You really can't compensate for the smaller sensor area with software that much.. considering how mature the 808's jpeg processing is, its pretty much maxing out what the Toshiba sensor can do.. so.. even if they get to that point with the 1020 in terms of processing, it will always be noisier overall.
And in terms of sharpness.. if you pump up the sharpness on the 808 to the max its pretty good, I doubt that the 1020 will be that much sharper.
How accurate do you think that 25% number is ? I am yet to see a high end camera use a BSI sensor and produce a decent image.. if they were 25% more sensitive to light, don't you think that we would've see at least one DSLR with a BSI sensor until now..
At this point, its just a way for the OEMs to be able to use the same small BSI sensor but pack more pixels into them so the consumer can be led to believe that they are buying a better camera. It also sounds good in the marketing materials..
Just like in the past, Nokia can always provide an update to fine-tune the overall result. Be it a bump or decrease in saturation, NR, sharpening etc or whichever they see appropriate to improve the result via firmware/sw update. If the response of initial field test (meaning reviewers and early users) are unfavorable, it is almost guaranteed that they will immediately provide the sw tweak. .
As for the sensor FSI of 808 vs smaller BSI of 1020. It is not just for marketing materials. BSI is the new technology in camera sensor and has been in development for more than 3 years now. That's the reason why you don't see DSLRs with BSI, yet.. The 25% advantage on light sensitivity of BSI I mentioned is actually a conservative value. Toshiba even claims that it can max upto 40% increase compared to FSI with similar sensor size.
A short summary, FSI needs larger pixel pitch like 1.4 micron to give space for "metal wirings" between lense and photodiode sensors, whereas BSI do not have these wirings in between since they are placed at the "back" of the assembly stack (hidden) and therefore no light obstruction. In essence, FSI is less "efficient" than BSI based on light capture per area because of the metal wirings. Since BSI sensors do not have this wiring layer, the pixel area can now be optimized to a much smaller size like 1.12 micron, still with the same light sensitivity.
Again this is not a gimik, it is a mature technology that is widely used. Lumia 92X series all have BSI sensors. Even the Sony RX100 II, which is very well known for its amazing quality, is boasting the Exmor R BSI technology, a 40% improvement from its earlier version RX100 with FSI in terms of light capture sensitivity. This new BSI sensor is rumored to be the same 20MP sensor of the anticipated rival of 808, the Sony Honami (Android).
And from Toshiba themselves, makers of 808 and L1020 sensors, here is what they have to say about their BSI development.
"BSI brings new levels of responsiveness to CMOS imaging. Lenses are deployed on the rear of the sensor on the silicon substrate, not on the front, where wiring limits light absorption. This positioning boosts light sensitivity and absorption by 40% compared to existing Toshiba products, and allows formation of finer image pixels."
Toshiba : Press Release 27 Oct, 2009
Therefore comparing sensors with different type of technology based on size alone is rather pointless.
But it is very possible that the smaller sensor of 1020 with BSI has the same light capturing sensitivity compared to a much larger sensor of 808 but with FSI.