I have used the ‘rtrace’ program to calculate the RGB values on different points at work plane height in different settings and I have a large sample. I have used Python to simply automate the process of creating and executing all the bash and RAD files. I tried re-running some simulations and I am confident that I never changed any of the parameters that I used in my previous simulation and I have realized that sometimes simulated values are inconsistent with the previous ones.
I would appreciate your insight and any tips on this issue.
The default rtrace -u+ option applies a random-number generator that is seeded by the clock and therefore sends different samples each time it is run. If results consistency is more important to you than accuracy, you can set -u- instead, but it is better to increase the other sampling parameters until your results are within your desired accuracy tolerance, instead.
If you are using an ambient file, that could also lead to your first simulation results differing from subsequent results because the file is populated during the first run. Without seeing the command you’re using, we’re just speculating.