[Parallax]

I discussed my appendectomy with my surgeon, that doesn't mean that I know how to perform one.
DPR is so biased against Pentax that there will never be one that they like. (Or admit to liking, anyway.)

[FantasticMrFox]

I discussed my appendectomy with my surgeon, that doesn't mean that I know how to perform one.

Some analogies make sense, others don't.

DPR is so biased against Pentax that there will never be one that they like. (Or admit to liking, anyway.)

Is that why they gave the K3 an 83% Gold Award?

[normhead]

I'd like to point out that a) the people at DPReview are no amateurs, they definitely have some experience shooting sports and b) they stated the following: "We tested continuous focus with a variety of Pentax lenses, having discussed relevant settings with Ricoh." Just claiming that they don't know what they're doing doesn't quite cut it here.

You do realize that you are claiming they do know what they are doing?
Well then, just buy the camera they tell you to buy and leave us alone. Obviously they are smarter than the collective wisdom of the forum.

We aren't in any way restricted by what you say we can't do, and we aren't going to stop doing it, because DPR claims some other camera is better. But you might want to consider, it's very rare that a camera is the best of anything in even one aspect of it's class. Claiming some other camera is better than the one you use doesn't mean a thing unless you buy it. So why not buy what DPR says is good and be done with it, if you trust them so much. Put your money where your mouth is.

It's no sweat off my back, but it does go to your credibility. If you think having the absolute best AF is important to your images and that some other camera is better, you're a hypocrite just owning a Pentax. (I don't believe that, I believe you can take great pictures with less than the absolute best AF and so does 98% of the rest of the camera buying public.)

[FantasticMrFox]

You do realize that you are claiming they do know what they are doing?

I am. Tracking a subject coming directly towards you is no rocket science. Just because half of this forum seems to dislike DPR for not praising Pentax like they'd like them to do doesn't invalidate them as a source. I used DPR as a source because they've written extensively about the K3's AF.C performance and provided their methodology and samples. I could just as well cite tons of other reviews that come to the same conclusion but aren't as well founded.

Well then, just buy the camera they tell you to buy and leave us alone. [...] So why not buy what DPR says is good and be done with it, if you trust them so much. Put your money where your mouth is.

I bought the camera that suits my needs best, and I get along with it splendidly, thank you very much. Pointing out a weakness doesn't mean I want to buy an entirely different product. Stop playing offended. Telling someone to just bugger off and buy another product just because he isn't satisfied with a particular feature is pretty fanboyish.

I trust them on this because it mirrors my experience, and that of many others. With a subject coming towards me in a predictable fashion the K3 would reacquire focus not often enough. It's exactly what DPR found. I and they used the settings Ricoh recommends.

[normhead]

Are you now going to point out the weaknesses in every other camera out there, or is it just Pentax that deserves this treatment? You did bring it up, so you just have thought is was important. Or were you under the impression that we all thought Pentax had the best AF out there?

I bought the camera that suits my needs best, and I get along with it splendidly, thank you very much.

But you love to point out other cameras that may have better AF, as if that were somehow relevant? What wid dat?

I'm not going to analyze DPR to find out exactly what I don't like about them, but I find them amateurish.

But if we're talking credentials, I have 50 years experience shooting Pentax, 15 years experience teaching photography, a year of school in photography, at least 10 working pros I've discussed photography with for years, and here's my take on this. I don't get caught up on the little stuff.

IN my sequence above, I used my F-70-210, because it's my fastest focusing lens, and despite what's been posted here, my reference for sharpness is the tennis ball in the dog's mouth and it stayed in focus. Which means the camera was tracking it. And I didn't even get a good focus lock on the first frame. The camera picked it up, which to me seems pretty incredible. But see here's the difference. I know which is my fastest focusing lens. It's either the Tamron 90 macro or the F 70-210.

I don't have to test a pile of lenses to know that, because I use the equipment all the time, and when I want fast focus for medium wildlife I shoot with the F 70-210. Sometimes fast focus trumps resolution. Did DPR shoot with an F 70-210? Then in my experience, they didn't test the AF system, they tested the relatively slower speed of the lenses they used. So technically, they didn't test the AF at all, they tested the speed at which lens could focus.

When a subject is coming straight at you, you are actually testing how fast the AF drive unit moves the lens elements. Not the AF. The problem is, if the company comes out with new lenses that move the elements faster, then the AF on the system will be better. So what exactly did DPR test? We don't know, they don't know.

Their test is completely irrelevant to me. And it's only relevant to anyone who owns both cameras and the lenses used on those cameras and can choose to use one or the other. SO why would DPR put 99% irrelevant data into a product review? Quite simple, they have an agenda. Why would they even mention one other product? That's tacky. If they wanted to mention all the products that have better AF than Pentax, fine but one? Sorry, completely unprofessional.

