A short, subjective, comparison for you:
I have the S5. I've never expected much from its C-AF, since I am also used to how Panasonic's CDAF M43 bodies behave in C-AF. As such, I pretty much always use S-AF (but it would be nice to have really usable C-AF for some bird work).
I've never been unhappy with the S-AF on the Panasonic M43 bodies. It's always worked very well for me.
However, the S5 S-AF is (again, subjectively) slower to AF and misses sometimes, where the M43 versions never do. This makes the S5 a little less pleasant to work with in my usual fashion, but with some adjustments to settings, I have gotten it to be acceptable.
Fast forward to the S9. (I have not used the S5II). ALL AF, both S-AF and (as expected) C-AF on the S9 is noticeably faster and better than the AF of any type on the S5. The S9 finds and locks on to things even in S-AF faster and more accurately than the S5 does. Of course, I am not sure if those differences are because of the PDAF, or if there are other CDAF tweaks Panasonic did But something is better.
Similarly, when I had the G9II last year, for about a month, I found all of its AF to be better, not just the C-AF. But, I didn't use that camera long enough to get any kind of detailed evaluation going, just a subjective one from use.
So, I don't know what Panasonic is doing with their PDAF bodies, but they do seem to work better in AF overall, not just C-AF...in my own personal experience with them.
-J
Panasonic S-AF has always been quite acceptable for my M4/3 camera bodies. I have not bothered with C-AF mainly because I have not ever needed video. But I can see some advantages for C-AF in tracking moving subjects.
My L-Mount S1 has been mostly used for adapted EF mount lenses - in which case the performance has been as per my well publicised, but aging, tests of multiple EF adapters. Overall the EF lens performance has been adequate and I have tended to not use the Sigma MC-21 as third party brands offer a wider range of compatibility.
The question now is whether or not these adapters have improved the performance of EF lenses on Panasonic bodies that now use PDAF focus assist in some form or another.
This related to the S5II/x, S9 and for M4/3 the G9II. I only own the latter.
Only Metabones (which does not make for L-Mount) seem to have specifically addressed PDAF support (in their M4/3 product). Starting at v4.10 on 12 July 2024 (installed) and an update as V4.20 on 22 November 2024 which I have not yet installed.
Viltrox firmware for M4/3 (M2 EF=M4/3 focal reduced) is fossilised at v4.0 last updated at sometime before June 2022. Their L-mount firmware is at v1.2 which I have not installed but suggest that it does not cover Panasonic PDAF focus ability.
Sigma MC-21 is at v1.4 from 22 December 2022 and none of their updates mentions PDAF support. The most recent update was pre-Panasonic PDAF.
Consequently we can only presume that neither Viltrox nor Sigma thought it necessary to adjust their adapter firmware for PDAF support.
Metabones who seem to work hardest at keeping their firmware sweet and trouble free has made two adjustments to their software which in part involve PDAF support. But this is M4/3 country and does not apply to L-Mount.
Testing the Viltrox and Metabones adapters with the "glory be" huge Sigma 105/1.4 DG lens on my M4/3 G9II with S-AF indicates that the Metabones firmware is giving the usual "double-clunk" reaction of CDAF. Finding focus reasonably quickly then "clunk-stop" and backup to best found focus point "second-clunk".
Using C-AF it seems to almost casually glide into focus when the shutter is soft pressed and then tries to glide into focus again. But draws the line as gliding into closest focus with a test jar of flowers. It needs a new shutter soft press to bring the vase and flowers into focus - which it again does smoothly and at measured pace. The occasional "clunk" but my general impression is "clunk-free". My guess it is using C-AF with PDAF in measured transition as perhaps the more video interested amongst us would appreciate. The S-AF "double-clunk" presumed CDAF focus is certainly faster. The C-AF presumed PDAF focus is smoother but I don't have a convenient moving subject matter to test to see if the "measured transition" I was seeing was actually "measured-enough" to keep up with the motion and provide adequate re-focussing.
But I don't have an L-Mount body that features PDAF to see what native EF-L electronic adapters can offer. Certainly there doe not seem to have been any effort by either Sigma or Viltrox to update their adapters to take advantage of the PDAF facility - perhaps it was not necessary for L-Mount?
But Viltrox has also not updated their firmware from V4.0 for its EF-M4/3 adapters. Both S-AF and C-AF work but my impression is that the Metabones product does work better.
Testing again with a happy more compact lens such as the Canon EF 50/1.8 STM might make for another perspective.