[BrianR]

I stutter a bit, which you can hear (kind of feeling exposed, as this is the first time I uploaded a video to Youtube with my voice in it). I hope it doesn't distract too much from the content.

No distraction on my end! Thanks for taking the time

I'm also surprised that the Pentax would give up and not be able to find a lock at times. I've actually been pretty amazed at how my k5iis can reliably lock onto something in low light levels (mostly non-moving targets). I've no experience with the k-3 (which should be improved), nor the 16-50mm.

[starbase218]

@Normhead

Here you go attacking again. I don't think my experiences are more important. But if you think mocking me is going to get me to say that, you are wrong. And given the tone of your initial reply, I could say exactly the same thing about you.

[BigMackCam]

OK folks - let's all keep this friendly and respectful, please. I'd rather not have to close the thread.

Thank you!

[starbase218]

@Normhead

I think we're going downward in a spiral where we keep making negative comments about each other, not unlike what happened before when I joined in discussions on this forum. I'm going to end that spiral now, because I don't think this is going anywhere positive. I do acknowledge your points and opinions, I just don't agree with them. Goodbye and good luck.

[normhead]

@Normhead

I think we're going downward in a spiral where we keep making negative comments about each other, not unlike what happened before when I joined in discussions on this forum. I'm going to end that spiral now, because I don't think this is going anywhere positive. I do acknowledge your points and opinions, I just don't agree with them. Goodbye and good luck.

Well, good luck to you too then.

It should be pointed out, the last time we went through this it came down to Nikon and Pentax having a different philosophy, with Pentax having a little hesitation where focus was checked at the end of the cycle, and Nikon didn't. The result was, with very similar lens focussing speed, Nikon attained a faster forcus lock, but it wasn't as accurate. The question we ended up with was, could Pentax achieve Canikon focus speed with the same accuracy, by sacrificing that last stutter step and taking a hit in the accuracy department. This was a few years ago, but I believe the D7100 was actually the camera tested for Nikon. The other question we asked at the time, was if Pentax did give up the stutter step to increase speed over accuracy, would we want that? Another question was, have Canon and Nikon programmed their AF systems to just fire off an out of focus frame when they can't lock focus to reboot the process on the next frame.

Needless to say the topic is just way more complicated than it looks on the surface.

If conditions are the same, and everything else is the same remains to be answered. But, what I've been getting at would be, there may have been negative trade offs to achieve the AF speed in both Canon and Nikon, parsonic, Fuji and others. Unless you can show Nikon have improved the number of images they achieve in acceptable focus in a significant number of trials, we are just rehashing what we already know. We know Nikon locks focus faster, and always did. The question is, does that make the camera more or less reliable.

That research suggested that if you were a one shot, take your time and focus type shooter you might be better off with Pentax. Because with no stress on the AF system, you'd have a 96% keeper rate, compared with a high of about 93% for the average system out there and under 90* for many systems.

Which is why I'm wondering why you didn't actually take any pictures. You ran the part of the tests where Pentax didn't do as well, but you didn't run the other half, where Pentax came out on top.

Hopefully that explanation will be negativity free enough for everyone.

I guess I tend to forget not everyone has not been around for all of this. When I think of it now, I might be the only person still here that participated in that discussion. The title was something about a German magazine that ran the tests.

[monochrome]

^^ I recall the discussion just as @normhead described it. It was speculated that both Nikon and Pentax estimate where to move the lens (equally quickly), then their process logic deviates. It was speculated that Nikon's AF algo quickly calculates whether the focus is above some slightly-less-than-perfect threshold to permit release, whereas Pentax algo logic always attempts to achieve perfect focus (so there is an additional 'bump'(s) to the lens before it permits release).

Nikon = ZIP -> shutter
Pentax = ZIP, Bump, Bump -> shutter.

