Should one interpolate between surrounding pixels to get the correction at a point? It seems one need not. I ran a test to see the effect of interpolation as a function of distance from the center. Near the center interpolated and uninterpolated magnitude changes due to vignetting correction match to well less than 0.01 mags and thus it does not matter there (points 8 9 10 11 12 below). They differ by a varying amount away from the center but the interpolated calculation is often wrong when there is a sharp change (points 4 5 7 8 below). Points too close to the edge are unreliable and neither method can do much (points 0 1 2 3).

ps version