[JJ_SO]

Exactly.   I'm exaggerating - focus is usually pretty good in non-critical situations. Moving people when the camera needs ISO 12800 and wide open at 1/40 is such a situation, together with artificial colored and mixed light sources pretty much a nightmare. What I said is as well for Sigmas as for Nikkors, I don't see much of a difference in these lenses focus reliability.

 

Subjectively I'd say the Sigmas get higher keeper rates, but this is no solid data. If we'd live closer, I'd be happy to borrow you a couple of those f/1.4 lenses. I had only briefly direct comparison with a 24/1.4 from Nikon and from Sigma but with those wide angles is hard to tell if I missed the proper pointing at a good target or the lens missed focusing to it or the camera was confused. All I can say is both cameras brought out an equal number of good focus, with maybe a slight advantage to the Sigma as I can adjust different distances but that is only visible in close-ups and can be solved better with live view.

 

Side note: in regards of the 85/1.4 and the truly shallow DoF I think, I'd be better of if I had known there's a 85/1.8 around the corner being better in some aspects as the 85 is - which already is a fine glass. And was in the repair because of a AF motor failure. No repairs so far on the Sigma side, but all of those are younger than the 85.

[thxbb12]

Let me clarify a few points.

I never said mirrorless is good at AF tracking. I only stated, like many people who use both DSLR and mirrorless, that mirrorless is more accurate at focusing a given target. Nothing more. It's not about the tracking ability, but the ability to reliably and consistently focus correctly.

 

I never claimed that a DSLR cannot focus properly. I only state that a DSLR is much less consistent than mirrorless when it comes to reliable focusing accuracy. Brightcolors, I can also show you tons of well focused images I shot with my Nikon or Pentax DSLR gear. That's not the point. What your pictures don't show is the percentage of accurately focused shots you get of the same subject, repeatedly.

 

As Joju said (as well as Nasim Mansurov in the link above), there are many factors at play when a DSLR focuses. Unless everthing is perfectly aligned, one is at the mercy of some focus inaccuracy.

Brightcolors is lucky to have a DSLR camera that focuses perfectly every single time with every lens in every condition. That's quite remarkable and borderline miraculous.

 

Now, I challenge anyone with a DSLR, preferably high-end, to perform the following test:

• Setup 3 still targets at 3 different distances.
  
• Shoot each target using your DSLR's PDAF.
  
• Before each shot, reset the lens twice: at close range and at infinity.
  
• Shoot every target several times.
  
• Use different lenses and shoot them all wide-open:
  <ul><li>Fast ones (f1.8 or below)
  
• Zoom lenses (preferably f2.8)
  <ul><li>Set your zoom lenses at 3 different focals: wide-end, long-end and in between.
  

</li>
</ul></li>
</ul>Once you've done this. Look at all your shots and get back here with your findings.

You'll notice that many shots will be off. I'd actually be interested to know what % of shots are actually dead on. I'm sure you'll be surprised by the number.

 

Now, if you do exactly the same with a Fuji system (for instance), pretty much all the shots will be dead on regardless of the focal or the target distance.

 

I'm saying this because that's exactly what I did a while back. With a friend of mine who is a wildlife photographer, we spent several days calibrating our lenses using FoCal. At the time, I shot with a D800 with the following: Nikkor 50 f1.8, Nikkor 85 f1.8, Nikkor 70-200 f2.8, Nikkor 20 f2.8, Nikkor 35 f2, Tamron 70-300, Tamron 90 SP f2.8, Nikkor 24-120 f4, Sigma 50-500. My friend had a D4 with several lenses, but I remember mostly the 300 f2.8, 400 f2.8, and 200-400 f4, Nikkor 85 f1.4 and Sigma 35 f1.4.

Let me tell you this: focus reliability, even after calibration, was much worse than what you get today with mirrorless. Mind you, this was with a D4. Is this body high end enough?

The worse part was that with zoom lenses, you needed different micro-ajustment at different focal lengths. After these findings I was so fed up that I decided to give mirrorless a try. What I found in terms of AF accuracy was a revelation. Some time later I decided to fully move to mirrorless.

 

I'm not saying mirrorless is the silver of bullet or that it's for everyone. Today, wildlife and action shooters are still better served with DSLRs because of their AF tracking ability and the lens selection.

However, in other use-cases, mirrorless is better IMO. Portraits shot at f1.2-1.8 is definitely one of them, unless you don't mind taking 10 shots and hope that half of them will be dead on.

 

Does this (long) explanation make sense?

[Brightcolours]

My images were not the best focussed ones out of a whole series of the same framing and subject, they were just one shot and done. The 55mm f2 with adapter will consistently nail focus, that is just the way it is.

 

I have never said that my 450D focussed well with every lens, stating so is again just mocking things. My 350D was very inconsistent with my Sigma 18-50mm f2.8. The Sigma does great on my friends (gave it to him) 450D, but the Tokina 12-24mm f4 was not good in focussing on the 450D (it did well on the 350D).

And above I stated already that my crappy 28-135mm does not focus accurately on my 6D. 

 

Stating that a MILC will focus "perfectly every single time with every lens in everycondition." is just nonsense too.

 

I have no confidence at all in your anecdotes, as you keep on even misrepresenting the posts I wrote that you reply to, so can't rely on what you say. I can however rely on Foto Magazin doing a very good test with "comparable" lenses in identical conditions with a 70D, D7100, K3, 77 II, E-M1, GH4, 6000. The Canon, with ancient EF 50mm f1.4 USM got 95.4% AF accuracy (images in focus) with PD AF and 95% with CD AF. The Oly 91.2%, the Pana 85.5%, the Sony 85%.

 

So, a methodical and objective test got those results. But my images from the 450D and that fixed focus lens are not good enough for you, nor the Foto Magazin test (from September 2014), nor the pro's who use the 5D mk III and 5Ds and tell about its focus acquisition performance. And no, they do NOT shoot 10 images in the hope that a few maybe nailed focus, with a 85mm f1.2, 50mm f1.2, 200mm f2.

 

No, when I go shoot flowers with my 6D + EF 70-200mm f4 lens I do not get a high percentage of images that are OOF. No, I don't shoot every subject 4 times in the hope at least one is in focus. And no, I have not done any MFA on my 6D. Maybe your friend should try a 1D-X.

 

On monday, I shot a quick snapshot of my girlfriend's dog while I was walking the dog. With my 6D and my EF 35mm f2 wide open, just to try out that lens again on my 6D. I just now checked how the focus was nailed:

[ATTACHMENT NOT FOUND]

 

Sorry that that lens is not sharper wide open. It is a very old design and can't keep up with the new 35mm f1.4's , f1.8's and f2's from Canon, Sigma, Tamron. Or the 23mm from Fuji or the 22mm f2 on my EOS M. It is sharp enough anyway (will sharpen fine in PP I am sure). I still keep it because it is so small and light, just a fun little lens.

Maybe I should have shot it 10 times in the hope the focus would be OOF?

[JJ_SO]

thxbb12, I read you and to me your explanation makes a lot sense. For critical focus DSLR users usually recommend LiveView, which on any DSLR is painfully slow. Maybe the mirrorless systems are judged by this experience, I don't know. I'm sure mirrorless will improve while DSLR is at the end of the possibilities.

 

There are so many factors, not the least mechanical tolerances of all the parts involved in AF process, that the conclusion can only be "indirect systems like PDAF will always suffer from more flaw sources than any mirrorless". It was easy to repair and build up this "superior AF" fairytale as long as the lenses were not that fast and the sensors were high resolution with 8 MP. These days it gets much harder.

[thxbb12]

Ah, here you go. So between your Canon bodies and various lenses, you pretty much had random luck in terms of accuracy.

This is exactly what I'm talking about.

I tested about 8 MFT lenses on more than 4 bodies. All lenses with every single body focused pretty much perfectly. There is simply no AF accuracy discrepancy.

Now I shoot Fuji. I have 5 lenses. All 5 lenses focus spot on. Again, no AF accuracy discrepancy.

With a DSLR it's just a nightmare to not know how well or bad a lens will focus before buying it. Fiddling with micro-adjustments, testing procedures, etc. With mirrorless the issue is simply non-existent.

 

This is what I mean by reliability. The same simply doesn't hold true with a DSLR, as you stated yourself.

 

Anyone who is objective and reasonable would come to the same conclusion.

Why is it so difficult for you to admit that AF on a mirrorless camera is more reliable than AF on a DSLR?

 

On a final note, can you please answer these questions in all honesty:

• On a DSLR, which is more accurate (when shooting a still subject): PDAF or CDAF?
  
• Isn't it true to say that Canon and Nikon DSLRs provide AF micro-adjustments as an attempt to correct AF accuracy issues?
  

