7D Mark II - Focus Discussions

I'm still going stand behind what I have been saying. The camera AF system does 80 to 90% of the work for you. You can only tweak those settings for the shooting conditions.

Even Canon says that in the first minute in the AI Servo I posted on a few threads. The presenter stated "you may never need to use more than case 1".

There was a thread where a member was shooting hockey games and stated the AF issues were user error. The system still focused on something, whether it was a player or the boards. The member got some good info and tweaked the settings to ignore the boards.

This is what the cases and parameter tweaks are all about. When your compare the parameter changes in each case to each other they are only tweaks. They all have the same 3 parameters.

If in case 2 and you set tracking sensitivity to -2 from -1 on the 7D2, the 5D3 and the 1dX all three will put more emphasis on ignoring everything entering the scene and keep AF locked on the subject.

However I think because of the processor speed and possible different algorithms in the 7D2 compared to 5d3 the settings may act diffently. Be more sensitive and not require as a big adjustment to achieve the same results, but they still do the same thing. So you may not be able to use the exact settings for both but they will be so close a user may not even know the difference. You would probably need equipment to measure that.

I'm just basing this on canons videos and my experience with my 5D3.

nsnqst wrote in post #17350833
Spot on. I second that. The moment I noticed focus issues with my second copy, I tested the lenses with dot tune and they were off on the 7D2; exact same lenses performed flawlessly on the 5D3/

Nsnqst- are you indicating that you dot tuned and then went on to further test and return the cameras; OR that once you saw mfa was required with dot tune you returned them because they needed adjustment on the 7dii (but not on your other cameras) so there was something wrong with the body just for needing adjustment?

Just in case anyone thinking I'm saying going past case 1 or 2 is easy for more complex shooting conditions it is not. It does take time to learn and pick a case that works best for you and your shooting style based on what you are shooting.

No matter which case you pick the camera is going to AF on something. The parameters just let you dictate how it reacts after it does. We only have some control over AF acquisition characteristics in 1st and 2nd priority.

gschlact wrote in post #17350934
Nsnqst- are you indicating that you dot tuned and then went on to further test and return the cameras; OR that once you saw mfa was required with dot tune you returned them because they needed adjustment on the 7dii (but not on your other cameras) so there was something wrong with the body just for needing adjustment?

With both the 7D2 bodies, I went out shooting. Static objects as well as birds in a pond and some BIF. the results were consistent no matter what situation - inconsistent focusing.

It is after that is when I did AFMA. I used a couple of methods (1) high contrast subject with a ruler & (2) dot tune. The issue with the first was that it was time consuming and subject to error (relatively speaking). Dot tune is more accurate because when you focus through LV, switch the lens to MF, change to VF and press AF on, it will instantly tell you whether it focuses or not. When it did not give me focus confirmation, I tried AFMA and that was inconsistent - meaning if I did a +15, that did not give me a high keeper rate. When I tried dot tune again, the MFA value had to be different.

Those exact same lenses perform well on the 5DM3 and need no AFMA (Canons - 50/ 1.4, 70-200 f4 IS, 24-105 f4, 400/ 5.6 and Tammy 150-600). That tells me the 7D2s are inconsistent in focusing.

If one of the 7D2s focused consistently after AFMA and gave me a better keeper rate, I would have had no reason to return it.

digital paradise wrote in post #17350842
There are only two differences. The obvious that is iTR. In single shot it face recognition and in AI Servo it along with whatever system it uses to acquire and lock AF it also remember colours.

Unfortunately canon provides very little information about pros and cons. Art Morris tried it and thought it worked well and others were not impressed. On another site someone mentioned it would be nice if someone like Art would do a write up on it.

The other non obvious is camera shake. Due to pixel density the 5D3 is more forgiving so the 7D2 requires the shooter to be more aware of that. A term they coined at FM as micro blur. Liquidstone commented a about that on that thread. Not a huge factor but a person should be aware you need to be steady.

The AF menus are identical with both bodies except for the processor speeds with 7D2, etc that the user has no control over.

I turned iTR off pretty soon after getting the camera. The blur from camera shake has been my biggest source of bad images. Not surprisingly, it's an issue with my non-IS 400 5.6, but not with my IS lenses.

Just from reading random bits and pieces from people who compared it to the 1DX are saying iTR is not great for fast objects coming at you but ok moving side to side. Theses are just people's observations so it is difficult to know for sure. Controlled tests with measurable results would be the best.

I would think it would do better with a person running wearing a red jacket than a grey bird flying in overcast sky but I did not design the system is this is just a guess.

Another person stated it was great for video recording. Again no detailed info for canon. I'm not into video but I imagine you would be on AI Servo so it would not be in face recognition.

Controlling camera shake a you noticed is important with a sensor packed with so many pixels.

iTR sounds good in practice, but it would have to be extremely advanced to be able to pickup on the eyes of a bird,
and keep tracking it properly, while the birds moving/ect.

Especially when birds often are the same/similar colors to the background, or there is stuff moving in the background.
I've had iTR work really well in a couple cases, but usually it likes to jump all over the place, or get stuck on the wrong spot.

The video AF however can track pretty well. You can lock onto something specific , and it is not bad.


