Major issue with eye-af and af-c with a7iii?

[ATLshooter]

The other thread filled up, so I was instructed to make another thread. Here's a link to the old thread here. (click me)

The issue still remains. I own 2 bodies now, both have the same issue. There are now 6 different people that have reported the issue. (between here, facebook, and fredmiranda)

We have tried all settings. We have tried everything. Eye-af just front focuses nearly 100% of the time, and using AF-C (in any focus area) results in a hit rate of about 5%.

This thread is simply a continuation of that previous thread for discussion and further testing.

 

[tesilab]

Two cameras and which lenses, again?

Is there a summary of anyone who reports a similar problem? My A7iii and Sony 85 1.8 nail it with no evidence of front focusing.

 

[ATLshooter]

Two cameras and which lenses, again?

Is there a summary of anyone who reports a similar problem? My A7iii and Sony 85 1.8 nail it with no evidence of front focusing.

Please see the link to the other thread. Copious amounts of information in there. I'd like to keep this thread about new information and new ideas if possible.

But:

- Two different bodies.

- Both bodies tested with 2 55mm zeiss 1.8s (I have two of those) and one Zeiss 85mm 1.8. Lens has no effect.

There are 6 total people reporting the problem, but I haven't compiled them into a list. The ones that have posted pictures look EXACTLY like the pics I'm getting.

 

[owlgroot]

6 people out of how much owners on here, Facebook and FredMiranda? If you have faulty device get it replaced by your retailer.

This thread is overly dramatic.

 

[ATLshooter]

6 people out of how much owners on here, Facebook and FredMiranda? If you have faulty device get it replaced by your retailer.

This thread is overly dramatic.

First off, it's only 6 people who have taken the time to come forward. I'm sure there are plenty more that either haven't discovered the issue yet, or don't post online.

Secondly, I own 2 different bodies now, and both do the same thing. The guy on fredmiranda has tested THREE different bodies, all with the same issue. It seems to be pretty widespread.

Thirdly, how exactly is this over dramatic? Is it over dramatic to expect my two 2000 dollar cameras to work? There is no evidence that getting a replacement would be any different.

If you only want to see praise and warm-and-fuzzies on this forum, then I don't know what to tell you. This is a genuine problem that needs to be addressed. If you don't like it, then why post in this thread?

It seems like you're the over dramatic one, since you came in here only to whine. Let's get back on track...

 

[tesilab]

You need to compile a list of serial numbers of affected cameras. Then you should look for possible ranges and ask others with cameras in that range whether they have the issue. Compare against all serial numbers of cameras that users are sure are unaffected.

It doesn’t sound like lenses are implicated in this case. If it was a lens firmware issue it would be extremely unlikely to affect a recent model Sony 85 1.8.

 

[ATLshooter]

You need to compile a list of serial numbers of affected cameras. Then you should look for possible ranges and ask others with cameras in that range whether they have the issue. Compare against all serial numbers of cameras that users are sure are unaffected.

It doesn’t sound like lenses are implicated in this case. If it was a lens firmware issue it would be extremely unlikely to affect a recent model Sony 85 1.8.

Honestly? I feel like Sony should be doing that. Troubleshooting Sony's cameras isn't my job. I have a full time job, side photography work, and a 13 month old daughter. Spending my (little) free time to compile and cross check lists for Sony is pretty low on my to-do list.

I have however been directing those who come forward to call Sony.

But, so far there doesn't seem to be any serial correlation. Probably because it's a FW issue. Regardless, I have two cameras and one starts with 337 and the other 377. The others that report the issue also fall into those numbers, but there are others with working cameras (who have posted pics to prove) that also have those numbers. So there doesn't seem to be a correlation so far.

 

[owlgroot]

Thirdly, how exactly is this over dramatic? Is it over dramatic to expect my two 2000 dollar cameras to work? There is no evidence that getting a replacement would be any different.

No but complaining on forum achieves nothing. Either replace your bodies or get refund and buy something different.

 

[ATLshooter]

Examples of how neither of my bodies can get shots in focus using af-c, regardless of focus area.

Both bodies perform the same. I do these tests back to back.

 

[ATLshooter]

Thirdly, how exactly is this over dramatic? Is it over dramatic to expect my two 2000 dollar cameras to work? There is no evidence that getting a replacement would be any different.

No but complaining on forum achieves nothing. Either replace your bodies or get refund and buy something different.

I beg to differ. I'm not complaining, I'm gathering information.

Around 2 hours ago I got a call from Sony and they actually thanked me profusely for making these threads and asked me to email the link to them so that they could review the information we had all gathered. So, while you may not like it (for whatever reason), Sony seems to be grateful for it.

Again, it's an easy solution for you. Don't like it? Don't open the thread. You're literally just coming in to complain about what you perceive to be complaining. Do you not see the irony in that?