[joachim]

 

On monday, I shot a quick snapshot of my girlfriend's dog while I was walking the dog. With my 6D and my EF 35mm f2 wide open, just to try out that lens again on my 6D. I just now checked how the focus was nailed:

donnacrop.jpg

I am sorry, but on my books, this would be a reject, due to either of:

• out of focus
  
• motion blur
  
• poor lens
  

For a keeper I am expecting better from my µFT, unless something else in the image is spectacular (e.g. timing).

 

I forgot, that I routinely use apertures in the 1.4 to 1.8 range.  2.2 is stopped down .  That gives, as you never tire, 4.4 in old money (FF).

[Brightcolours]

 

Ah, here you go. So between your Canon bodies and various lenses, you pretty much had random luck in terms of accuracy.

This is exactly what I'm talking about.

I tested about 8 MFT lenses on more than 4 bodies. All lenses with every single body focused pretty much perfectly. There is simply no AF accuracy discrepancy.

Now I shoot Fuji. I have 5 lenses. All 5 lenses focus spot on. Again, no AF accuracy discrepancy.

With a DSLR it's just a nightmare to not know how well or bad a lens will focus before buying it. Fiddling with micro-adjustments, testing procedures, etc. With mirrorless the issue is simply non-existent.

 

This is what I mean by reliability. The same simply doesn't hold true with a DSLR, as you stated yourself.

 

Anyone who is objective and reasonable would come to the same conclusion.

Why is it so difficult for you to admit that AF on a mirrorless camera is more reliable than AF on a DSLR?

 

On a final note, can you please answer these questions in all honesty:

• On a DSLR, which is more accurate (when shooting a still subject): PDAF or CDAF?
  
• Isn't it true to say that Canon and Nikon DSLRs provide AF micro-adjustments as an attempt to correct AF accuracy issues?
  

 

You are funny. The 350D was a DSLR from 2005, and it was famous for its temperamental AF (could have been a Pentax ). The lens was a Sigma with a simple DC motor, which gave terrible results with live view on the 450D.

The Tokina, what can I say. It was a Tokina. More crude AF motor action can't be found.

 

Yes, my 28-135mm Canon lens is defective. Is that defective old zoom a good example of how big aperture lenses on modern Canon DSLRs perform? No, of course not. 

[goran h]

Don´t discuss with BC, he is right even when he is wrong and will always have the last word.   

[thxbb12]

You are funny. The 350D was a DSLR from 2005, and it was famous for its temperamental AF (could have been a Pentax ). The lens was a Sigma with a simple DC motor, which gave terrible results with live view on the 450D.

The Tokina, what can I say. It was a Tokina. More crude AF motor action can't be found.

 

Yes, my 28-135mm Canon lens is defective. Is that defective old zoom a good example of how big aperture lenses on modern Canon DSLRs perform? No, of course not. 

 

Of course.

 

Where were we? Can you please answer these questions in all honesty:

• On a DSLR, which is more accurate (when shooting a still subject): PDAF or CDAF?
  
• Isn't it true to say that Canon and Nikon DSLRs provide AF micro-adjustments as an attempt to correct AF accuracy issues?
  

[Brightcolours]

 

Of course.

 

Where were we? Can you please answer these questions in all honesty:

• On a DSLR, which is more accurate (when shooting a still subject): PDAF or CDAF?
  
• Isn't it true to say that Canon and Nikon DSLRs provide AF micro-adjustments as an attempt to correct AF accuracy issues?
  

 

With above lenses, which you deemed a good example (the Sigma 18-50mm f2.8 and the Tokina 12-24mm f4)? PD AF is waaaay more accurate. And that is the whole truth and nothing but the truth.

 

Your 2nd question is a very leading one, and not a honest one. 

 

If you read posts on forums a lot, you will notice that very seldom there is talk about MFA in Canon FF forums (way less than for instance there is on Nikon FF forums with cameras like the D800). 

Besides that, Canon (and probably Nikon) says that MFA is there for emergencies, when you have a lens calibration issue and can't have it calibrated at that time.

 

It is also true that CD AF can deliver OOF results, and that that is not just in low light. 

 

This is getting tiresome, though. The simple fact is that (at least Canon DSLRs, but also Nikon DSLRs see Dave's experiences) can focus just fine with large aperture lenses with shallow DOF, and have no issue getting an eye in focus, not even with a 85mm f1.2, a 50mm f1.2, or a 200mm f2. And that one does NOT have to shoot several frames "in the hope that one is in focus".

 

I can't help that you have made bad gear choices, apparently, in the past. That does not excuse the FUD posts about DSLRs and PD AF in general, though.