I would love to know that as well. Would I be the guinea pig if I sent my lens into Zeiss?
I do have an "extra" Loxia 25 which I will use as the canary in the coal mine, if I get some more solid guidance that Zeiss has a fix.
LBJ2 wrote:
I haven't traversed the majority of this thread yet. Just curious has any one else had a successful Loxia lens repair to correct this issue? Or are you the first to consider sending in one of your lenses for evaluation and repair ?
tsdevine wrote:
I would love to know that as well. Would I be the guinea pig if I sent my lens into Zeiss?
I do have an "extra" Loxia 25 which I will use as the canary in the coal mine, if I get some more solid guidance that Zeiss has a fix.
Given you have multi lenses with the same issue, I'd go for it and let them evaluate. At a minimum you might be the one that exposes a chronic issue across the line and prompts Zeiss to start working in a fix. If it turns out to be a firmware fix they can roll-out a firmware update.
FWIW, my one Loxia lens the 21, does not have this issue when used on my Sony A1 camera. Which makes me wonder if this is a HW failure to a certain batch of parts used.
My Loxia 21 isn't as bad as my "newer" Loxia....but it had to have been from the first batch as I bought it upon release. In fact there are probably posts here on FM where I compared it to my ZE 21.
My son is going in for some pretty major surgery here in a couple of weeks.....so it may be a little while until I pursue things further with Zeiss. I have over 2 years worth of warranty on all but my Loxia 21 (warranty has long passed), so I'm not in an incredible hurry. I was hoping to not be the guinea pig, but it may work out that way.
LBJ2 wrote:
Given you have multi lenses with the same issue, I'd go for it and let them evaluate. At a minimum you might be the one that exposes a chronic issue across the line and prompts Zeiss to start working in a fix. If it turns out to be a firmware fix they can roll-out a firmware update.
FWIW, my one Loxia lens the 21, does not have this issue when used on my Sony A1 camera. Which makes me wonder if this is a HW failure to a certain batch of parts used.
tsdevine wrote:
My Loxia 21 isn't as bad as my "newer" Loxia....but it had to have been from the first batch as I bought it upon release. In fact there are probably posts here on FM where I compared it to my ZE 21.
My son is going in for some pretty major surgery here in a couple of weeks.....so it may be a little while until I pursue things further with Zeiss. I have over 2 years worth of warranty on all but my Loxia 21 (warranty has long passed), so I'm not in an incredible hurry. I was hoping to not be the guinea pig, but it may work out that way.
1. Sony can fix via a camera firmware update, and is wiling to do it.
2. Zeiss can fix it (via hardware and/or software), and is willing to do it.
3. It can't be reasonably fixed by either Sony or Zeiss.
I don't want to bother sending my lenses to Zeiss if #2 isn't a path to resolving the problem.
GMPhotography wrote:
Is Zeiss even available to repair is the bigger question
I am still certainly it is a Sony E-mount protocol tolerance problem:
The aperture value is not a discrete value: 1.4, 2.0, 2.8 but rather either an analog signal or digital continuous value in the spec of the protocol.
In either case the focus assist zoom mode needs to make a decision when at what delta a change of aperture is making it snap out of focus assist.
At some point in the chain, the analog signal of the position of the aperture ring needs to be transformed into a digital value. Either in the lens (if E-mount protocol for aperture is digital) or in the camera (if analog).
The camera needs to have a ‚threshold‘ at what delta a ‚change‘ happened.
The problem is likely that there is slight jitter (mechanical and/or electric) in the signal before digitalization.
A threshold window too narrow will make any form of random high jitter be interpreted as ‚change‘ and it will snap out of magnification.
Up to a certain camera generation, Sony cameras had a signal path and or jitter tolerance that did not break this function.
Starting with A1, A7s3, A7r5 (maybe others) this changed so that seem to have changed so that now manual lenses like Loxia or Voigtlander ‚slip through‘ in terms of their level of tolerance in regard to the aperture signal value stream.
Some lenses are within tolerance to still work, but might slip later. Others will stay in tolerance forever.
The theory explains everything anyone ever posted on the issue.
And it could be easily changed.
BY SONY.
They could change the tolerance value for aperture value triggering for the magnifaction feature so that it quite a bit more tolerant. To the point that it would work again for everyone like it did with older bodies.
BY FIRMWARE UPDATE
Neither Zeiss nor Voigtlander should need to change anything. Tolerance and jitter slip has to be accommodated for in the Sony E-mount spec and Sony firmware.
