Thestructured
Well-known member
I'm not asking for help in this thread, I'm just warning others to check their new cameras.
Last edited:
-
Welcome to the new forums! Please read this first. For known issues we are working to resolve, click here.
5d mark IV left side focus issue & poor calibration.
- Thread starter Thestructured
- Start date
chironNYC
Senior Member
I'm not defensive (I'm not buying the camera), but I am always a bit skeptical of certain kinds of posts, and yours has some of the warning signs:
a new camera comes out, and someone with a very brief posting history and who only owns lenses for another brand suddenly begins to post more in one thread than he has ever done before about any other topic, warning people about an obscure problem that doesn't really get reproduced in other shots and that no one else seems to have or to report. And the problem is one that the Nikon 800 actually had when it was first introduced.
Maybe its real, but none of us know you from anything before this and there is nothing wrong with some healthy skepticism either.
Also, I didn't quite get the logic of your not wanting to wait 2 weeks for Canon to repair the camera because you are trying to sell all of your Nikon gear, but you shoot a wedding a week? Don't you need to have a new system in place before you sell your gear, or did I misunderstand something?
No one else seems to have a camera that does this. There are some image threads at fredmiranda.com and elsewhere that people can check out.
BTW, who is the "other person who also confirms this is happening with theirs" that you mention in your first post? Forgive me, but what are the odds of you and one other person you know having this problem and no one else reporting it?
What you are reporting could be straight up, but it is reasonable to ask some skeptical questions.
Last edited:
Le Kilt
Veteran Member
Some folks lurk until there is a strong reason for them to post. And some people post without having much reason at all. This can sometimes mirror reality with those that are vocal and those that are not in company meetings. Yes there are folks with a penchant to just yap yap yap and fill in space at a company meeting, but sometimes the guy who rarely speaks up has the most important things to say.
This is how the internet has always worked. There is nothing strange or suspicious about his behavior. But then again people getting suspicious of 'new' posters is also not new phenomena. However I would say the likelihood of sophisticated espionage is little to none. Yes fake reviews get posted all the time, but they read much like you'd expect a third world data entry sweatshop would and they would not engage in any intelligent dialog of any sort.
You might as well say Samsung's battery fire problems is corporate espionage as well.
The idea of multi-point AF being off is not new...in fact it is probably more common than most people are willing to test for.
https://www.lensrentals.com/blog/2016/09/is-your-camera-really-the-best-optical-test/
Read Roger's most recent review on lens variance and testing. He pretty much admits that folks with discerning eye and good test technique combined with high MP sensors can reveal a lot of defects manufacturers are not even ready to correct for (in other words if you get a lens that just doesn't meet up to your standards, do not expect that a trip to the service center is going to repair it!). Same thing for multi-point AF...it's a *reality* that multi-point AF is not perfectly calibrated...it is something Rishi has commented on frequently enough but never so much as to publish an expose on it.
So the sad reality is - Many many lenses can have significant variation especially when wide open, and many many autofocus points have significant variation especially when tested wide open. Most people just have to dial in their expectations and realize they just are *not* going to nail tack sharp images on 36+MP FF sensors or 24MP crop sensors all the time and that both luck, persistence, and diligence (and unfortunately even chimping) is still very much a part of photography.
So PDAF mechanisms have extremely tight tolerances that manufacturers have yet to fully calibrate for (this is not speculation this is *reality*, they can't even do this for ONE center point with their lenses, what is the chance they do it for *EACH* AF point?). Just as there can be front/backfocus drift on PDAF point there can just as easily be *TILT* issues (which is also a very common issue in lenses). And beyond these tolerances is the *fact* that shipping companies are generally extremely rough with shipping.
This is how the internet has always worked. There is nothing strange or suspicious about his behavior. But then again people getting suspicious of 'new' posters is also not new phenomena. However I would say the likelihood of sophisticated espionage is little to none. Yes fake reviews get posted all the time, but they read much like you'd expect a third world data entry sweatshop would and they would not engage in any intelligent dialog of any sort.
You might as well say Samsung's battery fire problems is corporate espionage as well.
The idea of multi-point AF being off is not new...in fact it is probably more common than most people are willing to test for.
https://www.lensrentals.com/blog/2016/09/is-your-camera-really-the-best-optical-test/
Read Roger's most recent review on lens variance and testing. He pretty much admits that folks with discerning eye and good test technique combined with high MP sensors can reveal a lot of defects manufacturers are not even ready to correct for (in other words if you get a lens that just doesn't meet up to your standards, do not expect that a trip to the service center is going to repair it!). Same thing for multi-point AF...it's a *reality* that multi-point AF is not perfectly calibrated...it is something Rishi has commented on frequently enough but never so much as to publish an expose on it.
