because i have checked these 3 photos quite a few times, and i have seen unfiltered versions of all the ones without gnd usage because i did bracket to gauge optimal exposure before filtering. #2's horizon line was right smack in the middle, i just pushed the graduation line to the middle of the lens, for #3, i used live view to ensure the graduation was at the right part.
i am very puzzled as to where you perceive the filter graduation ends, because i am looking, and looking and looking at #2 and #3, and i can't imagine where you think it ends, other than the logical area where it should end, of course.
in fact, #2 and #3 are not even good displays of soft grad usage, unless you really **** up the placement or use a hard grad, wrong placement of graduation will not cause much visible effect.
#1 is the test of whether the graduation is obvious -
do you see an abrupt loss of detail in the temple? if not, then it is fine.
what you are paying for for more expensive filters is durability, build, and more variety, as well as branding. also, the neutrality, if it is really neutral. user error is not the filter's fault, and when the user error extends to the point where the user is seeing error in everyone else's shot because of the equipment used, then he/she needs a wash of cold water.
maybe, next time i will declare that i have bought an entire arsenal of singh ray filters, then all will GOOD GOOD GOOD very nice graduation for the filters. :bsmilie: