Originally posted by Rich_A It is my understanding, and
supported by Chuck Westfall's (Canon's US Technical Adviser) article here a little more than half way down the page) that each lens is calibrated at the factory and an MTF table based on a standard algorithm is flashed onto the ROM chip after it is tested. This is done under controlled conditions but often only at particular focal lengths (such as only 24 and 70mm on the 24-70mm zoom). If the lens is sent in for repair, this table can be recreated through retesting and re-programmed (re-flashed) onto the chip.
Interesting. This seems to me like way outside what anyone should expect a lens to be able to do, but it does at least suggest a sense in which it is theoretically possible. I don't see anything in the link that you referenced that suggests this is actually the case, though - I don't see any claim that MTF numbers are used in AF calculations, or that the "calibration" process involves altering these tables.
So in the absence of a credible explanation of this process, and evidence that is it anything but an extraordinarily rare situation, I'm still unconvinced tht any sort of "reprogramming" would be likely to have any effect on most focus problems. My assumption remains that *most* focus problems are due to the camera simply having misaligned AF sensors. Of the small number of problems that are lens related, most of those would be purely mechanical in nature. Of the small number of lens problems not mechanical in nature, msot would be optical in nature. Leaving only a tiny, tiny percentage that would be MTF-table related.
Last edited by Marc Sabatella; 11-12-2009 at 09:20 PM.