So the sad reality is - Many many lenses can have significant variation especially when wide open, and many many autofocus points have significant variation especially when tested wide open. Most people just have to dial in their expectations and realize they just are *not* going to nail tack sharp images on 36+MP FF sensors or 24MP crop sensors all the time and that both luck, persistence, and diligence (and unfortunately even chimping) is still very much a part of photography.
So PDAF mechanisms have extremely tight tolerances that manufacturers have yet to fully calibrate for (this is not speculation this is *reality*, they can't even do this for ONE center point with their lenses, what is the chance they do it for *EACH* AF point?). Just as there can be front/backfocus drift on PDAF point there can just as easily be *TILT* issues (which is also a very common issue in lenses). And beyond these tolerances is the *fact* that shipping companies are generally extremely rough with shipping.
Last edited:
MinAZ
Veteran Member
If that is the case, then why haven't I noticed the same problem in my camera? Is it because the AF alignment issue is very slight and I would need a sharper lens to catch the problem?
it is interesting to see you having issue with AF with all your canon and nikon cameras and me, on the other hand, have 4 canon cameras and never had any issues with my cameras and that goes for my 14 canon lenses as well! is this some kind of peculiar accidental or something else, i wonder?Exported it with exif with DPP. I'm not exactly sure what exif is so important to anyone other than the information I already gave. It's kind of funny how people are getting defensive. I've gone through a lot of equipment over my 6 years shooting weddings full time and this is far from the first time I've had a camera with AF misalignment. It's not uncommon. It's just annoying to get sorted out because some times the service center doesn't get it right the first..two times and then you're sending cameras back and forth for what seems forever before it's made right.
Because like the Samsung exploding battery, its a problem that has an overall low defect rate, but *unlike* the Samsung exploding battery, your DSLR does not explode if you have a bad one. Like lens variation what you get *varies*. Most DSLR's should deliver multi-point AF where each point should be +/- 5 MFA points away from the center as something being relatively 'in-spec'. When it becomes photographically relevant and more noticeable is when this skew approaches +/- 10 MFA points or more. Something like the D800 issue that was reported widespread probably went past 20 MFA points. That doesn't mean someone who is observant and *knows* how to test their DSLR might not notice their 5DMKIV being off by 10.
Also some people to this day including Tony Northrup still don't even believe you need MFA if ever...so IMO it sort of depends on how observant a pixel peeper you are. Despite being a charming talker, Tony mostly parrots DxOMark, doesn't believe too much in MFA or lens variation that I can see, and see's what he wants to see (i.e. anything supporting DxOMark like the lie that AA-less sensors resolve just so much more detail which he rates at 30-50% improvement in line with DxOMark). So to me he's not a very observant pixel peeper in my book...just good enough to know when something is in sharp focus or not in general. Fred Miranda by Roger's account seems to be a pretty observant pixel peeper and I bet he believes MFA makes a difference and is often necessary. I don't know how anyone can trust a guy in rating autofocus or lens performance or AA-less sensor spatial detail resolution when they don't even bother or believe in MFA....
You need a bright prime shooting at moderate distances. You most likely won't notice it as much shooting say a portrait prime at close up portrait distances, but shooting a portait prime as a landscape lens (wide open), and the bugs start to come out of the woodwork.
Last edited:
noms78
Well-known member
https://www.dpreview.com/forums/post/57851559
https://www.dpreview.com/forums/post/57858887
I don't want to dig up anything personal but Tony IMO backpeddles a bit from not believing in MFA to just stating that he has been enormously lucky as to never need it...ever in the history of owning DSLR's and lenses. Sure thing whatever. You see when you are charismatic and put video's on youtube you deserve both benefit of the doubt, and even having dollars just thrown at you for generating ad revenue. In fact you should be able to make a living just sitting on your ass playing video games and pull in millions and buy exotic sport cars if you wanted...that is the power of YouTube.
IMO Kris may not be as photogenic and as smooth a talker in front of the camera, but I'll trust his capabilities shooting birds behind the camera.
