First, the issue is real. The problem seems to be that if you repeatedly autofocus on a stationary subject, the camera will autofocus at different distances. I've found the inconsistency in focus is more the norm than the exception. This phenomenon happens with both AF-S and AF-C modes.
Using the shutter button (there is a reason to use the shutter button rather than back button focus) to activate AF and making repeated exposures (or even just autofocus activations) with mostly wide angle lenses, the camera will focus at different distances and not repeatedly focus on the correct distance of the subject. Most reports say the lens has to be stopped down to f/5.6 or f/8, but in my tests the focus inconsistency also occurred when the lens was wide open. In order to see this, you have to add the 'AF Distance Indicator' in your screen setup and display options. Set your camera up for AF-S and just repeatedly focus on a subject across the room and watch the indicator on the focus scale stop at different distances. There is no consistency. To me this is disturbing as I have never had an AF camera do this.
Also, it is interesting to focus on a subject in AF-C mode, half pressing the shutter button, and watch the distance scale on the LCD jump around when it should stay at one distance. Again, the camera should select one distance and stay there. Even using AF-C the camera should not continually vary the focus distance on a stationary subject.
My tests reveal two items of importance. First, I can directly see the variation in autofocusing distances. Second, when looking at my resultant images, I can see if that inconsistency has a perceptible impact on the image sharpness or is it inconsequential. My tests revealed that although the distance indicator was off most of the time, only about 40% of the images I made for the test were slightly but perceptibly out of focus. I suspect, in some images, depth of field compensated for any loss of sharpness. Additionally, this seems to occur using back button autofocus as well but with back button autofocus, the AF distance scale on the LCD doesn't activate so I had to look at my test images to see if any mis-focus occurred.
As an example, with a subject exactly 10 ft. from my camera, contrasty focus targets and good light (so the AF mechanism didn't have any reason to struggle to acquire focus), the focus distance indicator would vary from just over 5 ft. to 30 ft. Most (yes, most) of the images never actually focused at 10 ft. at all! Most showed just over 7 ft. Also, the wider angle focal length, the more the variation. Using a Fujifilm 10-24mm f/4 lens at 10mm, sometimes the scale showed infinity focus! Many showed a focus distance of 30 ft.
I also conducted tests outdoors on a sunny, blue sky day with the sun at my back. I photographed a house about 200 ft. from my position. The camera focus varied from 14 ft. to infinity. One exposure would have the camera focus at 30 ft., the next infinity and then a third at 14 or 15ft. Unbelievable!
Previously, I had not heard much about problematic and inconsistent focus with Fujifilm cameras but I think this issue is much more widespread than I previously imagined. I've found several places on the Internet with Fujifilm camera owners have discussed this same issue. Also, I thought maybe the issue was only with X-T4 cameras but I have found owners of X-T3, X100, X-H1 and XE camera bodies who have experienced the same issue.
As part of my testing procedure, I photographed the X-T4's LCD so I could record the variation in focus distances. Below are a couple of examples of what I saw.
 |
| Camera was on a tripod and the distance to the target was exactly 10 ft. Notice the focus indicator reflects about 20 ft. focus distance. 10-24mm lens @ f/4. |
 |
| Exact same shot. Focus is now at infinity. |
 |
| Focus is on the side of the house. Notice focus distance is at infinity. 16-55mm f/2.8 lens @ 16mm; f/8 |
 |
| Same shot. Now focus is at 15-17ft. |
 |
| Same shot, now focus is at 14 ft. Lens at f/2.8. |
At this time, I have no solution for this. I've tried configuring my camera in every possible way and the problem continues to exist. Evidently, it is a systemic AF firmware issue that can't be worked around by the user. It seems that the Fujifilm engineers will have to fix it. I recommend you set up and conduct a similar test to see if your camera's AF system is erratic. Just because some of us suffer with this, doesn't mean everyone's camera has a defective AF system.
