Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple and their surprises.. Call it strange, but i like what Apple do to make it easy (but only on the surface), but once you realize something is a-miss, your gone...

Kinda... interesting.. The whole "covering it up till you get into something that was totally unexpected" seems just wrong, but weird at the same time, Apple would even consider a camera..

Apple's game of going around all bits of a phone to decide "what non-reparable thing can we do next year" part is amusing.

Just to throw us all off. Even ifix-it.
 
How is it more speculative for me to say that your suggestions are baseless.
Because I linked you to a basis showing your speculation was wrong. You said it was baseless speculation that Apple does a "bespoke" per unit calibration after assembly. I linked you to the machine Apple uses to do a per unit calibration after assembly.

What I was taking issue with is your criticism of other people for having "opinions" while considering your own an "examination".
I've no problem with people having opinions. If you read back to the comment you started laying into me about, my issue was with stating opinions as fact and treating people with other ideas as tools.

As you point out everything I'm saying is wrapped up in "maybe", "might" and "I don't know". I'm not claiming anything I say is fact beyond an understanding of how the technology works and knowing that since the introduction of ARKit, Apple has individually calibrated each unit. At no point did I say "you have an opinion, I have facts", what I said is "you are stating opinions as fact and attacking others as 'pro Apple' for considering other possibilities".

What assertion is that? I never made any claims about Apple's motivation here, whether for profit or anything else.
The person I began by responding to and who you stepped in to defend did. You went on to call profit "a more simple theory", which also qualifies as a claim. I provided the basis for my suggestion by linking to photos of Apple's per unit calibration machine. Now, provide a basis for profit as a more simple theory.
You should leave the biology out of it too.
You were the one that asked me how I thought freaking God would do it! If you credit God with creating the human vision system, then God does exactly what I'm describing. Just because you don't like that the answer to your question undermines your narrative, don't tell me I shouldn't have answered it.

Another is that any part that still has some undesired variance would be considered a defect, and is more likely to end up in the bin than nurtured calibrated to meet its full potential.
Except, of course, I just gave you a link to a machine that calibrates out the variance. Whether it is the more likely or less likely approach in your opinion, it's how it's been done since at least iPhone 8 so it is now more than just likely-- it's true.

Your understanding of "manufacturing", "calibration", and even "computing" doesn't convince me of much either.
As someone who has actually worked in all three areas, I appreciate that you think everything that happens is near magic levels of sophistication barely held together by wishes and spiderwebs, but I can assure you that is not the case.
Apple knows how to make cameras and processors, without tripping over themselves to get them actually working once put together.
And yet, with all your supposed expertise, you're simply wrong as I've now linked twice. Apple does a post assembly calibration plain and simple. They've said it themselves, and we've seen the machine. And you've somehow stepped in to defend someone who thinks you can build a good depth map and perform multisensor integration without good extrinsics...

As for the profit motive, it's right there in the article you shared:
"Imagine if every car we bought, every time you wanted to change the oil you had to take it to the dealer," he added. "It's just not realistic, and it monopolizes aspects of their business that are rarely good for their consumers."
Not in the article I linked to, but in a Vice article about the same machine. So you did look at the article, and just chose to ignore everything about Apple doing a unit by unit calibration...

Oil changes are required maintenance, not repairs. I change the oil in every car at least twice a year. I've never had to swap a camera. So the question of profitability comes down to how many cameras are currently replaced each year specifically by unauthorized 3rd party repair shops, how much does the repair cost the customer, and how much does the repair cost Apple. If those numbers are sufficiently large to impact Apple's share price, then the "it's all about profit" argument might have some legs.

I've never said exerting more control over the camera repair isn't about security or profit, I've simply pointed out that it could be about other things as well. I state that explicitly in just about every post I've made in this thread. For some reason you think that's a character flaw...
 
Last edited:
And yet, with all your supposed expertise, you're simply wrong as I've now linked twice. Apple does a post assembly calibration plain and simple. They've said it themselves, and we've seen the machine. And you've somehow stepped in to defend someone who thinks you can build a good depth map and perform multisensor integration without good extrinsics...