 

[Lightfinity]

On our a7Riii I contacted pro support and did a test finding that from 113 images with three different lenses with a stationary subject and decent light, there was about a 15% overall "miss rate" with 5% bad miss rate (10% near miss where it appears to hit the eyebrow). They confirmed on their Riii and Rii that the rates are comparable and the camera is not considered "defective". My solution: if the shot is important take two captures if I can. Hopefully they will have an improvement in a future firmware update.

Thinking about it, this is probably what I used to get the with Rii, but the a9 is so much better my expectations have gone up. It's still at least as good and generally better than I would get using a center-recompose strategy trying to catch someone's eye with any other system, and most of the near misses are acceptable for smaller prints or web resolution. This means the "hit rate" is about 95%, but the "critical hit rate" is closer to 80-85%.

It seems maybe the 7iii is having a lower critical hit rate?

 

[owlgroot]

Around 2 hours ago I got a call from Sony and they actually thanked me profusely for making these threads and asked me to email the link to them so that they could review the information we had all gathered.

I stand corrected. I didn't know Sony cared about what people say on forums.

 

[kaku]

mobilejunkie wrote:
No but complaining on forum achieves nothing. Either replace your bodies or get refund and buy something different.

I think on the contrary. If anything, it brings awareness in case others may be impacted. At the very least, it gives people an excuse to further play with their new toys. =]

 

[Horshack]

I'm coming to this late but I just read the other thread so hopefully I won't be asking anything already known.

At first the issue was thought to be eye-AF specific, but after some experimentation it sounds like you've isolated it further to be an AF-C issue, which eye-AF happens to require as part of its implemention.

If that's the case then PDAF would be the prime suspect. AF-S uses PDAF as well (when adequate lighting is available), but combines it with a CDAF confirmation afterwards, correcting (and hiding) any misfocus from the PDAF operation.

If this is a PDAF-specific issue it could be anything in the PDAF chain, starting from the PDAF pixels on the sensor (including any factory/firmware calibration), up to the electronics and logic driving the lens.

Let me know if my summation is correct. If so I can think of some AF experiments to help isolate the problem further.

 

[The Lotus Eater]

Just for clarity, is this no longer about just Eye-AF, but AF-C in general?

My understanding from your previous thread and posts on other forums was that the issue was confined to Eye-AF focusing.

Based on your video, you seemed to be able to focus on your partner when she was stationary, yet others have reported focus problems with Eye-AF on stationary subjects.

Forgive me for being a bit confused.

 

[ATLshooter]

Just for clarity, is this no longer about just Eye-AF, but AF-C in general?

My understanding from your previous thread and posts on other forums was that the issue was confined to Eye-AF focusing.

Based on your video, you seemed to be able to focus on your partner when she was stationary, yet others have reported focus problems with Eye-AF on stationary subjects.

Forgive me for being a bit confused.

it started as being a problem with eye af. In trying different things with the cameras I realized that af-c doesn’t seem to be able to focus either.

So af-c is front focusing in any mode. Whether it be eye af or not.

 

[XeroJay]

My A7III is nailing eye focus most of the time using Canon lenses and AF-C.

 

[Marky12]

Examples of how neither of my bodies can get shots in focus using af-c, regardless of focus area.

Both bodies perform the same. I do these tests back to back.

Why are you using spot focus? My understanding if it's anything like the a6000, af-c and spot uses contrast detection. Afc and wide uses phase detect and flexible spot to some extent . Can you use wide/afc and then eye detect? I have mine set up with back button af so not sure if that is any different

The only other thing I can thing of is how face back your eyes are maybe affecting it?

I literally took a selfie with it with eye detect and it nailed my eyes.

--i

李 宝 華

Explore 李 宝 華’s 273 photos on Flickr!

www.flickr.com

 

[ATLshooter]

Examples of how neither of my bodies can get shots in focus using af-c, regardless of focus area.

Both bodies perform the same. I do these tests back to back.

Why are you using spot focus? My understanding if it's anything like the a6000, af-c and spot uses contrast detection. Afc and wide uses phase detect and flexible spot to some extent . Can you use wide/afc and then eye detect? I have mine set up with back button af so not sure if that is any different

The only other thing I can thing of is how face back your eyes are maybe affecting it?

I literally took a selfie with it with eye detect and it nailed my eyes.

--i
https://www.flickr.com/photos/22946100@N02/

Did you watch the videos? I’ve used them all. I’ve used small spot. I’ve used medium spot. I’ve used medium lock on. I’ve used face detect. I’ve used center lock on.

Ive used. Them. All.

The two videos above were specifically using spot AF and face detect af per the recommendations from forum members. To answer your question.

 

[RussKNY]

Examples of how neither of my bodies can get shots in focus using af-c, regardless of focus area.

Both bodies perform the same. I do these tests back to back.

Strong back lighting like that can often confuse AF (on any camera). Did you by any chance try having your subject walk towards that window instead, so she isn't backlit?