First of all, great work, the plugin is looking wonderful! It’s almost too good to be true, reducing the calculation time from 1 minute to 24 ms, I can’t believe my eyes! I can’t wait to integrate this into my workflows.
However before doing so I would like to see agreeing results between honeybee and cyclops. I have made a small comparison between the two scripts to better understand the logic behind cyclops and I’m failing to getting agreeing results, I see a whooping difference of 25% in the difference between both, even after simplifying the reflectance following the info provided here:
The maps seems to agree, that’s nice, but the amount of daylight inside the space is not the same. Here is the script used, which is based on the tutorial Daylight Factor from Cyclops:
Thank you for the kind words! Glad you see value in it.
Regarding the simulation, the results in Cyclops as with Honeybee/Radiance depend a lot on the settings that are used. In the file you have uploaded Radiance is running with 3 ambient bounces (-ab), 2048 ambient subdivisions (-ad), a 145 patch Tregenza sky and a traditional overcast sky. Cyclops is running at higher settings with 6 ambient bounces, 10,000 ambient subdivisions, an equal area/aspect sky with 512 patches and a standard overcast sky. The radiance settings also use ambient interpolation since the (-aa) parameter is set above 0. Cyclops does not use ambient interpolation and is using pure ray-tracing.
As an example of how the settings affect results, just by changing the detail_level in honeybee from medium to high you will get a 30% higher score.
All of the differences above are affecting the results. Please see attached file where the radiance parameters are updated to match the settings used in the Cyclops analysis. The average score that is compared in the script now differs by approximately 2% (7.98 vs 8.17).
I posted previously here regarding the relationship between Cyclops settings and Radiance parameters. There is also more info on the sampling settings in the Cyclops documentation.
You can ignore the Actual/Expected batteries you are seeing on the canvas. We only use those for testing in our internal CI/CD pipelines to ensure that the scripts produce the expected results as we implement new updates.