You might want to take a second to comprehend this.... a test is only good, if it tests what it claims to test. Many times people will devise a test for something that doesn't test what they think it tests, and then proclaim the test from the roof top. But then, back at Ryerson Politech, we spent a lot of time discussing system testing, blowing up the flaws in testing procedures, understanding why when hype made a system sound fantastic, the results rarely met the hype etc. My problem with DPR is they so often don't meet the minimum standards I'm used to, and can't see out of the box they think in.

I see way to many flaws of logic in their text, commentary and conclusions, it's painful to read them. Maybe the less educated will be hoodwinked. They do provide good descriptions of the cameras they review, just don't get caught up in their technical analysis. There, they are amateurs.

Results for these types of tests, are relevant only to the lens and camera tested. Generalizations based on that aren't worth publishing.

[Rondec]

I am. Tracking a subject coming directly towards you is no rocket science. Just because half of this forum seems to dislike DPR for not praising Pentax like they'd like them to do doesn't invalidate them as a source. I used DPR as a source because they've written extensively about the K3's AF.C performance and provided their methodology and samples. I could just as well cite tons of other reviews that come to the same conclusion but aren't as well founded.



I bought the camera that suits my needs best, and I get along with it splendidly, thank you very much. Pointing out a weakness doesn't mean I want to buy an entirely different product. Stop playing offended. Telling someone to just bugger off and buy another product just because he isn't satisfied with a particular feature is pretty fanboyish.

I trust them on this because it mirrors my experience, and that of many others. With a subject coming towards me in a predictable fashion the K3 would reacquire focus not often enough. It's exactly what DPR found. I and they used the settings Ricoh recommends.

I think truly testing auto focus is more demanding than you seem to think. Comparing one camera with another requires some standardization, both with regard to light used, glass, and rating. The folks who do this usually take a series of photos and rate them with regard to focus. Number of sharp shots is one way to compare, as is the distance to the last in focus shot the camera got.

I would think the best way to compare would be to use a lens available in mutliple different mounts, like a Sigma 70-200 and use that as your baseline.

Anyway, I don't really see DP Review having some sort of systematic evaluation of different cameras when it comes to auto focus -- unlike Chasseur d'Images, for example.

It doesn't mean that they are wrong, but it does mean you have to take their blanket statements with a grain of salt.

[Parallax]

Comparing one camera with another requires some standardization, both with regard to light used, glass, and rating.

.....and same subject.

[FantasticMrFox]

Are you now going to point out the weaknesses in every other camera out there, or is it just Pentax that deserves this treatment?

You mean the sub-par build quality and questionable IQ of the 70D? The D7100 weather sealing which is below the level of Pentax? The fact that the Canon 6D is absolutely castrated in most regards and they can only charge over a grand for it because it has a FF sensor?

This is a forum about Pentax cameras, why would I write about other cameras? I'm shooting Pentax and I'm active on this forum, what makes you think I'm only criticising Pentax? When I want to point out a flaw in something, do I automatically need to point out totally unrelated flaws in other products in order to proof my unbiasedness?

There is a guy in the Pentax section of the German DSLR Forum who seems to hate Pentax. He claims to shoot Pentax, but criticises every single aspect of the system. It's overpriced, the lenses are subpar, there's no FF, the new FF will suck ... Some other members and I keep proving him wrong, but he just labels us as fanboys.

Pan to Pentaxforums: I'm criticising one aspect of the otherwise amazing K3. I'm accused of only criticising Pentax and asked to just get another camera because apparently I must hate it.

Seriously? Why are so many people so unbelievably rigid in their brand perception?

But you love to point out other cameras that may have better AF, as if that were somehow relevant? What wid dat?

The fact that this thread's main content is comparing the D750 and K3 and the AF system is a major difference?

So technically, they didn't test the AF at all, they tested the speed at which lens could focus.

They tested a whole range of screw drive, SDM and DC lenses, including the DA 18-135 which is considered to be among the fastest Pentax lenses. They cannot test every single Pentax lens there is, by using a cross section of available lenses (including pricey high quality options like the DA* 50-135, DA* 55 and DA 70 Ltd.) they tested the system as a whole with lenses that shooters are expected to use. You can't expect everyone to acquire an out-of-production legacy lens.

When a subject is coming straight at you, you are actually testing how fast the AF drive unit moves the lens elements. Not the AF. The problem is, if the company comes out with new lenses that move the elements faster, then the AF on the system will be better. So what exactly did DPR test? We don't know, they don't know.

The issue they found was that the K3 was hesitant in initiating refocusing, which is unrelated to the in-lens AF motor. The DC motor in the 18-135 is certainly fast enough to refocus, but it won't if the camera isn't telling it to. This is a problem on the side of the camera, not the lens.

