The Amazing Sony A1/A7/A9/APS-C & Anything else welcome Mega Thread!

jj_glos said:
If you can manually focus a lens and it works, there isn't a physical problem with the lens. It means the lens is sharp and accurate. Once you start using AF it is driven by the camera, and the technical computational ability of the camera to determine if the subject is in focus. The camera controls the movement of the AF mechanism in the lens, it moves this mechanism until the camera thinks the image is in focus. The rules around how the camera decides the image is in focus is the crux of the issue as far as I can see?

I don't agree that any lens should be out on a mirrorless body, the tech should be preventing that if used properly. MFA on a DSLR, yes we know it happens but manual or live view AF have always been the accurate constant control that hit 100% every time.

I would love to know how a camera body knows where focus should be and tells the lens what to do, it needs to take the scene/AF point into account, focal length, aperture, subject etc. The lens feeds the camera information, the camera body makes the decision. This is based around me thinking that the end focus result is all based around the image and best focus of the highlighted area under the AF point, as you would manually... If the camera doesn't do this, then there will always be problems.

You may be right though, the camera may use prediction rather than verification, and the algorithm it uses just doesn't work for all lenses? That kinda makes sense to me. So possibly this is the root of the issue?

Edit: If the camera, in the pursuit of speed takes stock of the situation and decides the lens needs to move 'x' number of clicks and the scene will be in focus rather than actually verifying it is in focus, then yes if the translation of clicks to the lens is out for some reason, it's all going to end badly :D You'd think in single AF it would take it's time and make sure, having inconsistencies if C-AF I'd expect as that is generally about prediction. I wonder if pre-AF plays any part?

I've never seen this behaviour with Fuji, or Canon where using live view is the control that you set the MFA by when using software like FoCal.

Edit: Full disclosure, I'm about ten pints of Scrumpy Jack in, so this may all make no sense at all :D
Click to expand...
I agree that the camera decides where the focus should be, but as I mentioned earlier maybe certain lenses do not action this command correctly? So maybe the camera says to the lens right you need to focus at 3.25m yet the lens focuses at 3.22m? If a camera can correctly focus properly with one lens it doesn’t therefore make sense that it can’t with another one. I also have lenses that have the same front focus errors on the three different bodies I’ve used it on, how do you explain that? Why do sigma and Tamron allow you to change the focussing of the lens via a dock.

I’m not saying I’m right, but can you understand why I believe the lens may also be responsible for the final focus? Don’t forget mirrorless now use complicated phase detect systems and (in theory) are now prone to similar focus errors to DSLR.
 
So the E mount art lenses are out and still no reviews on them or mentions anywhere. Anyone here received theirs yet? I'll have mine on Tuesday, would love to see a review or users thoughts
 
snerkler said:
I agree that the camera decides where the focus should be, but as I mentioned earlier maybe certain lenses do not action this command correctly? So maybe the camera says to the lens right you need to focus at 3.25m yet the lens focuses at 3.22m? If a camera can correctly focus properly with one lens it doesn’t therefore make sense that it can’t with another one. I also have lenses that have the same front focus errors on the three different bodies I’ve used it on, how do you explain that? Why do sigma and Tamron allow you to change the focussing of the lens via a dock.

I’m not saying I’m right, but can you understand why I believe the lens may also be responsible for the final focus? Don’t forget mirrorless now use complicated phase detect systems and (in theory) are now prone to similar focus errors to DSLR.
Click to expand...

I just don't get the logic of the camera focusing by distance when it's using either contrast detect or pdaf methods? Pdaf is fast and potentially on it's own could cause the issue as it detects front/back focus and tells the lens how far to move in order to bring it into focus. However I was under the impression that contrast detect was used to fine tune the AF at the end? Perhaps this isn't the case with Sony though, or at least the A7III? Is it possible to force the camera to use contrast detect at all?

I don't know about Tamron, but Sigma have been producing the dock for a number of years for DSLR and it would make sense that the functionality would be the same.
 
F/1.4 said:
So the E mount art lenses are out and still no reviews on them or mentions anywhere. Anyone here received theirs yet? I'll have mine on Tuesday, would love to see a review or users thoughts
Click to expand...