IMHO obviously. I dont have insight knowledge but lots of experience in related fields.
Anyone supporting this? Would change who to put pressure on.
Bokehddicted wrote:
For me the problem is rather obvious.
The zoom function is designed to stay zoomed while you focus. And to get out of zoom when anything (or some defined things) change. For buttons it is easy. If you press some, it snaps out of it.
For things that have a more ‚analogue‘ readout, like the aperture on the Loxia. Even if the aperture is not in declicked mode, there seems to be a rather broad range of non-discrete values being communicates from the lens to the body.
The function that needs to make a call whether something ‚changed‘ needs to make ‚threshold‘ check, rather than a binary button check.
In my theory, they might have changed that threshold for those more non-discrete things, or only for the aprerture in particular.
That would explain many things. You might have a body or a firmware with wider or more narrow threshold.
And your particular lens, old or young, might be within the tolerance to trigger.
I pretty much agree what that assessment of the problem, it's along the lines of what I was thinking too.
My dilemma is that I find it hard to find a path to put pressure on Sony to fix, if Zeiss won't say anything as to whose problem it is to solve. If Zeiss said it was a Sony problem. Boom, I'd focus more on bugging Sony. Working with vendors, I would at least like Zeiss to point the finger at Sony, even if they won't work with Sony to resolve it. (Honestly I would expect Zeiss to work on behalf of their customers to pressure Sony to fix it...assuming it is Sony's problem to fix). If Zeiss did that, I think it would get fixed. If they do that and Sony doesn't fix it, I'm not sure the small niche of Loxia users are going to succeed in getting Sony to fix it. There are no warnings from Zeiss, that the lenses they are actively selling may not work with newer Sony cameras. That's kind of sucky.
I understand the argument that "Sony broke it". But Sony has shared the specifications with 3rd parties who have signed up for E mount. So it seems strange to me that they aren't following their own spec anymore. AND it only affects Zeiss Loxia (not Batis) lenses, as well as Voigtlander (who many people think manufacture the Loxia lenses for Zeiss.). And no other 3rd party manufacturer was affected. I suspect, like I think you do, that Zeiss (and Voigtlander) maybe got fudgy close to following the spec......but are close to the edge of being out of spec. And depending on individual copies, it can maybe fall into spec, or fall out of spec.... In other words Sony is following their own spec more closely with the new cameras, or the fudge factor/tolerance to the spec has changed, and that fudge factor was allowing the Loxia lenses to function properly on older cameras.
Do we have any examples where an OEM updated firmware to have their cameras work with 3rd party lenses that appear to have a compatibility issue, when that 3rd party doesn't seem to give clear acknowledgement as to whose problem it is to fix? I do have correspondence from Zeiss (via their Facebook account) that they are aware of the problem, and that the lens can be sent in for a fix. Granted in Europe....but you get the idea. So at this point it's hard to knock me off bugging Zeiss.
Bokehddicted wrote:
I posted the theory below one year ago.
I am still certainly it is a Sony E-mount protocol tolerance problem:
The aperture value is not a discrete value: 1.4, 2.0, 2.8 but rather either an analog signal or digital continuous value in the spec of the protocol.
In either case the focus assist zoom mode needs to make a decision when at what delta a change of aperture is making it snap out of focus assist.
At some point in the chain, the analog signal of the position of the aperture ring needs to be transformed into a digital value. Either in the lens (if E-mount protocol for aperture is digital) or in the camera (if analog).
The camera needs to have a ‚threshold‘ at what delta a ‚change‘ happened.
The problem is likely that there is slight jitter (mechanical and/or electric) in the signal before digitalization.
A threshold window too narrow will make any form of random high jitter be interpreted as ‚change‘ and it will snap out of magnification.
Up to a certain camera generation, Sony cameras had a signal path and or jitter tolerance that did not break this function.
Starting with A1, A7s3, A7r5 (maybe others) this changed so that seem to have changed so that now manual lenses like Loxia or Voigtlander ‚slip through‘ in terms of their level of tolerance in regard to the aperture signal value stream.
Some lenses are within tolerance to still work, but might slip later. Others will stay in tolerance forever.
The theory explains everything anyone ever posted on the issue.
And it could be easily changed.
BY SONY.
They could change the tolerance value for aperture value triggering for the magnifaction feature so that it quite a bit more tolerant. To the point that it would work again for everyone like it did with older bodies.
BY FIRMWARE UPDATE
Neither Zeiss nor Voigtlander should need to change anything. Tolerance and jitter slip has to be accommodated for in the Sony E-mount spec and Sony firmware.
IMHO obviously. I dont have insight knowledge but lots of experience in related fields.
Anyone supporting this? Would change who to put pressure on.
"I messaged Zeiss about this issue and received a response:
Dear David,
The issue is addressed and our developer team and supplier are currently working on a solution. We will start the testings when lenses with new firmware implemented are available.
Our website is always the best source for latest information because now it is hard to say when a solution will be ready for the market.
Thanks again for your request.
Mit freundlichen Grüßen - With best regards"
Bokehddicted wrote:
I posted the theory below one year ago.
I am still certainly it is a Sony E-mount protocol tolerance problem:
The aperture value is not a discrete value: 1.4, 2.0, 2.8 but rather either an analog signal or digital continuous value in the spec of the protocol.
In either case the focus assist zoom mode needs to make a decision when at what delta a change of aperture is making it snap out of focus assist.
At some point in the chain, the analog signal of the position of the aperture ring needs to be transformed into a digital value. Either in the lens (if E-mount protocol for aperture is digital) or in the camera (if analog).
The camera needs to have a ‚threshold‘ at what delta a ‚change‘ happened.
The problem is likely that there is slight jitter (mechanical and/or electric) in the signal before digitalization.
A threshold window too narrow will make any form of random high jitter be interpreted as ‚change‘ and it will snap out of magnification.
Up to a certain camera generation, Sony cameras had a signal path and or jitter tolerance that did not break this function.
Starting with A1, A7s3, A7r5 (maybe others) this changed so that seem to have changed so that now manual lenses like Loxia or Voigtlander ‚slip through‘ in terms of their level of tolerance in regard to the aperture signal value stream.
Some lenses are within tolerance to still work, but might slip later. Others will stay in tolerance forever.
The theory explains everything anyone ever posted on the issue.
And it could be easily changed.
BY SONY.
They could change the tolerance value for aperture value triggering for the magnifaction feature so that it quite a bit more tolerant. To the point that it would work again for everyone like it did with older bodies.
BY FIRMWARE UPDATE
Neither Zeiss nor Voigtlander should need to change anything. Tolerance and jitter slip has to be accommodated for in the Sony E-mount spec and Sony firmware.
IMHO obviously. I dont have insight knowledge but lots of experience in related fields.
Anyone supporting this? Would change who to put pressure on.
The one thing I am not certain on, is whether Sony actually really changed their protocol and tolerance in the way of a deliberate change in firmware code.
I am sceptical that Sony would actually change such a tolerance value, leading to numerous complaints from people using the focus ring magnified view feature for manual lenses. It does not make sense, and there is no real conceivable reason for de facto breaking a working protocol.
I wonder if the faster processors, most noticeably the A1 having a 10x faster processor, have not exposed a lingering and potential issue, that would not have appeared if it weren't for the vastly upgraded responsiveness of the later Sony cameras.
This would make coming to a solution complicated: Sony can lay the issue with Zeiss, seeing their lenses as being out of spec. Zeiss will not need to point elsewhere, as they have downgraded their support for the Loxia lenses to virtually non existent.
I am pretty certain they didnt change the spec of the protocol. Its likely an non-deliberate implementation problem, maybe due to rewriting some part of the software stack on a new processor.
Faster chip, different sampling rate, slightly different behavior. More bit depth to values, different rounding behavior. Something like that.
Maybe still in spec, but excluding some lenses on edge cases.
Just rumbling out loud.
ChrisMak wrote:
The one thing I am not certain on, is whether Sony actually really changed their protocol and tolerance as a deliberate change in firmware code.
I am sceptical that Sony would actually change such a tolerance value, leading up to numerous complaints from people using the auto magnification feature for manual lenses. It does not make sense, and there is no real conceivable reason for de facto breaking a working protocol.
I wonder if the faster processors, most noticeably the A1 having a 10x faster processor, have not exposed a lingering and potential issue, that would not have appeared if it weren't for the vastly upgraded responsiveness of the later Sony cameras. ...Show more →
But didn't we hear that Zeiss wasn't leaving the photography market...they stand behind their products...blah blah blah?
At some point you shouldn't knowingly be selling lenses that aren't fully compatible with Sony cameras.....at least without warning customers.
I wish Zeiss was more active in photography, specifically E mount. So I'm not trying to bash them. But at some point their name either means something, or it doesn't.
