Originally posted by jaad75 Yes, the AF-C is much worse, but still good enough to nail some basic situations
Yes, that's probably a fair way to put it. The AF-C yields a good percentage of properly focused and wonderful sharp images. It's just the failure rate which could be lower. I'd say it is between 20% and 50% (for f/1.8 low depth-of-field usage). Someone shooting a dog possibly wouldn't bother, he/she would just delete these and enjoy the sharp ones. But we (my fashion blogging model and I) are picky with regards to the expression and the flair of captured special moments. After every shooting there are always these disappointments, when a favourite cannot be used due to OOF blur. That's why I'm very ambitious with the keeper rate.
Yes a simple solution would to go for a pro body such as Nikon D750 or higher (or equivalent Canons). That'll bring down the failure rate to a small and reliable one-digit percentage. But it's much more money than I want to spend on the issue resolution.
Originally posted by jaad75 Optyczne/Lenstip testers had serious AF problems with their copy of the DA50/1.8
Yes I saw this as well before my purchase, but attributed this to the generally less capable AF of the K-5 (original) they used.
I'm impressed with my sample optically, no decentering (as if a Limited), good performance wide open. There's some focus randnomness at AF-S, but not severely, maybe just within the range -/+ 3 in AF fine adjustment terms, up to +/- 5 are rare outliers.
This is not perfect but in line with my other Pentax lenses.
I suspect that a main cause of the AF-C issues may be, that the AF-C may not be sensitive enough for small reductions in the object distance. I.e. the defocus threshold, which is required to recognise a distance change and trigger an adjustment, may be higher than what the shallow DoF can bear. I remember faintly from the times when the D7100 came out in 2013, and a couple of Pentaxians switched over from the K-5ii(s) to the D7100, one reported that the D7100 AF-C would be more sensitive towards small distance changes, and issue more frequent focussing updates.
If we then add a minor amount of general mechanical randomness on top, maybe this explains my issues (which may be applicable only to shooting subjects and styles comparable to mine) completely.