Perhaps I over-exaggerated how much worth Tony puts into AA-less sensors, I don't really pay attention to his DxOMark is all knowing banters to be honest. From *my* perspective he seems like a typical I shouldn't have to worry about MFA or multiple copies type of consumer. In a way I absolutely agree with him...as a CONSUMER I do *NOT* want to worry about copies of lenses or multi-point MFA focus skew issues. But when I see experienced folk like Kris, Fred Miranda, and Roger Cicala talk about not only lens variation but body variation, I can't help to take note especially having experienced similar throughout trying out various gear. Lets just say I also wish I was so blessed as to *never* need MFA...*ever* throughout any of my DSLR or lens purchases. Tony hints at buying a *LOT* of pre-owned gear which IMO makes it incredibly fortunate...not only does he have a strong youtube following, he is also incredibly blessed by the lens lottery gods for pre-owned gear.
https://www.dpreview.com/forums/post/57858887
I don't want to dig up anything personal but Tony IMO backpeddles a bit from not believing in MFA to just stating that he has been enormously lucky as to never need it...ever in the history of owning DSLR's and lenses. Sure thing whatever. You see when you are charismatic and put video's on youtube you deserve both benefit of the doubt, and even having dollars just thrown at you for generating ad revenue. In fact you should be able to make a living just sitting on your ass playing video games and pull in millions and buy exotic sport cars if you wanted...that is the power of YouTube.
IMO Kris may not be as photogenic and as smooth a talker in front of the camera, but I'll trust his capabilities shooting birds behind the camera.
Perhaps I over-exaggerated how much worth Tony puts into AA-less sensors, I don't really pay attention to his DxOMark is all knowing banters to be honest. From *my* perspective he seems like a typical I shouldn't have to worry about MFA or multiple copies type of consumer. In a way I absolutely agree with him...as a CONSUMER I do *NOT* want to worry about copies of lenses or multi-point MFA focus skew issues. But when I see experienced folk like Kris, Fred Miranda, and Roger Cicala talk about not only lens variation but body variation, I can't help to take note especially having experienced similar throughout trying out various gear. Lets just say I also wish I was so blessed as to *never* need MFA...*ever* throughout any of my DSLR or lens purchases. Tony hints at buying a *LOT* of pre-owned gear which IMO makes it incredibly fortunate...not only does he have a strong youtube following, he is also incredibly blessed by the lens lottery gods for pre-owned gear.
Last edited:
Thestructured
Well-known member
It's pretty much impossible to align it all perfectly, but there's an expectable tolerance. Some times some cameras get shipped out of that tolerance and some times the shipping company isn't so careful with packages (my box was in great shape so I doubt that's the case). Most cameras are going to be slightly off but 'good enough' and nobody will notice a problem. Mine just happened to be 'bad enough' that I noticed it without actually looking for a problem. This doesn't mean every single 5d mark iv is going to have this, that's just not likely. I'm not trying to say that they will. Mine was pretty off though, and there's another guy in a different forum that got one just as bad. So it's this whole thread is more of a 'hey guys, you might want to check yours' than anything else, and if you're not having any issues, great!
noms78
Well-known member
Come to think of it - for slow moving subjects or landscape photography I think mirrorless has an advantage in AF over DSLR in the sense that they do not require AFMA - their contrast detect AF is super accurate. DSLR require AFMA which is only as accurate in liveview with newer lenses (not a problem as all my lenses are recent) and at particular distances - this is a bigger deal. If I want super accurate AF in low light or at distances I did not calibrate the AFMA to, I would need to use liveview. Once again, not a big deal when shooting static subjects on a tripod...
The EVF showing live histogram is another benefit, but at this stage I am happy to shoot with Canon DSLR as I like their colour and proven quality. The weight is certainly noticeable when travelling but that doesn't detract from the experience.
The EVF showing live histogram is another benefit, but at this stage I am happy to shoot with Canon DSLR as I like their colour and proven quality. The weight is certainly noticeable when travelling but that doesn't detract from the experience.
Last edited:
Forgive me but this stance is irritating to anyone who has had camera gear issues and had that response thrown at them. Speculating that it's some kind of user specific peculiarity is rather what one attributes to the feableness of irrational conjecture.
Thestructured
Well-known member
I've owned a LOT of camera bodies and lenses, and the overwhelming majority of the time my lenses do not need any AFMA. And the few that have needed it, were not accurate all distances so they were sent in with the camera body to be fully calibrated together. I wish every manufacturer could adopt Sigma's take on this, allowing you to fully calibrate a lens yourself at all distance ranges with the dock. That's one reason I love the art line so much. AFMA in camera with one global value rarely remedies an issue.
chironNYC
Senior Member
Which other forum was it where another guy reported the same problem?
Thestructured
Well-known member
First, thanks to "Thestructured" for raising this issue. I read the thread a day before my 5D4 arrived, so I did focus testing early and often.
I can confirm a serious focus problem in my unit. I've now completed a rigorous series of tests, using a constructed target range. I've tested the 5D4 against my old 5D3, using a Canon 70-200 2.8 and a Canon 100mm macro 2.8. I've got a Canon 24-105, but I'm not going to bother. I've seen enough.
Significant focus problems on the 5D4, zero problems on the 5D3 using same lenses, same target, same exposures, same everything. No 3rd party lenses here, no Nikon bias, 40 years experience.
