I remain unconvinced that one stamp is actually wider than the other. The bottom stamp is tilted a bit to the left, giving the appearance that the right edges do not align. To be convincing, you need a template that show the alignment at both top & bottom.
Incidentally, I do have both a 563 & a 563b ... the widths & heights are slightly different. Note that my images, like yours, are slightly askew. But hopefully the differences are a bit more obvious.
Added: In my experience, you can frequently detect the varieties on special booklet paper by carefully measuring a stamp's aspect ratio (defined here as long-side/short-side). The aspect ratio for 563 is usually 1.16, for 563b 1.12. These can vary by a
hundredth or so in either direction. A stamp with AR of 1.14 would effectively not be identifiable via this method. Note that this method can be used to identify stamps without accurately knowing a stamp's dimensions. You can even use computer images to identify such stamps being sold on
eBay or HipStamp.
Added: I measure an AR of about 1.17 for both of your stamps. So your 563b is probably a 563a.
Added: The AR of rogdcam's example from the PF databse is 1.12.

