Increasing the level of detail in annual daylight simulations: major difference in results

Hello everyone,
I am simulating a deep classroom (7.7m x 9.3m x 3m) with a south facing windowed facade and extruded border shades to know the spatial daylight autonomy and illuminance values.
I noticed that there is a major difference in the results when I increase the precision of the simulation with the radiance parameter.

For example, when I run the simulation with a level of detail of 2 (and the recipe level of 2 for annual), I get a spatial daylight autonomy (sDA300,50) of 50.37%. When I do the same simulation, but without the “radiance parameter” , I get a spatial daylight autonomy (sDA300,50) of 33.33%.

Although I understand that the level of detail allows to obtain more accurate results, I wonder if it is normal to have such a major difference between the results ?

Here is the file : (350.7 KB)


The short answer, without having done any quality control of your model, is yes.

If parameters are not given, the precision is given by the following parameters:
-ab 2 -ad 5000 -lw 2e-05

Detail level 2 gives:
-ab 6 -ad 25000 -as 4096 -c 1 -dc 0.75 -dp 512 -dr 3 -ds 0.05 -dt 0.15 -lr 8 -lw 4e-07 -ss 1.0 -st 0.15

It’s not easy to understand the effect of all of these, but e.g.:
ab 2>6 increases the ambient bounces from 2 to 4

For more info on these you can browse these:


Hi @RachelLevesque,

@LittleBuddha is right that the ambient bounces will increase daylight levels and thus sDA.

You don’t necessarily have to run the simulation with detail level 2 to get good results. If the model is simple the -ad value of 25000 in detail level 2 is likely overkill. Try to select detail level 0, and then add a custom value for -ab as shown in the image.

1 Like

Thank you so much for your answer! Very helpful :slight_smile: