So does the battery last much longer when using MF only?

AF-S has been accurate and quick for a long time by any brand. AF-C is something that cameras are still working on even now. Panasonic is behind the pack in this which is reflected in the review.

Who is your comment directed to? To me? If so then I haven't used it. I was referring to oztuerts comment above where he had the K1 for a few months and sold it due to inconsistent AF-S performance.

Rob890, I used K1 with 50mm f1.8 AF, 70mm f2.4 Limited, 28-105mm (new kit lens) and 7-8 manuel lenses. Even in good light there were just too many AF errors and I always used single AF and center point. I tried AF microadjustment but didn't work. With the limited lens, I saw AF errors even at live view!!!
I had similar experience with K10D, K100D, K50, K5 II and K3 so I sold K1 and haven't bought any Pentax since then.
Maybe newer lenses and bodies are better but I will not spendy money on Pentax if I do not see visible improvement.
Maybe Pentax will take care of AF in mirrorless?

You are confused, because you have a per-conceived idea of what the performance should be , based on brand image etc. The frame rate drops when AF tracking on the S1R, 6 FPS is the theoretical max frame rate, not the actual frame rate when AF tracking is engaged. On the EOS R, the specified frame rate is 5 FPS, but practically it drops to 2.5 FPS with AF tracking, and this is not as much the case for the Pentax K1. The same goes for the S1R, especially with DFD and AF tracking. Boring to have to explain this.

Pentaust, S1R's theoretical max frame rate is 9 fps, not 6. 6 is with AF-C.
EOS R's fps with AF-C isn't 2.5, where do you get those numbers from? It is 5fps with AF-C. It doesn't drop to 2.5 fps. I have EOS R now so I can tell you right away.
See S1R:
https://www.imaging-resource.com/PRODS/panasonic-s1r/panasonic-s1rA6.HTM

Where do you get your numbers from? The Canon official EOS R user manual says 2.5 FPS when AFC is engaged and RAW 14bits. And this was confirmed by Tony Northrup when they evaluated the EOS R and Z7 for wildlife shooting. Do you research, read the details, and try the cameras. I've read the user manuals for each camera models, and while the spec sheets numbers look rather good, when digging the manual tells a different story. I can also tell that the Pentax K1 shoots 6.6 FPS by omitting what settings are used.

Just look at the Z7 spec, in DPR spec sheet, it says 9 FPS. Now read the Z7 official Nikon manual it says 5 FPS with mechanical shutter and full frame uncompressed RAW... so that's only 0.6 FPS more than the Pentax K1 in the same conditions.

Just look at the Z7 spec, in DPR spec sheet, it says 9 FPS. Now read the Z7 official Nikon manual, burst mode table page 84, it says "Continuous H, 14bits RAW, mechanical shutter 5 FPS, electronic shutter 3.5 FPS so that's even slower than the Pentax K1.

Just look at Canon EOS R advanced user manual. DPR spec sheet says 8 FPS. Canon manual page 145 says: "With Servo AF shooting:When [AF operation] is set to [Servo AF], the continuous shooting speed of max. approx. 5.0 shots/sec.". 5 FPS is the maximum (ideal conditions), in practice it drops further, down to about 2.5 FPS.

Just look at the S1R user manual from Panasonic, page 127., Burst Rate: 9 frame per sec. in AFS or MF modes and 6 FPS in AFC mode, 5 frames per sec. when use electronic FC shutter. And there is a note: "*The burst rate may be lower depending on the settings for recording such as [Picture Size] and focus mode.

ozturert, I took the lastest versions of manuals. Why are you suggesting I used an old manual? Strange. And basically, you are now writting that the frame rate isn't as high with AF tracking. So you are saying the same as I said earlier. Pentax K1 does 4.4 FPS with AF tracking, that's not far off 5 FPS and 6 FPS, in all cases those cameras aren't best for sports / dynamic shooting. What were you trying to demonstrate?

Pentaus, you said "Now read the Z7 official Nikon manual, burst mode table page 84, it says "Continuous H, 14bits RAW, mechanical shutter 5 FPS, electronic shutter 3.5 FPS so that's even slower than the Pentax K1."
I said with the latest firmware Z7 can take 9 fps with AF with AE.

Yes, I also wrote that the actual frame with the S1R and EOS R, are lower that specified. The specified frame rate for those mirrorless models are the max. The EOS R actually burst rate when tracking is between 2 and 3 frames per second, while the spec says 5. Is it a property of online forums that people argue using what's they've been sold and never used, against actual performance experienced by someone who used the camera? So, you've won, stay with your beliefs. As far as I'm concerned I'll focus on my practice, as it seems online discussion at pretty much a waste of time. Ciao.

The pupil is a circle with hazy, unclear borders. It's not clear to me how anybody is checking focus on that, unless you are interested in inspecting the capillaries on the surface of the eyeball. To me, the eyelashes are the instinctive thing to look on.

When you were in love as a kid did you look into her eyes or stare at her eyelashes.....