If anyone knows one of the “Fuji Guys” or a Fujifilm tech representative, please contact them and advise them of this widespread and unacceptable issue. Ask them to report a bug to the Fuji engineers.
This inconsistency issue has potentially two implications for my past work. First, images I thought I had missed focus or suffered from slight camera movement may have been a result of this focus inconsistency issue. It may have been the camera and not me!
Second, whenever I purchase a new lens, I thoroughly test it for proper working order and sharpness from center to edge to corner. It is possible that lenses I may have sent back were fine and the camera body was the problem. That really is unfair to the customer and retailer.
Several years ago, when I had my X-T2, I also purchased my first 100-400mm f/4.5-5.6 Fujifilm lens for wildlife and bird photography. My success rate was abysmal. I could only achieve proper focus on maybe 30% of my images—even tripod mounted using an expensive tripod and gimbal head. I thought it was me. Later, after I had moved to an X-T3 and purchased another 100-400mm lens, my hit rate was in the 80% and higher range. I remember telling friends that my particular X-T2 and lens just didn't work well together as evidenced by the newer combination. I wonder if the cause was this issue?
I am quickly losing confidence in my Fujifilm camera to provide me with the best possible and consistent autofocus. I'm starting to have thoughts of just selling it all rather than worry about whether or not my images will be consistently sharp or marginally out of optimum focus. Additionally, I haven't seen anything from Fujifilm about this issue and I don't understand why they haven't come forward in a very public way to advise users of this issue and let us know they are doing everything they can to correct it.
For now, I plan on setting my camera for manual focus and use "Instant AF" in the AF-C mode with back button focus activation. With this configuration the Manual Focus scale shows in the EVF as well as the LCD so I plan on glancing at it after initializing AF to ensure the distance displayed is consistent with the distance of my subject. I then can continue to re-initialize AF until it is correct before fully pressing my shutter button. That said, this is far from how things should be. So far in my tests, this method seems to be as stable or maybe just a bit better, but doesn't totally fix the problem.
If anyone knows of a successful workaround, please comment so all of us Fujifilm users can benefit.
More about this issue on Tuesday.
Thanks for looking. Enjoy!
Dennis A. Mook
All content on this blog is © 2013-2022 Dennis A. Mook. All Rights Reserved. Feel free to point to this blog from your website with full attribution. Permission may be granted for commercial use. Please contact Mr. Mook to discuss permission to reproduce the blog posts and/or images.
Odd situation. The Lenstips site tests lenses for repeatability of focus. The Fuji 16-55 test indicated near perfect focus.
ReplyDeleteMy Olympus consistently throws in out of focus images in a series taken at a high rate with a telephoto lens. I have not seen the problem at single exposures with a wide lens. I have found the single dot focus does not always perform focus where indicated in the viewfinder.
Dear Dennis — been following you since the early days where Olympus and Nikon were your main systems to talk about — I really appreciate your work and efforts not only talking about gear but also having some very meaningful thoughts about the creative process and photography per se.
ReplyDeleteThanks a lot for this very important yet eye opening article addressing a serious problem I noticed even back in older Fujifilm cameras though not being a real issue. I'm a press photographer and in most assignments I have to work fast and accurate. With my Fuji cameras I always noticed that in a series of pictures of a let's say a portrait session each and every shot was slightly different with the focussed area even though face recognition was on and the person stood perfectly still. With objects – as you mentioned in your article – the sharp area of the photo taken uses to shift back and forth ... :-( ... recently I did a test with the 16-80 mm f/4 zoom lens to find out whether it's sharp enough to use wide open or better stop down to f/4.5 to get sharper results. Photographing different objects with tripod at the same distance I got an annoying mixed bag of results: either the picture at f/4 was sharper than at f/4.5 or the focus was slightly off on both aperture settings.
I made some short tests with my X-T4, X-T2 , X-100T and a X-T1 and the problem with the focus accuracy is much more prominent on the X-T4. But it is present on all my cameras I tested with the X-100T having the least problems.
