a simple film speed test question. filter or no filter?

having a brain fart. I should know this but my mind is blank.

I'm gonna do some simple EI testing for the films I use but test for reversal processing (which look best projected). this being unscientific, basically shoot the frames, mount em and then project them. I was wondering if I should shoot without any filters, or use the filter/filters I would normally use when shooting. some films I use a yellow and others an orange.

so, to filter or not to filter, that is the question.

my $0.02:
If you're always going to be using the exact same filter for this film in the same situations and same light all the time, then you may as well put the filter in place.
If you might one day use no filter, one day use yellow, one day use green, then it might be best to work out the ei with no filter, and then work out the filter factors individually.
The first way, you get an ei for the film/filter combo in a certain light, and then you'll get a different ei in a different coloured light (think a blue-light dominated noon-landscape vs a yellow-dominated indoors or sunset-shot). Or the second way, you get one ei for the film in all lights (presuming a nearly-flat spectral sensitivity curve) and then have different filter factors for different colour filters in different light. Either way, it's a whole lotta work (and why most of regular people just use 'ballpark' figures for B+W negs and hope it's within the latitude, but I know reversals are a bit tighter...)

IMO, all film speed tests are without filter. Do that first. Coming up with filter factors afterwards is secondary. That should be the easy part after establishing the former. No other way makes any sense. Any other way, you're starting out by introducing a variable, which to me is not sound procedure. JMO