Cheap and false. Look into the mirror, inspect your pupil and tell me what you want to focus on.

And by the way, sure the eyelashes can be sexy.

I misspoke, I said pupil but was actually thinking iris....my grade 10 biology course was many moons ago:) Either way I stand by my point, the eyes are the focus, no pun intended.

I was looking for a photo next to the photo credit and found none, but if you just have a photo at the top you are doing ok.

Panasonic US just dropped the price by $700

@shadowwalker - "PDAF is constrained by the silicon. What is there when the sensor is made, is all you get."

no, updates on sony firmware that added things like real-time tracking and animal eye-af prove that there is a lot you can do with ospdaf.

"That's why PDAF mirrorless cameras use CDAF for the last bit of accuracy after the PDAF gets them close."

the a9 is the best ospdaf camera on the market, and there are many situations where it doesn't use cdaf, with xlnt af accuracy... for example, with some adapted lens situations, there is a menu item that forces you to choose between pdaf or cdaf, so it's easy to actually test that, which I have done.

"DFD/CDAF OTOH is completely software based "

Panasonic proprietary dfd is not desirable for adapted lens use, in part because it forces you to use cdaf on lenses that weren't designed to be used with cdaf… companies use ospdaf because it is far superior to cdaf.

Lets start by saying the most important: When the AF of a camera is good enough for you it is just that, good enough for you, no need to want better then that.
Then the AF systems. ALL AF systems are limited by hardware. The DFD/CDAF is limited by processor speed and the speed of the sensor just like PDAF. PDAF can be used in a hybrid system where it uses CDAF for the last perfection.
Both systems hang on software algorithms to create the results, bot systems can be software upgraded to some point (look what Sony did with their A7III) .
In the end all cameras can focus fast enough for most people, not many needs 20 FPS with tack sharp pictures...

CDAF limitation is not the processor speed but the algorithm itself. The only way it can tell the distance is to focus in and out, which takes time and is even mechanically limited. PDAF can tell the distance instantly and very accurately, and it also requires less cpu, and doesn't need to actuate the lens motors, saving previous battery. TOF has limitations, the resolution is very poor, it's less precise at judging distance than PDAF and CDAF, but at least it doesn't need to do focus hunting, and works great in the dark. Probably the best focus system in the future will be PDAF assisted by TOF for low light.

Um no, you are almost certainly wrong. The history of computer-assisted automation has shown over and over again that as processing power grows, restrictive hardware implementations invariably give way to flexible software implementations.

PDAF is a restrictive hardware solution that is very fast but not very precise. That lack of precision is why it will go away and something, probably a descendant of CDAF will rule, possibly working with TOF to reduce bellowing. The only thing that I'm certain of is that aside from adding more sensors, PDAF hasn't improved in years. It's at the end of its development cycle and still far from perfect. Better, more customizable software solutions WILL crop up within the next few years.

Virtually everything you just wrote is wrong to the point that it isn't even worth replying.

Enjoy your camera, whatever it is. Take beautiful pictures that give you artistic fulfillment, and don't worry about the technical issues that you don't understand.

Okay, fine. Most of the answers to your statements were already presented above, but since you didn't read them, here goes. If this takes more than one post, I apologize in advance.

PDAF is not accurate, it is not new, it is old technology that is not getting better. The current implementation of fast cross type sensors goes back over a decade. The first implementation that I remember was on my Canon EOS 40D from 2007, but there may have been earlier ones. The only changes to today are placing the sensors on the main camera sensor (too many uses of the word sensor, sorry), and providing more of them. More sites are a refinement, not an improvement, and putting them on the main sensor is problematic as regards picture quality and light gathering.

PDAF is not accurate. Or to be more correct, it can be accurate if your lens is perfectly calibrated to your camera. Unfortunately, this doesn't happen in the real world. Lens calibration shifts and PDAF loses accuracy. On mirrorless cameras, this is normally cured, for people that care, by using a hybrid system that uses PDAF to get close and CDAF (there it is), for fine focussing. On DSLRs we had to use micro-focus adjustments to over come this PDAF deficiency (and it was/is a royal pain in the butt). BTW - the CDAF solution doesn't work on DSLRs because, except for live view modes, there's a big mirror in the way, thus the micro adjustment solution.

So, for all intents and purposes, you can think of this as analogous to focusing a microscope. PDAF is the rough focus knob, and CDAF is the fine adjustment.

The fine focus in my last post, is based on the fact that CDAF doesn't require a calibrated lens. It is literally doing the focus from the image on the main sensor. This will ALWAYS be superior to any other method for absolute photo focus. The issue is getting CDAF to be as fast as PDAF for video and continuous focus.

This is where my statement about faster processors and better algorithm come in. The trends in computer development that I detailed half-a-dozen posts up are facts, not opinions. Generalized programmable solutions ALWAYS replace rigid hardware solutions as soon as sufficient software driven processing power becomes available. We are seeing the onboard image processing power of all brands of cameras increasing every year. This will happen.

And finally, your comments about TOF. You are completely correct, today. But unlike PDAF sensors, this is a technology that is being actively developed and improved. It has the promise of being the only focusing engine in the future. But today, that is only promise and may not come to pass.

But guess what? Even if it never gets better than a 1 cm accuracy, it would make a very reasonable replacement for PDAF as a fast mechanism to get a near focus so that CDAF can do the final refinement. And guess what, with your 1 cm accuracy of TOF, the TOF/CDAF hybrid would be able to track motion very accurately without any noticeable focus bellowing. The TOF engine would be off the main sensor, so would not interfere with images and would be upgradable or even left off of different SKUs without changing the main sensor.

Pretty much superior to PDAF in every respect.

Sorry this took so much space, and I really do hope that you have a great day.

Excuse any typos because I typed fast.

shadowwalker - "PDAF is not accurate. Or to be more correct, it can be accurate if your lens is perfectly calibrated to your camera. Unfortunately, this doesn't happen in the real world. Lens calibration shifts and PDAF loses accuracy."

no, that is confusing dslr pdaf with ospdaf, which never needs calibration, because with ospdaf the focusing is done with the same sensor that the picture is taken on.

You are correct, but not entirely. You are correct that the PDAF sites are located on the main sensor, but you misunderstand what that means. They are on a totally separate logic path and have no knowledge of the actual image data. Also, the calibration issue is not in the camera and never has been. The calibration issue is in the lens, or more specifically the interface between the camera and the lens.

The way it works is that the PDAF system decides on a focus point and determines the distance–this is mostly accurate. Then it tells the lens to move to the correct focus for that distance, and here is where the fun begins. If the lens is not calibrated exactly, it will front or rear focus at that point. That is the reason for micro adjustment or a PDAF tuning or whatever solution is used for the final correction.

That single distance is both the strength and the weakness. There isn't any hunting or uncertainty, but there is also a reliance on precision that is probably not present.

@shadowwalker - "If the lens is not calibrated exactly,"

ospdaf does not need lens calibration, as I already explained.

ospdaf is extremely accurate on my a9, including with adapted lenses... that is actual experience, and pretty much everyone who owns an a9 agrees that a9 ospdaf is far superior to any cdaf/dfd system on the market.

that is the bottom line, it proves that ospdaf is the best.

I'm getting really bored with this.

The A9 uses a hybrid focusing system with 693 phase-detection points which cover 93% of the image area. These do the OSPDAF focus, as I described above.

Additionally, there are also 25 contrast-detection autofocus points (CDAF) to fine tune the OSPDAF, also as described above.

Reread, or really read for the first time, what I wrote above. Lens calibration and micro adjustment is for DSLR PDAF systems. CDAF is used for the same purpose on OSPDAF systems.

And just to put it bluntly, there is no functional difference between DSLR PDAF system and mirrorless OSPDAF systems. Having the sites on the main sensor does not give them any info about the image. That is not how PDAF works. OSPDAF simply means that they don't need one surface for the image and another surface for focusing. It simplifies the light path and it prevents alignment errors, but that's it.

The A9 uses contrast detection to tune the OSPDAF output, as I described. Period!

shadowwalker: 2 mistakes, first you said that the A9 is doing contrast detection to tune the PDAF, that's not true. If you use the A9 in AF-C mode, it's able to focus without using CDAF at all. Second, you said it was 25 contrast detection points, that's not true since firmware 5, there are now 425 contrast detection points. Yup, they can be added in software, since it's just software.

If the eye lashes are in focus but not the pupil, it could be that you are photographing an Homo Neanderthal model. We don't do it very often, only once in a while.

But then again, maybe portraits with only a thin slice of the model's face in focus will be the next big thing. We may be on the cusp of a new art form that will eclipse even "Lifestyle and People" photography!

I have - but once in a while I really have to hunt for that elusive setting. That said I have not used the latest menu system, but it looks a lot like the old one to me. I have always felt that the (fn) menu UI was far superior to the main menu - and Sony should use it as model for the main menu UI.

Exactly, the customization of the 3rd gen models is incredible, I'm not even entering the menus ever again, except to go to "My Menu" sometimes where I put 6 items. Everything else is mapped in the Fn menu and all the buttons on the camera. Sometimes I have to use a Canon or Nikon and it's actually a step backward to have to enter the menus, yes they're better, but it's even better when you don't need to use them.

I'm sure it is, but not when it comes to video.

MikeRan, I know. I have an A7R3. It is amazing AF'ing indoors. I imagine the A9!!. But I'd loved to have double axis tilt screen like the one in Panasonic or Fuji.
Maybe the A9II will finally have one. Fingers crossed.

Irakly, I'm glad but wonder that you have no problems with AF indoors. I have watched a few S1/R videos and mainly the cameras struggle with low light levels. Best

I use AF only when facial recognition could be used. It works well in low light. Otherwise, it is old school follow focus on MF. There is no other way to get precise focusing in video, no matter how good the AF is.