No, I'm saying that calibration doesn't do what you seem to think it does, and the mere presence of calibration doesn't explain locking down camera hardware as neatly as you seem to think. It's a very common technical procedure that comes in different forms, but essentially amounts to checking that measurement systems work to a prescribed level.
It's not really a part of putting things together, but a check against some specification. Simply knowing that they do it doesn't tell you what that specification is or why it's necessary. It also doesn't indicate defective performance if it is not performed.

Basically, Apple can have very high internal standards for new and repaired products, and use calibration as one part of guaranteeing those standards for themselves and customers.
But regardless, if at some point you want to repair your busted iPhone, you should have the right to a crappy repair if you want. To your own standards. Barring any actual danger or security risk etc, which is something Apple would have to explain.

You were the one that asked me how I thought freaking God would do it! All I was doing was giving you an answer to your question. If you credit God with creating the human vision system, then God does exactly what I'm describing. Just because you don't like that the answer to your question undermines your narrative, don't tell me I shouldn't have answered it.

It was a reference to the omnipotence paradox. The joke being that your expectations of Apple are quite high, and might be lowered if you had a better understanding of manufacturing, or any type of engineering.
 
No, I'm saying that calibration doesn't do what you seem to think it does, and the mere presence of calibration doesn't explain locking down camera hardware as neatly as you seem to think. It's a very common technical procedure that comes in different forms, but essentially amounts to checking that measurement systems work to a prescribed level.
It's not really a part of putting things together, but a check against some specification. Simply knowing that they do it doesn't tell you what that specification is or why it's necessary. It also doesn't indicate defective performance if it is not performed.

Basically, Apple can have very high internal standards for new and repaired products, and use calibration as one part of guaranteeing those standards for themselves and customers.
But regardless, if at some point you want to repair your busted iPhone, you should have the right to a crappy repair if you want. To your own standards. Barring any actual danger or security risk etc, which is something Apple would have to explain.
What you're describing sounds more like validation than calibration. I'm not sure what systems you're most familiar with calibrating, but these types of multi-view vision systems can be extremely sensitive to small variations. A pixel is a bit over a micron, sits on a moving image stabilization platform and behind I think 7 optical elements now, some of which move axially for autofocus? You're trying to localize image features to a fraction of a pixel if you can. The ability to do so, drives the performance of your depth mapping and SLAM solution.

And you're not only worried about lateral variations in spacing between the sensors, but the 3D translational offsets and 3D of rotation, plus the sensor to lens offsets and alignments, and the inevitable distortion of the lens itself. All of which will have unit to unit variation.

It is essentially impossible, or at the very least cost prohibitive, to meet that performance through controlling manufacturing tolerances alone. Fortunately almost all of those errors, if well modeled and measured accurately, can be corrected for in software. So the calibration procedure measures all of those errors and records them for software to apply later. Calibration adds manufacturing steps, which adds cost, but is relaxes the specified manufacturing tolerances and improves yield which reduces cost. And it allows the system to perform at a level that you'd never achieve if you were limited by manufacturing precisions.

Calibration in this context is not a pass/fail. It's the collection of information critical to the functioning of the product.

But regardless, if at some point you want to repair your busted iPhone, you should have the right to a crappy repair if you want. To your own standards. Barring any actual danger or security risk etc, which is something Apple would have to explain.
This is a different argument. On the one hand, I see your point-- if I want a crap repair for my device, so be it. On the other, I understand why Apple may not feel the same. It's clear that most people don't understand how these systems work, and you can read through these forums and find it full of people blaming Apple for things that are outside Apple's control. As I said earlier, if I were Apple, I'd rather be blamed for something I control than for something I don't.

I'd also have to think through what the impact is on Apple if you could do an unauthorized repair that Apple can't prove.

It was a reference to the omnipotence paradox. The joke being that your expectations of Apple are quite high, and might be lowered if you had a better understanding of manufacturing, or any type of engineering.
I love it when people think they know something about anonymous strangers...

Anyway. If an omnipotent being could make a device with zero manufacturing tolerance, then no calibration would be necessary. For us mortals, we need to put in the work. This isn't about expectations of Apple-- this is standard practice.
 
