The setting of evalglare

Hi list,
I am learning the “evalglare” to analyse the glare sources in the HDR images. I have met several problems about the masking: I have found some ways to mask the HDR images and there existed some difference between their results like the glare metrics and the glare sources. And which method is the right or recommended one? Besides, is there any difference between the option “-A” or just analyse the masked HDR images directly(still using the -vta -vv 180 -vh 180) ? Any advice is appreciated!
method 1:





method 2:



method 3: