Holger wrote:
tn1krr wrote
echelonphoto wrote
MY pumping is occurring at wide apertures like 2.8 That really does not exclude PDAF giving up and falling back CDAF. It is all about light and horizontal contrast; if using flex spot small only single PDAF spot used does not get enough light/contrast say hello to pumping/pulsating. The PDAF points in sensor are not cross type and are very small (and A7RIII has smaller pdaf points than A9 for example) compared to the one used in DSLRs so they will "give up" sooner than large cross type point in the DSLRs do.
The strength of on-sensor PDAF is using several points (they are rather densely grouped) so using Flex spot Medium/Large or Expand Flex spot will likely to help in many cases. These are not DSLRs, so best results are not obtained by using them like DSLRs which often is using center-spot only.
Like I commented above: put the camera to tripod and point it to a low light object any test will show that larger spot is likely to pulsate less than small and you'll find a spot that will start pulsate when you close the aperture step by step. This should have been similar with the A7rii. But in my experience I only rarely saw that, usually in bad light. Here the user reports even tell it is happening in o.k or good light. I am not sure that flex spot small does mean that you are only using a single PD point pair on the sensor. Could be, but that would be very inefficient, being half masked pixels, when I have time I will calculate it.
I'm not claiming this has not changed at all from A7RII. A pile of AF related things propably have changed since there supposedly is not 425 CDAF points (whatever that means) instead of 25 we had in A7R II etc. I did not notice this in 3 weeks of active use of A7R III until I specially looked for it; all I saw was AF-C that was overall heaps better than my A7RII and passed in flying colors test where subject walks towards me in quite low light; same test where my old A7RII rarely got more than one shot out of 6 or 7 in decent focus.
The PDAF grid on both A7RII and A7RIII is 21 x 19 points and the visible AF point grid that gets displayed in LCD/EVF seems to be 15 x 15 (when Phase Detect AF mode is used with adapted glass of wide/zone AF mode) so what one sees on the screen does not map 1:1 to actual PDAF points. I'm not so sure this matters actually so much since we've already established that smaller aperture, smaller AF point and lesser light/contrast leads to this pulsing behavior.
I'm also not convinced yet there is anything to fix with FW upgrade; we have not seen good or heard of a side-by-side comparison between R2 and R3 or observed bad behavior in a very sensible use case; to me using small AF point with AF-C on a static low contrast subject is not one; although that is the way my Canon-using buddy shoots almost exclusively. If we for example see this pulsing behavior on stopped down AF on a subjects face in a case where A7R II still works perfectly ok I'm gonna be the 1st one to hop into "fix this ASAP" train.