People compairing phones and laptops to cars: stop it. The car industry has a completely different approach - te sell you the car cheaper, but spare parts are expensive. People keep saying they can work on their cars but do they really? Oil, batteries and brake pads are just consumable parts, they do need to be replaced often and yes, you need to go to the dealer to maintain warranty for a lot of repairs. But do you swap electronics from your cars? do you replace engine components? do you work on your servo motor? do you swap transmission systems? don't think so.

The whole right to repair movement doesn't really make sense in the way we consume products nowadays and implementing that would change everything - for better or for worse. Do we undersand the impact? would it really be better for the consumer? I love Hugh Jeffrey's videos and repairs, but how practical is it though? It's nice that you can upgrade an old iPod, but would you use today? it's big, old software and a much worse experience than an iPhone. You CAN use it of course, but would you really?
 
Your description of "computation" above doesn't do much to convince me
You don't pay me for teaching or even convincing you. Use some DIY solution instead

I've shown that Apple does a per-unit calibration
Ohh, really? You've only shown you think it is such, in the same time being unable to reply simplest question why same tech in former models had no such a calibrations.

Stop murmuring this b#$#****, no one cares
 
Is it worth adding the cost of the components to every device that ships just to make it easier for unauthorized 3rd party repair of the much smaller number that need the camera replaced when when you already have plenty of space to store that data right where it's being used on the processor side of the connector?

Every component added is another potential point of failure, so is it worth increasing the overall failure rate just to make it easier for 3rd party repair?

Space is tight inside the phone, and every cubic mm is fought for. What features should be sacrificed to add the storage making it easier for 3rd party repair of the small number of phones that require camera repair?

I don't think the accelerometers and gyros are in the camera module, so even if the camera module itself were factory calibrated, there's another set of extrinsics that may still need to be done for motion tracking.

And then, as the module is installed (which has tolerances to account for) and ages, the calibrations can be expected to shift. So a precalibration may not be sufficient, and may require special equipment that certified techs are known to have but other 3rd parties may not so the software check-in may just be a way of ensuring that the calibration is done at all.
Yep, I agree, these are the questions I was asking as well. Is it worth it? Unlike the assumption that seems to load your questions, I think it might be well worth it. It might be worth adding a tiny bit more complexity to make the phone more repairable, even by third parties. Since this device is expected to last beyond the 1-2 years warranty, it shouldn't matter how "authorized" is the repair. These things are those that make "authorized" matter, and it shouldn't matter as much. That's part of the problem.

Cameras already include an integrated circuit and logic, which I imagine is designed by Apple and it's specific to the model. Flash memory could be part of the IC. Why are we assuming that it wasn't done because it was the best technical choice? Product decisions are complex, and not all of them are made to maximize the benefit to the customer. It could be that when the camera was designed the idea wasn't there, and then they had no time to iterate on the design to make it better. Or it could be that the idea was there, but implementing the logic and the software was deemed an unnecessary cost.

Or—hear me out—the idea was there and it was implemented. Possibly it's been there since iPhone 4 or dunno. Whatever hardware-specific calibration data had to be stored is wired in the camera itself. And this has nothing to do with calibration! This sounds just as likely as the above hypotheses, if not more. Perhaps we don't have enough information to establish that. Then, maybe, the manufacturer could be a tiny bit more transparent about this. This is not an unreasonable demand.
 
What you're describing sounds more like validation than calibration. I'm not sure what systems you're most familiar with calibrating, but these types of multi-view vision systems can be extremely sensitive to small variations. A pixel is a bit over a micron, sits on a moving image stabilization platform and behind I think 7 optical elements now, some of which move axially for autofocus? You're trying to localize image features to a fraction of a pixel if you can. The ability to do so, drives the performance of your depth mapping and SLAM solution.

And you're not only worried about lateral variations in spacing between the sensors, but the 3D translational offsets and 3D of rotation, plus the sensor to lens offsets and alignments, and the inevitable distortion of the lens itself. All of which will have unit to unit variation.

It is essentially impossible, or at the very least cost prohibitive, to meet that performance through controlling manufacturing tolerances alone. Fortunately almost all of those errors, if well modeled and measured accurately, can be corrected for in software. So the calibration procedure measures all of those errors and records them for software to apply later. Calibration adds manufacturing steps, which adds cost, but is relaxes the specified manufacturing tolerances and improves yield which reduces cost. And it allows the system to perform at a level that you'd never achieve if you were limited by manufacturing precisions.

Calibration in this context is not a pass/fail. It's the collection of information critical to the functioning of the product.


This is a different argument. On the one hand, I see your point-- if I want a crap repair for my device, so be it. On the other, I understand why Apple may not feel the same. It's clear that most people don't understand how these systems work, and you can read through these forums and find it full of people blaming Apple for things that are outside Apple's control. As I said earlier, if I were Apple, I'd rather be blamed for something I control than for something I don't.

I'd also have to think through what the impact is on Apple if you could do an unauthorized repair that Apple can't prove.


I love it when people think they know something about anonymous strangers...

Anyway. If an omnipotent being could make a device with zero manufacturing tolerance, then no calibration would be necessary. For us mortals, we need to put in the work. This isn't about expectations of Apple-- this is standard practice.

I wasn't suggesting that calibration is pass/fail, just keeping it simple. Because everything you said here still speaks to high quality, not necessary quality.
And still goes back to the earlier thing, that you aren't actually adding legitimacy to speculation simply by adding more detail to the speculation.

The real difference between profit and security explanations, and a quality explanation, is that the first two make it very clear why third party repair should be prevented and how Apple or customers would benefit.
Quality as a motivation is more vague, and you aren't addressing why Apple should suddenly start enforcing this by guessing how difficult it could be to ensure. Which means a lot of your argument boils down to "look how difficult this thing I made up sounds, no way anyone else could do it" which is giving Apple a lot of credit and third parties the opposite.
 
And how much BS is this? All of it. All the BS. To the folks wondering why this is a big deal, or defending Apple, imagine if Honda announced today that going forward all 2020 or later vehicles they sold could only have the batteries, bulbs and tires replaced at “authorized” repair facilities. If the vehicles computer detected a replacement headlight that wasn’t “calibrated” with proprietary software, the car would lock out the ignition entirely.
Physically the new part is fine, even genuine from the same brand your vehicle came with stock. Functionally the part works identically to that which it replaced, and the process of replacing it was 99% the same as with prior models.
But now, thanks to Honda, unless you pay three times as much to have the repair/replacement done by a “certified” facility than if you did the work yourself you no longer have the right to drive the vehicle you paid for and legally own.

That is what Apple is doing. That is why Right to Repair laws that prevent just this exact type of nonsense MATTER.
Many of you will hand gesture this away because you don’t change your own oil anyway, but the fact you prefer to pay extra for the convenience of having someone else do it for you does not justify preventing another owner from doing it themselves, either because it is cheaper or just because they enjoy being self sufficient and take pride in such things. Your case use doesn’t have to be anyone else’s, and vice versa. Try to understand that when people like me get frustrated with moves like this from companies we have been loyal to and given thousands of dollars to. This is about us, not iFixit or their business model, anymore than it would be about AutoZone or Pep Boys in the above analogy.
Things being more difficult to repair due to increased complexity is understandable, and we don’t care about that. I guarantee my 1985 F-150 was a thousand times easier to work on than my 2011. But when Apple INTENTIONALLY makes their product difficult to repair by glueing/soldering things in, or requiring proprietary silicon or software to “authorize” a functionally sound repair, that is absolute money grubbing profit margin pumping anti-consumer monopolistic BS!
This. And it can be applied to anything. Right to repair is the right to repair. can you imagine if your Ac at your house went out and you could only use the blower produced by the Manufacture and installed by an authorized repair center or the ac would quit working. A 250.00 self repair just hit 1k. And I know this first hand because this is a repair I did myself in the last year.
 
You don't pay me for teaching or even convincing you. Use some DIY solution instead


Ohh, really? You've only shown you think it is such, in the same time being unable to reply simplest question why same tech in former models had no such a calibrations.

