Correct me if I'm wrong: masked area = some part of the sensor that is physically blocked from light but still records brightness values. By definition, this would be the darkest part of any exposure.
Allow me to correct you. Although the OP certainly implies that is the case, the 'masked area' the OP is referring to is not physically blocked from light - lenses project an image circle, and for an EF lens, a FF sensor is the largest 3:2 rectangle than can be inscribed within that circle, such as the red box below:
So, what he's calling the 'side masking area' - the regions to the left and right of the red box - are actually being illuminated by light from the lens. Why, then, are those regions 'black' in the RAW file? Because those regions of the sensor are electronically turned off (technically, set to an arbitrary value). Given that, I remain unconvinced that the method described by the OP is valid.
Those regions are being illuminated, but what's there to record those values beyond the area of the sensor?
Does the sensor physically extend beyond the 36x24 area and they disable the band of pixels outside the perimeter? I guess that makes sense from a design standpoint (to ensure you get the full desired frame, you cut sensor a little larger and hold back the outliers).
I'd never really given a lot of thought to sensor design before. Thanks for bearing with me.