -aa set to 0 the -ar settings has no effects.

Optional readings:

  1. See this discussion for an example.
  2. rcontrib documentation.