[starbase218]

Alright, since you continue on about this, I'll also post my thoughts, since I made the video. The objective of this test was not to test AF accuracy, but AF speed. I understand that you might find the test of limited value, but if I wanted to test AF accuracy, I'd have taken multiple images with both cameras at low ISOs, instead of conincidentally looking at a noisy, smudged liveview image of only one of the cameras with the brightness levels causing blooming on my iPhone sensor (btw, images look softer on the D7100 screen, even though sharpness is a wash between it and the K-3). Also, I haven't had the opportunity yet to properly set AF finetune for this lens. With regard to the blooming: the metal of the speaker stand below the logo does look pretty sharp. With DOF being about 14cm, and the metal being at most 2cm further away from the sensor.

In my use so far, I have found it more difficult to achieve precise focus with the K-3 because, from my observations, it seems that the AF points on the Pentax are far larger.

Still, I do acknowledge there might be something in this. But more testing would be required before I would be comfortable to say anything more about AF accuracy. I'm not planning to do it though. I think I've spent enough time on this already. I have other, more important things to do.

---

Last edited by starbase218; 03-23-2017 at 11:28 AM.

[biz-engineer]

Then it may be that the SDM drive in the 16-50 sometimes overshoots where it is supposed to be focusing, and keeps overshooting from that point on until the camera gives up.

There is the effect of the SDM drive , for sure. And also the Pentax algorithm that never changes...Pentax focuses in two stages: 1) direct move up to a point (like Nikon and Canon) and 2) fine-tuning and lock (what Canikon aren't doing). Stage 2 is the problem every time, the focus is already well reached and the Pentax camera tries to tweak the last bit. They should give the choice to users (e.g C menu setting) to disable this tweak. When the conditions are good, the tweak may give a tiny bit more accuracy, but when in poor lighting (weddings etc), not doing the last tweak would be more beneficial than doing it: without the tweak you get a rapid focus of 99% in focus shot before the subject - photographer distance changes, and with the last tweak you get worse out of focus image because the subject-photographer distance may have changed by the time the camera is able to lock focus. The issue with Pentax is that they can't care less about what's I'm writing here.

It was speculated that Nikon's AF algo quickly calculates whether the focus is above some slightly-less-than-perfect threshold to permit release, whereas Pentax algo logic always attempts to achieve perfect focus (so there is an additional 'bump'(s) to the lens before it permits release).

That's it.

It is possible to emulate the Canikon way by setting AF to FPS prio, or use the back button AF, and make the decision of when to shoot for the camera.

[normhead]

In my use so far, I have found it more difficult to achieve precise focus with the K-3 because, from my observations, it seems that the AF points on the Pentax are far larger.

Yet in indépendant tests, their AF was found to be more accurate, so for some of us, that's an acceptable trade off. I'd agree there may be a problem with some images, but a lot of the the time there isn't. In the image below I locked focus on the goose head behind the grass. despite there being grass very close to the goose right in front of it that could have thrown off my focus. One of the Nikon shooters had trouble with focus lock with nothing between him and the otter. I absolutely refuse to draw conclusions because I can't sit down and analyze exactly what happened. But from my experience anyone can have these kinds of issues. They aren't just happening to Pentax people. Same with AF. No camera achieves 100% correct focus, but most of them do the focus confirmation thing and take an out of focus image. My Pentax doesn't set the shutter off. It tells me it can't lock focus, as opposed to pretending it can, and snapping off a worthless exposure.



I'm not cherry picking these images, I'm drawing from what happened to me and around me in the last week. That image is a few days old and therefore fresh in my mind as an example.

The other reason the AF speed is pretty much irrelevant to me, is I shoot first frame focus, not coninuuso focus, because I want that fast burst. The camera doesn't refocus until I lift my finger off the shutter release button, and I don't do that until I see the animal move. Personally I find asking the camera to check focus between shots when the animal hasn't moved is a huge waste of burst speed. For example in your polar bear image. Their feet are planted, they aren't going to move out of your DoF. GO for burst speed. At that point, that's all that's important. You've got your focus point established.