Yeah, Im surprised by this, unless UK/US are getting stock first so there aren't any in the wold. Sigma hasn't given them to testers yet which is a bit odd.
 
twist said:
Yeah, Im surprised by this, unless UK/US are getting stock first so there aren't any in the wold. Sigma hasn't given them to testers yet which is a bit odd.
Click to expand...

I've alot of work this month. I'm praying they're alright haha
 
Hi folks, looking for some ideas re a problem with my A7rii...

I'd been having problems with the Auto function that switches viewing from momitor to evf automatically. Due to this, I selected either EVF or Monitor in the menu when shooting. Now, when I switch on the camera, I get no image either on the monitor or EVF. The camera seems to work fine, it makes all the right noises when I press the shutter or shoot a video, and takes pictures of whatever the camera is pointing at. I'm not sure which was selected last time I used the camera, but I am assuming either the EVF or the monitor has packed up, with the faulty display selected.

I can't find anything online about how to do a hard reset. I'm assuming a hard reset would reset the monitor/evf selection to Auto. Does anyone have any ideas? For the moment, the camera is effectively bricked.
 
jj_glos said:
Sorry, I didn't realise that it was an assumption that the camera body determines whether AF has been secured or not whilst controlling the lens. It's how all AF cameras I have ever used have worked, sorry I assume they have worked. Must be some other non AF acquisition related way that they work then. :D
Click to expand...

You keep saying this yet still not knowing how it actually works, don't misunderstand I am not claiming to, I'm simply saying we don't have enough information.

jj_glos said:
I am interested in how you think they work, purely in a functional layman's way.
Click to expand...

I don't know, why waste time assuming how it should work?

jj_glos said:
The AF performance, with regards to speed and operation in low light can obviously be down to a lens as that is where the AF motors are and the lens is responsible for transmitting the light/image that the camera uses to determine focus. It's easy to see how this might result in no focus, or missed focus but consistent front focus? Not sure about that myself.
Click to expand...

Yes but we don't know how the two communicate. If it's as described and the camera does everything and just reads focus from the sensor then it shouldn't be possible to get front/back focus should it?

jj_glos said:
None of the mirrorless systems or even non mirrorless systems using live view have ever shown anything but 100% accuracy in single shot AF with good light. Not needing MFA with mirrorless is a headline feature, that and WYSIWYG is pretty much the main reason I like mirrorless.
Click to expand...

Certainly.
 
simonbarker said:
You keep saying this yet still not knowing how it actually works, don't misunderstand I am not claiming to, I'm simply saying we don't have enough information.



I don't know, why waste time assuming how it should work?



Yes but we don't know how the two communicate. If it's as described and the camera does everything and just reads focus from the sensor then it shouldn't be possible to get front/back focus should it?



Certainly.
Click to expand...

There are plenty of sources to read up on how pdaf/cdaf work and are employed, but you are right in that I am assuming Sony do the same. They don't appear to document how their implementation works. That is what I am questioning though, at a basic level a DSLR uses cdaf in live view and whilst slow is very accurate. The cdaf process is pretty straight forward and pdaf improves on this speed wise.

Found quite an interesting blog (also talks about the 35mm f1.4 copy variation) here:

https://blog.kasson.com/category/a7riii/

On page 2 there is an AF accuracy summary which mentions the focus shift on a number of lenses. Maybe this process might be part of the problem? I dunno, we'll see how things pan out.
 
jj_glos said:
There are plenty of sources to read up on how pdaf/cdaf work and are employed, but you are right in that I am assuming Sony do the same. They don't appear to document how their implementation works. That is what I am questioning though, at a basic level a DSLR uses cdaf in live view and whilst slow is very accurate. The cdaf process is pretty straight forward and pdaf improves on this speed wise.

Found quite an interesting blog (also talks about the 35mm f1.4 copy variation) here:

https://blog.kasson.com/category/a7riii/

On page 2 there is an AF accuracy summary which mentions the focus shift on a number of lenses. Maybe this process might be part of the problem? I dunno, we'll see how things pan out.
Click to expand...

