I tried a mirrorless. Have you?

MarBa wrote:

Roland Karlsson wrote:

MarBa wrote:

unhappymeal wrote:

Every time I feel nostalgic and want to pick-up a KP or K-1, I have to remind myself how awful calibrating lenses is and how much harder manual focus using the OVF is compared to an EVF.

I hate this too ... But Pentax could improve the process! This could be entirely automatic with little user input. Simply use LV to focus compare this with AF measurement and calculate correction .. they can even take photos and compare sharpness in the center etc ... All this could be automated for calibration of different focal lengths, or AF points.

So .. yeah .. with bit of software development it could be an easy proces .. put camera on a tripod and press a button!

I am a programmer, and I have been thinking about this one. If I can write such a program.

There are some problems.

1. The communication protocol is too simple. It do not support what you want/need.
   You cannot do micro movements.
   You do not really know where you have focused.
   You cannot set the adjustments.
2. There is really no way to calibrate zooms over the entire interval.
3. Focusing is not accurate. It varies from try to try.
4. Evaluating if you are in focus is not easy.
5. Focus shift makes it sometimes nearly useless to calibrate for full opening and then use F5.6. They need separate calibration.
6. How about calibration not only center spot?

Number 1 is a blocker for totally automatic calibration I would say. But you could do it by manually set the different calibration steps, and for every step automatically take a number of pictures and analyze the result into a number. Calibration with highest sharpness wins.

Number 2 is a blocker for zooms really.

Number 3 and 4 makes it difficult.

Number 5 makes calibration of some fast lenses almost impossible. Actually even AF.

Number 6 I think is something that those that calibrate chose to ignore.

Nikon was able to do this! I see no reason why this would not be possible. They managed to get astrotracer without GPS... Calibration of lenses seem easier than that.

I always wondered if Nikon got a patent for that, which would explain why nobody else has tried it.

There can be multiple correction numbers stored for zooms, different AF points or even apertures.

LV can focus really well with contrast focus. So it is NOT a problem to find out if things are sharp. And then, when things are in focus using LV, you just need to read out what the AF module is reading and use this information to calculate a proper correction.

I can't be worse than when you do this manually by trial and error and have just ONE number per lens! The current implemention is terrible if you think about it.

#3 is a really big problem.  I have a Tamron 70-200 that I never use because I can't get consistent focus out of it.  Even on the K-01 which does CDAF I'll get one shot front focused and the next back focused.