Stop murmuring this b#$#****, no one cares
If you look above, I've linked to both the keynote where Schiller says every unit is calibrated at the factory, and to the machine Apple uses to calibrate each individual unit after repair. If you have a citeable reference that shows Apple doesn't calibrate per unit, then please quote it and then maybe there's a reason to keep discussing it. Otherwise, I take that much as fact.
 
  • Like
Reactions: skari
Imagine your brand new $70k Tesla has its tires remotely locked by Elon Musk, and swapping the tires is going to cost you an arm and a leg.
if you need to replace a display or a a camera in a Tesla yes you need to contact a Tesla authorised shop. You really don't get it.
 
And how much BS is this? All of it. All the BS. To the folks wondering why this is a big deal, or defending Apple, imagine if Honda announced today that going forward all 2020 or later vehicles they sold could only have the batteries, bulbs and tires replaced at “authorized” repair facilities. If the vehicles computer detected a replacement headlight that wasn’t “calibrated” with proprietary software, the car would lock out the ignition entirely.
Physically the new part is fine, even genuine from the same brand your vehicle came with stock. Functionally the part works identically to that which it replaced, and the process of replacing it was 99% the same as with prior models.
But now, thanks to Honda, unless you pay three times as much to have the repair/replacement done by a “certified” facility than if you did the work yourself you no longer have the right to drive the vehicle you paid for and legally own.

That is what Apple is doing. That is why Right to Repair laws that prevent just this exact type of nonsense MATTER.
Many of you will hand gesture this away because you don’t change your own oil anyway, but the fact you prefer to pay extra for the convenience of having someone else do it for you does not justify preventing another owner from doing it themselves, either because it is cheaper or just because they enjoy being self sufficient and take pride in such things. Your case use doesn’t have to be anyone else’s, and vice versa. Try to understand that when people like me get frustrated with moves like this from companies we have been loyal to and given thousands of dollars to. This is about us, not iFixit or their business model, anymore than it would be about AutoZone or Pep Boys in the above analogy.
Things being more difficult to repair due to increased complexity is understandable, and we don’t care about that. I guarantee my 1985 F-150 was a thousand times easier to work on than my 2011. But when Apple INTENTIONALLY makes their product difficult to repair by glueing/soldering things in, or requiring proprietary silicon or software to “authorize” a functionally sound repair, that is absolute money grubbing profit margin pumping anti-consumer monopolistic BS!
you don't know muds a bout cars do you. this is the same on modern cars all components need to be registered in to the computer to work, and naming tiers and lightbulbs is not comparable to computer sensors and moving components from car to car is not possible wit out authorisation.
 
Imagine your brand new $70k Tesla has its tires remotely locked by Elon Musk, and swapping the tires is going to cost you an arm and a leg.
Tesla is known for locking down features of repaired cars in non authorized service locations. There has been a lot of complaints about people loosing the fast charging on their cars due to this. Then there are those so called salvaged cars that Tesla won't even touch because someone else repaired it.
"...but iFixit notes that the camera module is not a security component."

I assume that iFixit did not get the point here.
Part of the system is LIDAR, and this will certainly become a security relevant tool soon.

Everyone who has a deeper insight into the matter knows this.

Back to this topic, it is quite weird how by the test from iFixit, the cameras on the Iphone 12 aren't swappable, but the camera on the 12 pro is. A bit strange here.
 
I think is a little short sighted defending a company that will prevent to do what ever you want with your already paid phone, cause in the long run it just seem that you are just leasing a very expensive phone. Is a monopolizing practice, that only the selling company can repair your phone. It doesn't matter if the repair is done in a kiosk at the mall, a 3rd party technician established store or even yourself, but Apple every time is taking the choice away from the customer.
 
Tesla is known for locking down features of repaired cars in non authorized service locations. There has been a lot of complaints about people loosing the fast charging on their cars due to this. Then there are those so called salvaged cars that Tesla won't even touch because someone else repaired it.