So I am hopeful to at least get an answer from them, even if I don't like it.
ChrisMak wrote:
The one thing I am not certain on, is whether Sony actually really changed their protocol and tolerance in the way of a deliberate change in firmware code.
I am sceptical that Sony would actually change such a tolerance value, leading to numerous complaints from people using the focus ring magnified view feature for manual lenses. It does not make sense, and there is no real conceivable reason for de facto breaking a working protocol.
I wonder if the faster processors, most noticeably the A1 having a 10x faster processor, have not exposed a lingering and potential issue, that would not have appeared if it weren't for the vastly upgraded responsiveness of the later Sony cameras.
This would make coming to a solution complicated: Sony can lay the issue with Zeiss, seeing their lenses as being out of spec. Zeiss will not need to point elsewhere, as they have downgraded their support for the Loxia lenses to virtually non existent. ...Show more →
ChrisMak wrote:
The one thing I am not certain on, is whether Sony actually really changed their protocol and tolerance as a deliberate change in firmware code.
I am sceptical that Sony would actually change such a tolerance value, leading up to numerous complaints from people using the auto magnification feature for manual lenses. It does not make sense, and there is no real conceivable reason for de facto breaking a working protocol.
I wonder if the faster processors, most noticeably the A1 having a 10x faster processor, have not exposed a lingering and potential issue, that would not have appeared if it weren't for the vastly upgraded responsiveness of the later Sony cameras.
This would make coming to a solution complicated: Sony can lay the issue with Zeiss, seeing their lenses as being out of spec. Zeiss will not need to point elsewhere, as they have downgraded their support for the Loxia lenses to virtually non existent. ...Show more →
Yeah, I think you're on the right track. The protocol might have stayed the same, but the data (via lens contacts like EXIF) is simply polled at a higher frequency on current gen cameras. So, let's say prev gen polled 1 time per sec, the new gen bumped it up to 10 / sec.This means that any kind of signal triggered by unstable/faulty/twitchy/oversensitive element in the aperture ring mechanism is now 10x more likely to be registered by the camera... Given that nudging the aperture ring helps in most cases, I think it might be a mechanical issue with the ring mechanism itself or more of an electronic issue with the aperture ring movement detector/converter/encoder.
I'm probably an outlier, but I had this actually happen first on my A7III with CV35/1.4, and then reappear on A1 with CV65/2, which never showed the issue before on A7III...
I did some searching for info on the E-mount protocol. There are no analog signal lines, it is all done with serial communication with the lens. Manual chipped lenses regularly report aperture and focus position to the camera. It would appear if the camera has concluded the aperture changed it is because the lens digitally reported the aperture changed. I don't see why this would manifest on the newer cameras unless the newer cameras talk more often with the lens and it catches a change of state that would otherwise go unnoticed, but It would seem the frequency would just be less on older cameras not no problem at all.
I would think Sony could fix this by adding hysteresis to how much aperture change is required to cancel MF magnification. I would think this would be best as a configurable option as it would reduce correct functionality for well behaved lenses.
To me it makes more sense to fix this via a firmware update to the lens. The logic would be a bit tricky as the lens would need to track past aperture position and decide if a significant change actually happened to allow reporting a new value. This would essentially be asking the lens to not actually report the current measured aperture value in many cases. I don't see a big downside to an incorrect reported aperture as it is just for exif info and not used by the camera except to toggle out of MF magnification.
So assuming Zeiss or CV comes up with new firmware, how does that get loaded into the lens? I assume there are no USB ports on these lenses. Sony lenses can update firmware via the body, but I suspect this is not supported in these lenses. So maybe the lens needs to be disassembled to upgrade, or maybe it can be done via a dock for the lens. Anyone know about firmware upgrades for zeiss or CV?
I did some searching for info on the E-mount protocol. There are no analog signal lines, it is all done with serial communication with the lens. Manual chipped lenses regularly report aperture and focus position to the camera. It would appear if the camera has concluded the aperture changed it is because the lens digitally reported the aperture changed. I don't see why this would manifest on the newer cameras unless the newer cameras talk more often with the lens and it catches a change of state that would otherwise go unnoticed, but It would seem the frequency would just be less on older cameras not no problem at all.
I would think Sony could fix this by adding hysteresis to how much aperture change is required to cancel MF magnification. I would think this would be best as a configurable option as it would reduce correct functionality for well behaved lenses.