And it's only relevant to anyone who owns both cameras and the lenses used on those cameras and can choose to use one or the other. SO why would DPR put 99% irrelevant data into a product review? Quite simple, they have an agenda. Why would they even mention one other product? That's tacky. If they wanted to mention all the products that have better AF than Pentax, fine but one? Sorry, completely unprofessional.

99% of the review is a test of the K3's capabilities. If you actually read the whole review, you'd notice all the praise they have for the camera and Pentax as a manufacturer. But you notice two sentences in the AF section that mention another camera and you immediately lose it about some 'agenda' and how suddenly 99% of the information in the review is irrelevant. Seriously?

You might want to take a second to comprehend this ... a test is only good, if it tests what it claims to test.

You can save your lecture, I study a degree that's all about test validity, reliability and making sure you actually measure what you want/claim to measure.

My problem with DPR is they so often don't meet the minimum standards I'm used to, and can't see out of the box they think in.

They wanted to test a camera's AF system. They used a wide variety of lenses, some of which are among the system's fastest. They explicitly wrote that they thought it's the lenses' AF speed which in many instances is holding the camera back, at the same time they also pointed out flaws in the AF system itself. They did exactly what they wanted to do and stated exactly what they measured and which extraneous variables (lens AF) there were and how they might've interfered with the results.

I see way to many flaws of logic in their text, commentary and conclusions, it's painful to read them. Maybe the less educated will be hoodwinked.

I see someone who doesn't like the results of the test and thus tries to deliberately pick it apart and undermine it's credibility with questionable arguments, trying to construct some 'agenda'.

[Winder]

I would think the best way to compare would be to use a lens available in mutliple different mounts, like a Sigma 70-200 and use that as your baseline.

You would expect that Canon AF would be faster with Canon glass, Nikon AF would be faster with Nikon glass, Sony AF would be faster with Sony glass. Sony's AF-D mode only works with about 5 Sony lenses and zero 3rd party lenses and AF-D is their best mode for tracking a moving subject.

These are systems and should be tested as such. My favorite Ricoh lens on my K-3 is the 31m and if I were to buy the Fuji I would test it against the Fuji 35 or Zeiss 32mm. I think there are big advantages to testing the camera with the lenses that people are most likely to actually be using. Many people buy into a system specifically for the glass. I would test the DA* 55mm vs the Fuji 56mm on their respected bodies, because that is how they would actually be used.

Ever since I started using K-mount I have been saying they need to update the AF motors. Screw drive is slow, loud, and I don't think it has the precession necessary for consistent AF with fast FF glass and todays really high resolution sensors. SDM is simply slow.

[FantasticMrFox]

These are systems and should be tested as such. [...] I think there are big advantages to testing the camera with the lenses that people are most likely to actually be using. Many people buy into a system specifically for the glass. I would test the DA* 55mm vs the Fuji 56mm on their respected bodies, because that is how they would actually be used.

Exactly. Perfectly objective testing is an illusion, even in the most exact of natural sciences. Equipment tests need to have practical value and the best in-camera AF system is worthless if the surrounding system cannot properly support it.

[rawr]

unlike Chasseur d'Images

Or Pop Photo, who do a regular chart of AF speed (seconds to lock) by light-level (bright to dim) test.

But overall, it is surprising that there is very little rigour in AF testing by review sites. I can see why DxOMark sticks to simply testing camera imaging sensors and lenses - AF testing is just too hard.

On the subject of Pop Photo, in their K-3 review they did note:
'Burst shooters should be happy with the drive speed of 8.3 frames per second. That’s fast enough to handle almost any sport and, in our trials at the marathon, the continuous AF did well at tracking as runners approached the camera and then rounded a corner.' The K-3's Nikon peer at the time of the review was the D7100, so they did mention: 'but the Nikon has slightly better AF tracking—we did see a few less-than-sharp images in our marathon images.'

That's perhaps the difference between 51 and 27 AF points at work. But it was only 'a few' unsharp shots, not an 80% OOF rate or anything.

Similarly Imaging Resource, when they went to the GoKart races with a K-5 and K-3:

I sat down and figured out how many of my photos -- I'd shot a few hundred, between both cameras -- were in focus, nearly in focus, or out of focus. Here the K-3 had a clear advantage, especially when shooting with the 18-135mm lens. With my K-5 body, perhaps one in four burst shots with that lens was significantly out of focus, and around a third of the remainder were close, but didn't quite nail the focus. With the Pentax K-3, though, about 90% of my shots with the 18-135mm lens were in focus or very close to it, and for the great majority of those, focus was spot-on.

Anyway, this is a D750 thread ...

[Rondec]

You would expect that Canon AF would be faster with Canon glass, Nikon AF would be faster with Nikon glass, Sony AF would be faster with Sony glass. Sony's AF-D mode only works with about 5 Sony lenses and zero 3rd party lenses and AF-D is their best mode for tracking a moving subject.

