To close the VignettingCorrUnknown subthread, here are some findings (none of which is really related to this topic).
The routines which gather data for this tag only exist on cameras with native raw. There are several tables that are used for this tag and not referenced anywhere else (that means in-camera processing does not use them).
The tag's content depends on the current zoom step (unsurprisingly), other parts depend on aperture related propcases. Some of the data comes from uncached RAM (!) - the base address of that RAM area is also found in the commondrivers task and elsewhere.
That's it.