Wasn't focus shift with the Samyang 35 1.4, it was at all distances. I could fine tune the AF using the dock so I could force a huge front or rear focus which nobody thought was possible with mirrorless. Tolerances just like DSLR lenses it seems, maybe more so with 3rd party because the QC isnt as good and they dont have all of Sonys AF algorithms figured out yet.
 
twist said:
Wasn't focus shift with the Samyang 35 1.4, it was at all distances. I could fine tune the AF using the dock so I could force a huge front or rear focus which nobody thought was possible with mirrorless. Tolerances just like DSLR lenses it seems, maybe more so with 3rd party because the QC isnt as good and they dont have all of Sonys AF algorithms figured out yet.
Click to expand...

Yeah, I don't think it would be focus shift to be honest. As when shooting wide open it wouldn't be a factor.

That's the thing though, as you mention DSLR it's completely different in that context. You can have a whole bunch of lenses that need MFA, and the way you work through that process? You use live view with on sensor AF as it's accurate and consistent (at least in my experience with Canon, haven't used MFA with Nikon). Basically no matter the MFA setting, using live view to focus on a DSLR still hits every time. You've proved with your testing there is a disconnect in this process with Sony somewhere as you can force front/back focus. I might test this with my Sigma lens on Canon (use the dock to see if I can make it front/back focus even when using live view).

Were your AF problems the same whatever the distance? (Edit: sorry, you already said that!) At least if that's the case any MFA fix would at least be easy to apply.
 
Last edited:
jj_glos said:
Yeah, I don't think it would be focus shift to be honest. As when shooting wide open it wouldn't be a factor.

That's the thing though, as you mention DSLR it's completely different in that context. You can have a whole bunch of lenses that need MFA, and the way you work through that process? You use live view with on sensor AF as it's accurate and consistent (at least in my experience with Canon, haven't used MFA with Nikon).

Were your AF problems the same whatever the distance? (Edit: sorry, you already said that!) At least if that's the case any MFA fix would at least be easy to apply.
Click to expand...

Yup, the MA fix is very easy via dock, set from +4 to -4, click of the mouse, its just the testing that takes a little time but at least its not between +20 and -20!
 
twist said:
@jj_glos Heres an example, 3 clicks difference....

My custom adjust was -3 (left), right is at 0. Night and day difference. This was one of hundreds of shots at various distances. Focus on left hand side branch.

https://www.flickr.com/photos/87090421@N02/40575895340/sizes/o/
Click to expand...

Yeah you've basically proved that the lens can be made to front/back focus, so I'll shut up now :D I think it's a worrying development but at least with Samsung, Tamron and Sigma it looks like you can fix it. I wonder how Sony will address it?
 
jj_glos said:
Yeah you've basically proved that the lens can be made to front/back focus, so I'll shut up now :D I think it's a worrying development but at least with Samsung, Tamron and Sigma it looks like you can fix it. I wonder how Sony will address it?
Click to expand...

Tamron and Sigma dont have FE docks afaik, Samyang only so far, I think the others will follow. Its not great but fixable, I dont mind spending time on a lens if its spot on when tuned.
 
Last edited:
twist said:
Tamron and Sigma dont have FE docks afaik, Samyang only so far, I think the others will follow. Its not great but fixable, I dont mind spending time on a lens if its spot on when tuned.
Click to expand...

Ah right, are the Sony fit docks available just in A mount at the moment then? (Edit: quick Google and it looks like that is the case). Yeah, if it's a one time fix I can live with that.
 
Last edited:
jj_glos said:
I just don't get the logic of the camera focusing by distance when it's using either contrast detect or pdaf methods? Pdaf is fast and potentially on it's own could cause the issue as it detects front/back focus and tells the lens how far to move in order to bring it into focus. However I was under the impression that contrast detect was used to fine tune the AF at the end? Perhaps this isn't the case with Sony though, or at least the A7III? Is it possible to force the camera to use contrast detect at all?