When you consider how many images it is spread over, my AF is really fast. I may be .2 seconds behind acquiring focus on a low light shot, but I do it once. Divide that time by 23 shots and I'm way ahead of the game. And a camera with slower FPS rates and smaller buffer have no chance to catch up even if they do what I do. So technically the way I shoot, my camera blows your D7100 out of the water, just calculating how much time the camera spends on AF for each image.

Had I saved everything I shoot I could easily pick out a few 20 shot burst where the keeper iis in between the 16th and 20th shot in the burst. IN other words, there are very few cameras other than a K-3, and certainly even fewer at the same price, that would have given me the keeper. Fast AF is not the only way to achieve "capturing the perfect moment". In fact from my pers[ecive, there are more times when it's a hinderance because people leave it on, than it is a detriment. My small bird images are filled with images taken every where in the burst, and it's painful when using the K-1 to miss those images.

Much of the time I'm getting better images with the AF turned off, than the "faster AF " guys are with it turned on, simply by understanding where my DoF is and when I have to refocus, instead of just focusing everything and relying on AF to get the shot they want. As far as i can tell, Pentax gives me a more accurate AF lock, and that is really useful, because I'm going to use that AF lock up to 23 times, and If you have a fast shutter like the K-3, having the AF on will usually slow down your FPS.

Maybe if I had a !Dx I'd do it differently, but my goal is to do better than those around me by taking advantage of what my camera does and how it does it. I guess I could just throw bucks at it like so many others, but this is ultimately much more satisfying, at least for me. Maybe a pro would see it differently.

The other day with the K-1 and 60-250 on the ground in front of me, one of the otters popped out of the water right in front of me, A guy with a 5Dmk4 beside me asked me what I was doing. I said "this is my full frame". He said," I shoot full frame all the time". I said "my APs-c gives me about the same of MP as your full frame, and this camera gives me 36MP. If one of those shots turn out and one of your shots turns out, mine is going to have noticeably higher resolution and dynamic range." If you work with your gear, you figure out how to use it to your advantage. I won't come out on top every time, but there are enough times I do that you can't tell me, spending a lot of money would get me significantly more.

Sure there are lots of way I could use my gear where they wouldn't be as good as the competition. My attitude to using my gear not to it's strengths is "don't do that, buy the appropriate tool for the job."

[Nicolas06]

Of course Pentax could, but we wouldn't buy it if they did.

There's no magic to cameras. Pentax could immediately create a D500-spec. body and Nikon class strobe. It's just math, engineering and capital. What they can't do is recapture the time lost since 2000 doing nothing while CaNiFuShy were incrementally developing their products.

Ricoh could throw capital at the problem and catch up 15 years of investment in 2 - 3 years. But we wouldn't pay the price required to recover the development cost, tooling and materials.

The problem is us.

If they put the money, the right team and give them the time, yes they can. So let say they make it their #1 priority, they spend tons of money on it and in 5 years, they may have the D500 performance and maybe even match the new version of that D500.

This is not even sure. They may do that, and they don't really invest well, there management issues, their employee do a good job on the topic, but not a great job... And so they would get better AF, but even with all the time and money they would not catch up.

What they could do maybe is to ask Nikon to provide them their AF sensor and all and pay a lot for it. It may be then instant and ready for the next model, but you need good negociation skills and the cost Nikon might ask may be exchange of technology (like SR + pixel shift) AND sell their AF sensor for a high sum meaning it could be incomporated on a 1800€ APSC body or 5000€ FF body, but not in anything cheaper.

No things are certains in life. AF is Nikon asset. They are the best at it. Canon has bigger budget and does good, but is behind on that topic. This is like being the worldwide champion in a sport. You can't just wake a day say you'll train for 2 years and that you'd win the Olympics. You might achieve it but that very unlikely. What you can do is improving but if you are too old, you know your results may be quite limited.

The same is likely for Pentax. Would the most experienced people in AF would accept to work for them and leave Nikon and Canon? Would Canon or Nikon simply agree to sell them their existing system, even the N-1 generation that would still be a huge progress for Pentax? Not that likely. Would just asking the employees that didn't manage that well on that area on the past 15 year a warranty of success if you just give them more time? Or hiring many newcomers, new in the field without much experience? Again far from being sure.

