how to use pixel math to lower the ADU count on a master flat?

[prefetch]

i think my flats are over exposed.  the corners of my calibrated image are too bright.

i'd like to experiment and my thinking is that if i can use the pixelmath function to artificially change the exposure of my master flat frame by doing a blanket reduction of ADU count and seeing if the corners of my images start to look more normal.

is this a reasonable approach, and how would i do something like this with pixelmath?

edit: i think i figured out that using something like this: flat_FILTER_O3_BINNING_2/5 would divide the ADU by 5.

btw, here is an example of my master flat and a master light from my o3 filter.  any advice on how i can help my corners?

master flat file:https://dl.dropboxusercontent.com/u/4404304/flat-FILTER_O3-BINNING_2.fits
master light file:https://dl.dropboxusercontent.com/u/4404304/light-FILTER_O3-BINNING_2.fits

[prefetch]

i think i found the solution to my problem:

http://pixinsight.com/forum/index.php?topic=6210.msg42200#msg42200

[Carlos Milovic]

The best solution is to take the flats again.
Linear modifications of the flat won't do anything, because they are normalized before applying. Normalization means different things, depending on the software (and options they give). The most common choise is to set the mean value at 1.0, so there are values greater and lower to 1.0 that either attenuates or amplifies the signal.
A nonlinear adjustemt with curves could do the trick, but wou should know exactly how to correct the nonlinearity of your sensor at those high intensities. At the end, the best thing is to take the flats again.

[prefetch]

okay, i think i got it figured out.  my ccd sensor was not aligned with my square filters, producing lights/flats that had screwed up corners.

more details here: http://www.cloudynights.com/topic/495046-cropped-jellyfish-too-much-vignetting-solved/

not a PI problem at all - an image train problem.  so many variables!