Back to this topic, it is quite weird how by the test from iFixit, the cameras on the Iphone 12 aren't swappable, but the camera on the 12 pro is. A bit strange here.
maybe is just a bug
 
Yep, I agree, these are the questions I was asking as well. Is it worth it? Unlike the assumption that seems to load your questions, I think it might be well worth it. It might be worth adding a tiny bit more complexity to make the phone more repairable, even by third parties. Since this device is expected to last beyond the 1-2 years warranty, it shouldn't matter how "authorized" is the repair. These things are those that make "authorized" matter, and it shouldn't matter as much. That's part of the problem.

Cameras already include an integrated circuit and logic, which I imagine is designed by Apple and it's specific to the model. Flash memory could be part of the IC. Why are we assuming that it wasn't done because it was the best technical choice? Product decisions are complex, and not all of them are made to maximize the benefit to the customer. It could be that when the camera was designed the idea wasn't there, and then they had no time to iterate on the design to make it better. Or it could be that the idea was there, but implementing the logic and the software was deemed an unnecessary cost.

Or—hear me out—the idea was there and it was implemented. Possibly it's been there since iPhone 4 or dunno. Whatever hardware-specific calibration data had to be stored is wired in the camera itself. And this has nothing to do with calibration! This sounds just as likely as the above hypotheses, if not more. Perhaps we don't have enough information to establish that. Then, maybe, the manufacturer could be a tiny bit more transparent about this. This is not an unreasonable demand.
Yep, certainly all possible. I'm not sure I want to carry the cost of making phones more repairable, because I've never once need to repair a phone, but there is a tradeoff to be made there.

I don't think Apple designs their own sensors. That may have changed in recent years, but earlier phones were using sensors by companies like Omnivision, On, and Sony I think. I haven't seen the sensor manufacturer called out in the tear downs of the last few generations. That said, I'm sure Apple carries the weight to specify and shape the design.

It may cost a bit to embed flash memory into an image sensor process, assuming anyone was willing to make such a thing. TSMC, for example, has an eFlash process but I'm not sure if it's compatible with their image sensor process.

I think there's two separate conversations here. One is about what may have motivated Apple to start restricting components, and the other is whether we're ok with them doing that. Maybe the answer to the second depends on people's assumptions about the first. You're suggesting you'd be willing to pay more for the phone if you could get the camera swapped in a back alley. That's not my personal priority, but I think there are a lot of people who think that way, just from a "hey it's mine to do with as I wish" perspective.
 
everything you said here still speaks to high quality, not necessary quality.
That seems reasonable when we’re discussing Apple’s flagship phones, doesn’t it? If your point is that Apple should make lower quality products, or that people want lower quality results after repair, then that’s a different discussion.
And still goes back to the earlier thing, that you aren't actually adding legitimacy to speculation simply by adding more detail to the speculation.
If the detail I’m adding is specifically refuting your counter argument, it does.
The real difference between profit and security explanations, and a quality explanation, is that the first two make it very clear why third party repair should be prevented and how Apple or customers would benefit.
Quality as a motivation is more vague, and you aren't addressing why Apple should suddenly start enforcing this by guessing how difficult it could be to ensure. Which means a lot of your argument boils down to "look how difficult this thing I made up sounds, no way anyone else could do it" which is giving Apple a lot of credit and third parties the opposite.
I didn’t make anything up. I also didn’t say no one else could do it. I did say that specialized equipment is needed to do it.

If you‘re asking whether Apple or its customers should be motivated by quality then, again, that‘s a different discussion.
 
Imagine your brand new $70k Tesla has its tires remotely locked by Elon Musk, and swapping the tires is going to cost you an arm and a leg.
Tesla is one of the worst examples for repair of any vehicle. Try googling the ridiculous wait times some people get for replacement parts on their vehicles. The cost would be the least of your worries, as sometimes you'd be SOL for months.
 
There was another guy on youtube that discovered this before them. Maybe it would be nice to give him credit for it.

They did give him credit
 

Attachments

  • E8D2D9AB-DAE7-488F-9CA5-DB9F2157DE14.png
    E8D2D9AB-DAE7-488F-9CA5-DB9F2157DE14.png
    1.2 MB · Views: 95
So your telling me only a Certified Apple Authorized Service Provider can properly fix apple products which anyone can get with a little studying not really a big deal
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.