For a test bed I made a large flat target with high contrast black on white text. A flat field covering the entire range of focus points. Carefully positioned to be parallel and level with the camera. Brightly lit with flood lights. Exposures at 1/640, ISO 400, f/2.8, 16 feet to target.
I shot in Liveview, Single-point-precision, and Large Zone multi-point. For Liveview and Single-point precision I shot at frame center, center-left, and center-right. For each position I shot 3 to 5 identical frames. Between 21 and 40 frames for each round. For each frame I manually de-focused before auto-focus. Converted raws to tifs with ACR at defaults, loaded multiple frames as layers in Photoshop, toggled layers to compare.
The 5D3 nailed every shot. I have to blow up to 300% and pixel peep real close to see minor differences.
The 5D4 is all over the board. Even the single-center-point-precision shots vary. Three to 5 shots made back to back as fast as I can defocus, auto focus, and fire all show significant differences. They range from good to maybe OK, to bad, to really bad. Liveview has a higher success rate than single-point. Large Zone multi-point was by far the worst, and no better at center, left, or right.
I called B&H pro tech support. They said they had no similar complaints and no returns due to focus problems. They asked me to call Canon. Canon said the same. I felt both agents were fair and forthcoming, not trying to hide anything. Both encouraged me to send the camera in. I'll definitely do that, but am in a quandry. Should I swap if for a different body from B&H and risk getting the same thing back? Or send it to Canon and pray they can fix it? I'm leaning toward the Canon route.
I can confirm a serious focus problem in my unit. I've now completed a rigorous series of tests, using a constructed target range. I've tested the 5D4 against my old 5D3, using a Canon 70-200 2.8 and a Canon 100mm macro 2.8. I've got a Canon 24-105, but I'm not going to bother. I've seen enough.
Significant focus problems on the 5D4, zero problems on the 5D3 using same lenses, same target, same exposures, same everything. No 3rd party lenses here, no Nikon bias, 40 years experience.
For a test bed I made a large flat target with high contrast black on white text. A flat field covering the entire range of focus points. Carefully positioned to be parallel and level with the camera. Brightly lit with flood lights. Exposures at 1/640, ISO 400, f/2.8, 16 feet to target.
I shot in Liveview, Single-point-precision, and Large Zone multi-point. For Liveview and Single-point precision I shot at frame center, center-left, and center-right. For each position I shot 3 to 5 identical frames. Between 21 and 40 frames for each round. For each frame I manually de-focused before auto-focus. Converted raws to tifs with ACR at defaults, loaded multiple frames as layers in Photoshop, toggled layers to compare.
The 5D3 nailed every shot. I have to blow up to 300% and pixel peep real close to see minor differences.
The 5D4 is all over the board. Even the single-center-point-precision shots vary. Three to 5 shots made back to back as fast as I can defocus, auto focus, and fire all show significant differences. They range from good to maybe OK, to bad, to really bad. Liveview has a higher success rate than single-point. Large Zone multi-point was by far the worst, and no better at center, left, or right.
I called B&H pro tech support. They said they had no similar complaints and no returns due to focus problems. They asked me to call Canon. Canon said the same. I felt both agents were fair and forthcoming, not trying to hide anything. Both encouraged me to send the camera in. I'll definitely do that, but am in a quandry. Should I swap if for a different body from B&H and risk getting the same thing back? Or send it to Canon and pray they can fix it? I'm leaning toward the Canon route.
MinAZ
Veteran Member
Sorry to hear about your problems with AF, it sucks for sure. One thing very interesting to me is that one of the lens you chose to use for testing (the 100mm macro) is well known to be incompatible with dual pixel CMOS AF. I wonder if that might be a contributing factor to your focus issues?
Loading…
www.canon.co.uk
Personally, I would exchange it once. If the second has the same issue, I'd send it to Canon.
noms78
Well-known member
So what you are saying is that you are seeing "inconsistent" phase detect AF using different AF points?
Assuming you MFA'd the 70-200 and 100 macro to the 5d4 body?
The fact that the 5d3 nailed the exposures with the same lenses, while the 5d4 did not suggests there is a serious AF problem.
Out of curiousity was the focus accurate in liveview when zooming in 5x or 10x and focusing?
Last edited:
Thanks, didn't know there were incompatible lenses. If I understand, that would only affect Liveview focusing. Strangely, my Liveview focusing with the 100mm on the 5D4 is fairly good and consistent. Better than the 70-200.
Attached is a sample. 5D4 with the 100mm. The three boxes are 200% screen captures. All are single center point precision focus (thru the viewfinder), made back-to-back within a few seconds. Defocus between each shot. This sample is very typical of what I'm seeing with both lenses on the 5D4.
Similar threads
Keyboard shortcuts
- f
- Forum