More generally we all know that we are not equals, not as creative. Experience also count a lot but there also politics, corporate policy and many things that could go wrong. If people were always doing their best, not only for themselve but for the whole community, there would be no war, no starvation, no offence... We are far from that perfect world. And people would find it stupid to redo in Pentax what is already in Nikon. They have 1 camera system compatible with everything that would do a lot for a cheap price... Because the goal would not be to have your own private echosystem to begin with and people would have other priority than to be focussed on a brand.

[Nicolas06]

It is possible to emulate the Canikon way by setting AF to FPS prio, or use the back button AF, and make the decision of when to shoot for the camera.

Back button AF is very powerfull tool. Pentax AF maybe not the best, but it is quite capable in the hand of people that know what they are doing.

[starbase218]

There is the effect of the SDM drive , for sure. And also the Pentax algorithm that never changes...Pentax focuses in two stages: 1) direct move up to a point (like Nikon and Canon) and 2) fine-tuning and lock (what Canikon aren't doing).

Actually this is not entirely the case. You may even be able to tell it from my video. The Nikon does at least some fine-tuning. It just takes less time than with the Pentax. I actually also have a Nikon 70-200/4 which, when focused closely at 200mm in dim light, sometimes keeps jumping around for multiple seconds before settling. The difference is, it jumps around much faster than I ever experienced with my Pentax gear. Still, I actually wonder if it's normal for that lens to be tick-tick-ticking away for multiple seconds, so I asked around on a Nikon forum. They advised me to contact Nikon about it, which I still have to do.

[biz-engineer]

Actually this is not entirely the case.

Ah ok.

Anyway, AF on static subjects with Pentax or any other camera, frankly speaking , checking for differences is splitting hairs. Then you have the low light cases etc... you can always put ANY the camera are the border of capability, you'll also put the best one to fail. Someone tested 1Dx or D5 at -2ev and was disappointed. I've also been taking photos in ugly lighting , and whatever the autofocus performance was, the results were never so great. Doing test shots at home in the dark after work with artificial lighting , berk, just more photos to delete. When the lighting is good to produce nice shoots, everything works out, I tell you.

[monochrome]

At some point it comes down to the person behind the viewfinder.

[starbase218]

Ah ok.

Anyway, AF on static subjects with Pentax or any other camera, frankly speaking , checking for differences is splitting hairs.

I have absolutely no problem with Pentax AF for a lot of types of photography: landscapes, stil lifes, posed portraits, and I managed to get good shots out of my K-5 and DA 55-300 when tracking birds and jets as well (though I am disappointed that the much more expensive DA* 60-250 seems less capable in this regard).

But sometimes I wish the AF would just respond a little faster. In my PPG there's a photo called "Gift from Laos". I took it with my K-7 and DA 17-70/4 at the time. These kids were moving all the time. Sure, they were aware of me, but that's something else to them being aware that I needed some time to compose, focus etc. In the end, I just shot away with them moving in front of me, hoping that it would yield a nice result. And it did, but sheer luck had a lot to do with it. To me, that's not a very satisfying way of taking photos.

Then you have the low light cases etc... you can always put ANY the camera are the border of capability, you'll also put the best one to fail. Someone tested 1Dx or D5 at -2ev and was disappointed. I've also been taking photos in ugly lighting , and whatever the autofocus performance was, the results were never so great. Doing test shots at home in the dark after work with artificial lighting , berk, just more photos to delete. When the lighting is good to produce nice shoots, everything works out, I tell you.

Hmm, not sure I agree with you on that. I mean, light is everything (or at least it's very important). But low light can still be good light. Unless I'm the only one who appreciates the high-ISO capabilities modern bodies are capable of?

-2EV though... that's really low I think? I'm a bit confused as I looked up a table relating exposure settings to EV values, but it didn't include ISO.