This has never been an issue for me as with LF I meter independently and apply the filter factor BUT I have recently been shooting a fair bit of 35mm SLR using filters for B&W and metering thru the filter (as speed has required).
When I initially got the filters (a Hoya Orange 'G'-med orange and a B&W 091 - a wratten #29 deep red) I checked the listed exposure factors and unless I am very mistaken it was factor 4 for the orange (2 stops) and 8 for the deep red (3 stops). This seemd to make visually very little sense as the orange is not even close to the deepness of the deep red. Upon metering thru the lens with the filter ON, I lost one stop with the orange (well below the 2 stops of the filter factor) and 2&2/3 stops with the deep red (ie very close to the 3 stops stated. I have also read that meters (mine is and eos 3) can be more sensitive to red thn other wavelengths resulting in not enough increase in exposure when red filters are used with TTL metering, but in this case the camera was close to smack on with the deep red and a stop out on the medium orange.
I realise that bacground would have an effect so chose a neutral grey wall to do the initial test.
As a fudge I have bracketed a stop ether side as well as applied exposure compensation where required for backlit scenes etc to hopefully keep me on 'zero' and thus allow the +1 frame to save teh day should my orange filter really need 2 stops.
Personally, I doubt that this orange filter needs two stops, but also doubt that one would be enough. As a related aside I have stopped giving a full 2 stops to my lee orange polyester filter as i kept marginally overexposing and now give 1.5 stops.
Any idea what is going on here? Why does the deep red meter seemingly accurately (I have never used a deep red before so should be interesting...) when one hears that metering with deep reds is notoriously erratic TTL and so far off manufacurers exposure factor when using the orange?