I hope that there will soon be a solution as a problem with focussing gets me in a shaky and uncomfortable situation having to work fast and precise. But maybe it's a little bit like in the older days where I new that I had to take some more shots with my DSLRs to have a couple of keepers for shure ... ;-)
Greetings & have a good time
Enzo Lopardo from Switzerland
Hi
ReplyDeleteI have had this problem on all 4 fuji cameras owned.Two Xpro 1,XT10 and XH1.
Many articles say that the scale is variable but the focus stays good. If only this were so !
As you demonstrate any variation on the scale distance gives a variation in focus also.
My worst focus variation was an older 18-55 on an Xpro 1.But later lenses on the XH 1 still show the problems.Im guessing most people dont use the scale and are unaware of the focus variations until editing.Thanks for bringing this issue to light.In short the focusing is poor.
Stupid question, do y use shutter priority or focus priority?
ReplyDeleteAF-S is set for focus priority and AF-C for release priority. What is distressing about AF-C is that if I press and hold the shutter button (or back button) to autofocus, the distance indicator jumps around as though it can't find anything upon which to focus. On occasion, it will settle down to one distance after a few seconds.
ReplyDeleteI had the same issue (inconsistent S-AF during landscape photography) with my X-t4 and the X100V. I already found a report (DPreview forum) about similar looking issue with the X-H2.
ReplyDeleteI think the most interesting question is: why not a single reviewer talks about this issue? They don`t realize? They don`t care? Anyhow, in this way there is no pressure on Fuji to get this issue solved...
I have shot with Fuji since x-t1, then x-t3, x-t4, and x100v. I have focus issues and in the past was wondering if it's my fault. but more and more I see, on sessions, sequences of photos where some have very weird, irrelevant distance of focus.
ReplyDeleteAs much as I LOVE Fuji and the look+ease of their film simulations, I've started on migrating. hard thing, when you own so many lenses, but... focus is a hard constraint!
Rob, thank you for your comment. I, too, started to fully migrate away from Fujifilm but I hung on to my Fujifilm gear until I could test the new X-H2S. With the last firmware update Fujifilm offered on the X-T4 (and other cameras) and the improvements with the new X-H2S, X-H2 and X-T5, I have not found any AF issues like the ones I experienced in the past. These new cameras are consistent and accurate in their AF behavior. I have now keep my Fujifilm gear and have added one of the new cameras. In fact, I’ve ordered the X-T5 as well with full confidence that the previous AF consistency issues have been solved.
DeleteI would like to see some images than, I think you just don't notice the problem. Even if it gets the focus right it is still less sharp than a nikon for example.
DeleteI can tell you this problem is still there even with the 5.0 firmware update for xh2s I have 2 bodies that have exactly the same issue.
DeleteI've just run additional tests with both my X-H2S and X-T5. See tomorrow's blog post (August 11, 2023) for my findings.
DeleteGreat but also dissapointing article as i also face the same problems with pumping, inconsistent autofocus on my X-H1 and by now i have not find any solution for that problem so i switched back to my old canon equipment which does not have these problems.
ReplyDeleteUnder the given circumstances the camera is far away from the expectation you have if you look on the price and is not for professional use where you always need crispy pictures.
I have the same issue. The software on Fuji camera never feels polished. It's like they're cramming as many new tech in their new cameras and never sat down and test out what is working and what is not. My other complaint is the touch screen. I have touch to focus on, but there is always a delay. I touch the screen, the AF box goes to where I touched, but the AF always only starts working 1/2 second later. This is true on my XT3, XT4, XA7. It is so annoying. I'm think about selling all my Fuji gear all-together and going back to DSLR
ReplyDeleteI have seen this with other Fuji cameras especially the XT2, XT3. I have never trusted the digital readout of the distance to subject!
ReplyDeleteA friend has noted this as well with his XT2 cameras (he has three) and has moved to shooting in manual focus.
