Customizing dmtx_par (with -sj) causes error in DCoeffGBRecipe

Dear all,

I was trying to set customized dmtx_par in Daylight Coeff Grid Based Recipe, and it came out with error. However, when I used the default dmtx_par, it worked fine.

  • default dmtx_par: 0.25 -ab 3 -ad 5000 -ar 16 -as 128 -dc 0.25 -dj 0.0 -dp 64 -ds 0.5 -dr 0 -dt 0.5 -I -lr 4 -lw 2e-06 -c 1 -ss 0.0 -st 0.85
  • customized: -aa 0.15 -ab 5 -ad 2048 -ar 512 -as 1024 -dc 0.25 -dj 0.0 -dp 64 -ds 0.2 -dr 2 -dt 0.5 -I -lr 6 -lw 0.004 -c 1 -ss 0.0 -st 0.15 -sj 1

I basically follow this website to set rendering options https://floyd.lbl.gov/radiance/refer/Notes/rpict_options.html

My questions:

  • Are there any other required thresholds of these Radiance parameters in the Daylight coefficient recipe other than the Radiance website? If not, what is the possible reason to cause this error?
  • Regarding -sj, I also decreased it to .8, however, same error still came out. So should I remove -sj and what is the possible reason behind it?

I am in the entry-level of using Radiance, any suggestion will be appreciated.

Below are error messages.


runRadiance error message:

Found 14 opaque surfaces.
Found 1 fenestration surfaces.
Found 0 window-groups.
Writing recipe contents to: c:\ladybug\TESTROOM\gridbased_daylightcoeff
Calculating solar values…
Number of sun up hours: 4432
Writing sun matrix to c:\ladybug\TESTROOM\gridbased_daylightcoeff\sky\sunmtx.smx
Unloading the current values from the analysis grids.

loading total and direct results for 4d99cbc9-acb2 AnalysisGrid from scene::default
c:\ladybug\TESTROOM\gridbased_daylightcoeff\result/scene…default.ill
c:\ladybug\TESTROOM\gridbased_daylightcoeff\result\sun…scene…default.ill

Runtime error (Exception): Failed to load the results form an empty file: c:\ladybug\TESTROOM\gridbased_daylightcoeff\result/scene…default.ill
Double check inputs and outputs and make sure everything is run correctly.

Traceback:
line 67, in script
line 326, in set_coupled_values_from_file, “C:\Users\Di Liu\AppData\Roaming\McNeel\Rhinoceros\6.0\scripts\honeybee_plus\radiance\analysisgrid.py”
line 397, in results, “C:\Users\Di Liu\AppData\Roaming\McNeel\Rhinoceros\6.0\scripts\honeybee_plus\radiance\recipe\daylightcoeff\gridbased.py”

Thanks!

Hi,
I would like to share an update regarding this issue:

The above error is disappeared after removing -sj from radiance parameters, although the reason behind it remains unclear for me yet, particularly when -sj is necessary in DIVA.
I also discovered that Radiance parameters applied in DIVA cannot be directly applied in HB+ even with removing -sj, otherwise it could cause abnormally low DA, UDI, and sDA outputs.

Currently, I am still confused about how to set reasonable Radiance parameters in HB Plus. Any comments or further information will be appreciated :slight_smile:

Most of those parameters are not relevant for rfluxmtx and rcontrib, which are what HB+ is using to do raytracing.

Yup, that is expected and exactly how it should be. You’d notice that for the same input params HB+ takes a lot less time too. DIVA uses an older, not necessarily inferior, set of Radiance programs that incorporate ambient caching algorithms to generate the ray trace results. As I mentioned previously, HB+ uses rfluxmtx and rcontrib, which rely on pure Monte Carlo method. Basically, you need to set really high (like 10 to 20 times higher) value for ad when running HB+. The values for aa, ar, are not relevant for HB+.

( I got tired of typing this on my cellphone…I will try to update this post in a couple of days when I am back from my travel)

2 Likes

Hi @sarith
So sorry for the late replying.
Thank you so much! It made me feel much more clear right now. I am also wondering why low -ad value doesn’t really have an influence on ASE results.

Hi @catsquito, I am sorry I forgot to update my post from last week. Anyway, I am glad that you have figured things out.
The value of Ambient divisions(-ad) is meant to control diffuse inter-reflections by increasing or decreasing the amount of random sampling in the scene with higher or lower values respectively. The calculation for ASE is not affected by -ad for two reasons: firstly the calculation itself is meant to be direct as per LM-83-12…
image
…secondly, for the calculations done specifically within HB[+], we also convert all the geometry to black. This ensures that even the specular reflections are turned off. As a result of this none of the ambient parameters have any bearing on the results either. You can read more about the rationale for turning geometry black in the description for the direct sun part of the Five-Phase Method and also in Andy Mcneil’s presentation from the 2014 Radiance Workshop (pages 123-128).
Just keep in mind that unlike the 5165 sun set for the direct part of the Five-Phase Method, in HB[+], we use actual sunpath positions.

Regards,
Sarith

2 Likes

Hi @sarith,

Thank you so much for the detailed replying!
I am reading through the paper and the Radiance Workshop presentation file.
Thank you for providing nice resources!
:slight_smile:

Di

Hi @sarith,
Hope you are doing well!
I tested 5-phase method, ASE still gets zero :no_mouth:
I am not sure if it’s because BSDF material I used is a single clear glazing? is it possible to get ASE in 5-phase with using single clear glazing? or it has to use a more complex window system?
I also read this post, but couldn’t understand it well…