These are systems and should be tested as such. My favorite Ricoh lens on my K-3 is the 31m and if I were to buy the Fuji I would test it against the Fuji 35 or Zeiss 32mm. I think there are big advantages to testing the camera with the lenses that people are most likely to actually be using. Many people buy into a system specifically for the glass. I would test the DA* 55mm vs the Fuji 56mm on their respected bodies, because that is how they would actually be used.

Ever since I started using K-mount I have been saying they need to update the AF motors. Screw drive is slow, loud, and I don't think it has the precession necessary for consistent AF with fast FF glass and todays really high resolution sensors. SDM is simply slow.

Maybe. Should you just choose the fastest lens on each model or, a 70-200 f2.8 or, a 50mm f1.4 just for the sake of parity? I don't know. Primes in my experience often have longer focus throws and so take longer to get where they are going, but maybe we should throw the Canon 50mm f1.2 in to the mix, as it does not seem to be super speedy...

It takes time to learn a camera's auto focus system well enough to get the most out of it. Hard to believe that someone could pick one up in a few hours make a real judgment about accuracy or speed of the auto focus system.

I think Pentax still does lag behind Canon and Nikon, but I have seen significant improvements with the K3 and hopefully the K3 II improves those even more. Hopefully as well, there are more speedy lenses like the 70-200 and 150-450 in the works.

[normhead]

Or Pop Photo, who do a regular chart of AF speed (seconds to lock) by light-level (bright to dim) test.

But overall, it is surprising that there is very little rigour in AF testing by review sites. I can see why DxOMark sticks to simply testing camera imaging sensors - AF testing is just too hard.

On the subject of Pop Photo, in their K-3 review they did note:
'Burst shooters should be happy with the drive speed of 8.3 frames per second. That’s fast enough to handle almost any sport and, in our trials at the marathon, the continuous AF did well at tracking as runners approached the camera and then rounded a corner.' The K-3's Nikon peer at the time of the review was the D7100, so they did mention: 'but the Nikon has slightly better AF tracking—we did see a few less-than-sharp images in our marathon images.'

That's perhaps the difference between 51 and 27 AF points at work. But it was only 'a few' unsharp shots, not an 80% OOF rate or anything.

Similarly Imaging Resource, when they went to the GoKart races with a K-5 and K-3:



Anyway, this is a D750 thread ...

Ya, it's a D750 thread, so make your D750 points without spreading mis-information or partial information, or information of limited use about my K-3. I came here to hear what people had to say about the D750, because of its much faster focus lock.... not to hear people tell me what can and can't be done with my K-3, which I can decide for myself.

So far I'm seeing trade off. Faster initial focus lock, slower frame rate, in my mind, pretty much a draw, so far. And here is the problem with that, when tracking an animal the K-3 is not slow, because often the animal has only moved forward or back a few feet from the last time I had him in focus. So real world these numbers are ridiculous. The number of times I have my lens at infinity and I go all the way from there to 12 feet away , it almost never happens. What is more typical, is the animal moves backwards or forward a few feet. And the focus lock is almost instantaneous, even if I'm focusing manually. So then the difference is probably like .05 for the D750 and .07 on the Pentax, and the focus lock number becomes meaningless, but the K-3 still shoots 8 frames per second. The whole argument that fast AF is critical, is based on the erroneous assumption that there are a lot of situations that require it. Honestly, these picture of runners in a race running towards you or cars coming at you on a street, I have not one image like that. What is critical is that focus lock is accurate when achieved, and dated but most recent information, I've seen would suggest a Pentax focus lock is more accurate than any other brand, although it might take a little longer.

I wish I could lend my dog to one of the D750 owners. I'd love to see what they come up with in a similar sequence. Although, the frame rate has seriously killed much of my interest in the camera. I hate paying more for less, of anything. For every argument that frame rate is not that important, I'll give you two reasons why AF speed is even less important.

[rawr]

AF-C and tracking is interesting. There's probably a lot of variation out there about how people use it, hence differences in results.

And in terms of fps, I always shoot in small bursts, for example, so 8 fps or 5 fps is usefully fast, but it's not as if I aim to use up all those frames every time I press the shutter.

[normhead]

The only time I use up the buffer is when anticipating a move... you have a stationary subject that you think will start walking away, and you want the moment when it turns. Anticipating movement, you press the release and hold it.... but often you fill the buffer, then the critter moves and you have 23 wasted frames, which you delete with gusto, sometimes damaging the keyboard.

Not that i've ever done that.

Oh by the way, about the soft focus thing... a little pixel peeping...

4th frame and 10th frame....





For those of you who saw some really soft out of focus images, looks bad on you. .

Not only that, the tenth frame is one of the sharpest.... so much for lack of tracking ability. I'd like to see hand held panning images, from another camera that are as sharp. Someone has to step up and make a case for the D750.