To me it makes more sense to fix this via a firmware update to the lens. The logic would be a bit tricky as the lens would need to track past aperture position and decide if a significant change actually happened to allow reporting a new value. This would essentially be asking the lens to not actually report the current measured aperture value in many cases. I don't see a big downside to an incorrect reported aperture as it is just for exif info and not used by the camera except to toggle out of MF magnification.
So assuming Zeiss or CV comes up with new firmware, how doe that get loaded into the lens? I assume there are no USB ports on these lenses. Sony can update firmware via the body, but I suspect this is not supported in these lenses. So maybe the lens needs to be disassembled to upgrade, or maybe it can be done via a dock for the lens. Anyone know about firmware upgrades for zeiss or CV?...Show more →
Very informative. For what it's worth, Zeiss Batis lenses can receive firmware updates via a Sony body, much the same as for Sony lenses. Zeiss has issued a firmware update for the Batis 2/40 CF and I installed this on my sample, worked flawlessly. As far as I know, Zeiss never issued firmware updates for Loxia lenses.
Why do you come to the conclusion after your findings that it should be fixed in the lens? It just reports its aperture value.
Its the function in the camera that gets triggered by various inputs that triggers magnification to go off. The logic and threshold should be a function of the implementation of the feature. Not snap our at all. Snap out only after changing 1/3rd of a stop… i personally would not only like to see this fixed, but also to be able to switch this off altogether. Only to snap out by pressing the focus zoom button again.
That's my fear, quite a few 3rd party lenses use the Sony mechanism via the camera, including the Zeiss Batis series. (Which Tamron does as well, so I can't help tying Zeiss Batis to Tamron....but for this conversation, it's meaningless). I'm not aware of any end user upgradable firmware capability for Zeiss Loxia or CV (again...can't help but draw the parallel to the rumor that Cosina makes Loxia lenses for Zeiss.). I would send my lenses in for a firmware update, if that was the only option.
I can see a timing issue if the camera is polling the lens. Maybe it overwhelms the lens and it has trouble responding in the allotted cycle...and the camera reacts poorly to that. Or it may create a jitter scenario, which could explain how it can be really cranky at one point and fairly well behaved at others. If the lens is responsible to report a change, only when a change occurs (not polled by the camera)...then it's harder for me to understand how this is a problem exists only when using newer Sony cameras.
One thing that is interesting, I see the aperture as red I via the EVF when the lens is acting whacko. For example, I don't believe I can release the shutter in timer mode if the lens is acting up.....doesn't matter if I'm in magnified view. Sometimes the aperture shows red and the shutter button works in timer mode, sometimes it doesn't. But it doesn't 100% not work or 100% work. When it's not red it always works.
And sometimes the lens is super cranky and it's hard to find any aperture where you can use magnify. Sometimes the lens is pretty well behaved. Sometimes when it's cranky, if I change from one of my saved modes to another and back again...it gets less cranky. Sometimes turning off the camera and turning it back on makes it seem less cranky. Sometimes it lets me go into magnified view, but as soon as I turn the focus ring it bounces out. Sometimes it bounces out immediately. Sometimes it gives me a second or two, then it bounces out.
I swear I didn't feed it after midnight.
tschopp wrote:
I did some searching for info on the E-mount protocol. There are no analog signal lines, it is all done with serial communication with the lens. Manual chipped lenses regularly report aperture and focus position to the camera. It would appear if the camera has concluded the aperture changed it is because the lens digitally reported the aperture changed. I don't see why this would manifest on the newer cameras unless the newer cameras talk more often with the lens and it catches a change of state that would otherwise go unnoticed, but It would seem the frequency would just be less on older cameras not no problem at all.
I would think Sony could fix this by adding hysteresis to how much aperture change is required to cancel MF magnification. I would think this would be best as a configurable option as it would reduce correct functionality for well behaved lenses.
To me it makes more sense to fix this via a firmware update to the lens. The logic would be a bit tricky as the lens would need to track past aperture position and decide if a significant change actually happened to allow reporting a new value. This would essentially be asking the lens to not actually report the current measured aperture value in many cases. I don't see a big downside to an incorrect reported aperture as it is just for exif info and not used by the camera except to toggle out of MF magnification.
So assuming Zeiss or CV comes up with new firmware, how doe that get loaded into the lens? I assume there are no USB ports on these lenses. Sony can update firmware via the body, but I suspect this is not supported in these lenses. So maybe the lens needs to be disassembled to upgrade, or maybe it can be done via a dock for the lens. Anyone know about firmware upgrades for zeiss or CV?...Show more →