I really hope when I send mine in, they find the AF sensor or something not aligned properly and get it fixed.
If it comes back the same way, I'm probably going to have to sell it somehow.

Canon should be able to correct it and I would not just settle for it if it is still off. Send it back and push them.

When talking to a canon rep about my 7D2 I mentioned my 700-200 II. It was never quite right from day one and I did send it to NJ about 18 months ago under warranty. I'm not what they did to it but it still seemed off. He suggested I send it to Newport which has the latest facilities.

I did and since off warranty it would be costly. As he suggested I escalated it and I brought up the old RN number and they gave me a very good reduction. I got the shipping notice the other day so I called them out if curiosity. A circuit board was out of calibration causing mis-focus. I'm picking it up on Monday.

Just to point out this was not my primary test lens, it was the first time I have ever sent a lens in for warranty work and the first time NJ let me down. They have been pretty good in the past.

So never give up and stand up for your right to receive what you paid for. That whole thing with canon went well. They were very accommodating. By escalating I did not mean getting mad, just moving up the ladder until someone could do something about it.

Yeah, I've sent things to Canon and Tamron for repair, and both were very quick to repair and send back, with no issue.
I've been sending my stuff to California I believe (Canon).

But I have to send Tamron to NY, which is too bad. I'm in Seattle, so it's the total opposite side of the country.

digital paradise wrote in post #17350915
Even Canon says that in the first minute in the AI Servo.

Here it is. The presenter states optimize and I said fine tune.

https://m.youtube.com/​watch?v=7sgRZBo-WKI

This is the kind of result I can consistently re-produce with my 2nd Mark II.
Random focus in Live View, on a tripod, center point only , One Shot, on a static target. I rack focus, try to refocus, it says "it's good", but it's not.

I wasn't seeing issues like this with my 60D or Nikon D750 with the same lens.

I will get varying degrees of focus. Sharp, very soft, soft, sharp, ect. It's definitely going in. No more testing.
I'll keep using it, but when I return the Nikon D750 rental, this is going at the same time. Probably my Tamron 150-600 as well.

I know the shutter speed is slow, but this isn't motion blur. I was using mirror lockup and a timer as well, with nothing moving around me.
I could also see it being outta focus on the LCD before taking the shot.

100% view. Target about 12ft away or so

I posted this on the 1D4 vs. 7d2 thread, but I'll add my second thoughts here.

Today was the first day I've had the chance to get out and shoot my 2nd 7d2. Overall, I thought focusing was much better this time with the same settings. However, even after MA'ing my 600 with and without TCs, the image quality remains subpar in my opinion.

Here's my comparison with the 1D4 - https://photography-on-the.net …showthread.php?​p=17353462.

After examination of what I brought home today and comparison with previous shots from my 1D4 under the same shooting conditions, I've decided the 7d2 simply isn't for me. Be it the AF inconsistencies or the IQ (IMO lack there of), it just isn't ringing my bell.

Kickflipkid687 wrote in post #17353302
This is the kind of result I can consistently re-produce with my 2nd Mark II.
Random focus in Live View, on a tripod, center point only , One Shot, on a static target. I rack focus, try to refocus, it says "it's good", but it's not.

I wasn't seeing issues like this with my 60D or Nikon D750 with the same lens.

I will get varying degrees of focus. Sharp, very soft, soft, sharp, ect. It's definitely going in. No more testing.
I'll keep using it, but when I return the Nikon D750 rental, this is going at the same time. Probably my Tamron 150-600 as well.

I know the shutter speed is slow, but this isn't motion blur. I was using mirror lockup and a timer as well, with nothing moving around me.
I could also see it being outta focus on the LCD before taking the shot.

100% view. Target about 12ft away or so

Was this shot of a wall and perpendicular to the camera.?
If so, what were you trying to test? The dof at 10ft 600mm and f7.1 is 0.02inches total. Even assuming you did mfa, probably not the most realistic test. If no mfa was done all this shows then is that mfa is necessary. If you did mfa, try to repeat but at an angle to the wall, I bet you'll find focus much close than you thought.

Yeah, the tripod was level, camera was the same height as the railing I was looking at,
and straight on from what I could tell. But I can try to be even more accurate.

I can't MFA the lens, since it's random focus. It's never the same amount outta focus.
I can re-tap the AF button, and basically roll dice and see what it does.

Kickflipkid687 wrote in post #17354091
Yeah, the tripod was level, camera was the same height as the railing I was looking at,
and straight on from what I could tell. But I can try to be even more accurate.

I can't MFA the lens, since it's random focus. It's never the same amount outta focus.
I can re-tap the AF button, and basically roll dice and see what it does.

That's probably the surest way to confirm a defective system, so long as you have a very high contrast focus target that the system should not be bouncing from front to back on. I've found that just shooting outdoors stuff is not a good way to test, because the AF system can be working perfectly correctly and still jump around from the front to back of something like a leaf, depending on where it gets the highest contrast at the instant of focus; that, combined with the fact that the focus sensors are larger than the focus squares in the viewfinder, may account for a lot of people's concerns. I don't think, however, it's the reason you're getting such poor results, especially since your sample above is a flat surface, and is out of focus then in focus with the same setup. It seems like there are some 7DIIs that got out with incorrectly adjusted or possibly even incorrectly installed AF components.