I don't know about Tamron, but Sigma have been producing the dock for a number of years for DSLR and it would make sense that the functionality would be the same.
Click to expand...
My understanding is that PDAF uses distances and movement as well as a way to help with tracking motion, it uses complicated software to 'predict' where the next shot should be. It's not simply a case of bringing two images together (although this is part of the process). I thought that Sony finished off with contrast detect too but my recent reading suggests that it's purely PDAF on the latest bodies which is what allows it to be as fast as DSLR, contrast detect still can't track as fast as PDAF. In AF-S it may well still be finished off with contrast detect, I don't know.
 
snerkler said:
My understanding is that PDAF uses distances and movement as well as a way to help with tracking motion, it uses complicated software to 'predict' where the next shot should be. It's not simply a case of bringing two images together (although this is part of the process). I thought that Sony finished off with contrast detect too but my recent reading suggests that it's purely PDAF on the latest bodies which is what allows it to be as fast as DSLR, contrast detect still can't track as fast as PDAF. In AF-S it may well still be finished off with contrast detect, I don't know.
Click to expand...

In AFS it uses contrast detect to finalise, thats why its much slower in afs.
 
Last edited:
snerkler said:
My understanding is that PDAF uses distances and movement as well as a way to help with tracking motion, it uses complicated software to 'predict' where the next shot should be. It's not simply a case of bringing two images together (although this is part of the process). I thought that Sony finished off with contrast detect too but my recent reading suggests that it's purely PDAF on the latest bodies which is what allows it to be as fast as DSLR, contrast detect still can't track as fast as PDAF. In AF-S it may well still be finished off with contrast detect, I don't know.
Click to expand...

If the issue was only in AF-C I definitely agree there lots more than can go wrong, but this front focus issue is happening in AF-S with static objects.
 
Trifid said:
It does. Keep them coming people. Especially those getting the Sigma Art lenses next week. :)
Click to expand...

qXDh6uF.jpg


qUQcLFw.jpg


9U6gGrY.jpg
 
Looks like I should skip the tamron and go for the sony 24-105, especially if the tamron needs to be at f4 or 5.6 to sharpen up the edges
 
Raymond Lin said:
That’s the risk of 3red party lenses, and it’s not going to be fixed unless the lens’ firmware is updated. Sigma lenses are fine but unless Tamron makes a dock, you need to send it back.
Click to expand...

How do you know Sigma lenses are immune?
 
twist said:
Er but it is broke in video, have you looked at what it fixes o_O
Click to expand...

I apply that to everything. Not just camera.

As far as I’m concern, my camera aren’t broke so I’m not going to fix it.

What is broken in my A73?
 
Raymond Lin said:
That’s the risk of 3red party lenses, and it’s not going to be fixed unless the lens’ firmware is updated. Sigma lenses are fine but unless Tamron makes a dock, you need to send it back.
Click to expand...

It's not just 3rd party lenses having problems though. Assuming Tamron and Sigma do release e mount versions of their docks then at least you can fix it.

To quote from snapsy on the FM forums:

If the misfocus is unique to AF-S and does not occur in AF-C then that sounds like a focus shift issue, just like what was discovered on the initial copies of the Sony 24-105mm. For the Sony it wasn't optical focus shift but instead algorithmic focus shift, ie the lens dynamically adjusted focus based on the dialed-in aperture and some copies went out with their aperture-to-focus look-up table misconfigured.

Maybe the Sony 85mm f1.8 can be fixed this way as well.
 
jonneymendoza said:
Macs are pos I agree:)
Click to expand...

Lol that’s a wild statement, my mid-2010 Apple MacBook Pro is still running today, it’s outlasted any of my previous Windows based laptops by years!
Never had a issue updating my Sony bodies on the MacBook. :)
 
jj_glos said:
It's not just 3rd party lenses having problems though. Assuming Tamron and Sigma do release e mount versions of their docks then at least you can fix it.

To quote from snapsy on the FM forums:

If the misfocus is unique to AF-S and does not occur in AF-C then that sounds like a focus shift issue, just like what was discovered on the initial copies of the Sony 24-105mm. For the Sony it wasn't optical focus shift but instead algorithmic focus shift, ie the lens dynamically adjusted focus based on the dialed-in aperture and some copies went out with their aperture-to-focus look-up table misconfigured.

Maybe the Sony 85mm f1.8 can be fixed this way as well.
Click to expand...

I’m talking about that Tamron not focusing at all.

What are you taking about?
 
You must log in or register to reply here.
Back
Top