I always assumed it was down to either some micro-movement on my part as I seldom use my tripod for my Fuji X cameras, only for GFX.
I havent noticed it on my XH2, XH2S or XT5. So at this point I would watch for it but not do anything with the camera.
I think that if it happens consistently then Fuji should own this as an AF problem they need to fix urgently.
The data from 'thewanderinglensman.com' does look convincing but I would prefer to see it repeated with a 100-400 or 150-600 as these will provide a more accurate determination of where the focus point is located on the subject.
Unfortunately or fortunately it can be something as simple as trying to focus on a glass of a window or a fuzzy edged object and mirrorless AF accuracy goes to pot. I tested this with my Nikon Z7 v Fuji XT3 v Nikon D850. There are ways to trick the AF on all the cameras.
At present I am vascillating as I really do prefer the physical layout of the XT5 but the customisations offered by the XH2 and XH2S around the fn buttons allows for very easy use as well.
Thank you for your comment. I’m not sure what Fuji did, but after the last firmware update for the X-T4, the AF issue was largely corrected, at least in my camera. I have since sold that camera so I can’t test it any longer. I now have both the X-H2S and the X-T5 and I’m having no AF accuracy or repeatability issues that I can discover, and I have tried. As far as testing the 100-400 or the 150-600, I only have the 100-400 and the accuracy and AF repeatability was much better at telephoto focal lengths than at wide angle ones. In fact, I found the shorter the focal length, especially stopped down at the10-20mm range, the problem was at its worst. Long telephoto focal lengths had very few errors. I bought the X-H2S first, mainly for my occasional bird and wildlife forays as well as family gatherings, grandkids sports, etc., i.e., when speed and subject tracking is beneficial. I fully programmed the C1-C7 dial settings for birds-in-flight, birds-in-flight with pre-shot, people and events, trains at high speed, etc. Those custom settings are very useful. That said, my planned use for the 40mp X-T5 is for more deliberate and measured photography such as landscapes, travel, close-ups and macro work, etc., where fine detail, texture and high resolution made better sense. As much as I enjoy using the X-H2S, I’ve fallen in love with the X-T5. It may just be that I’ve found the perfect camera for me.
DeleteI do have this issue with Fuji X-t5 and both XF 10-24 V1 (especially at 10mm) and XF 16 1.4. Very inconsistent results at 8.0
DeleteThe solution (a workaround indeed) is instant focus. Put the camera in M and use AF-ON button
DeleteThe workaround is to use instant af: put the camera in M and use afon button
ReplyDeleteI have an X-H2S and I still have these extremely frustrating AF issues on the latest 5.0 firmware. I mostly use the 150-600 for birding and I often get out of focus or very soft shots. The camera says the subject is in focus, but it's not accurate. I call it false positive auto focus locks. I find myself having to take multiple bursts, reacquiring focus each time, and hoping that I'll get a very sharp photo.
ReplyDeleteSame problem and findings here! This has been a problem for quite a while now. Maybe, Fuji cannot afford good software developers, or are they about to go bankrupt?
ReplyDeleteXF 10-24, XF 14 2.8, XF 16 1.4 and XF 16-80 are all extremely unreliable save for 16-80 above 50mm!
They are useless if I can't trust them. By the looks of it, Fuji is not going to fix the problem any time soon since it is present in all generations of their cameras.
Decided to swallow up the costs and started moving back to Canon. Just ordered R6 and RF 24-105. If I can cope with the weight, I'll sell off my Fuji ecosystem in favor of Canon.
It’s now mid September and the latest Fuji summit revealed nothing about releasing a solution to the problems we are having with autofocus. My camera has been returned to Fujifilm UK twice and is now back at firmware 3.0 but this hasn’t solved the issue, it only reduced the problem minimally from the dismal 5.02 release. I invite Fujifilm to officially make a press release telling their customers (for the time being) what they intend to do and by when. I fear there is either a lack of talent in the Fujifilm software people or an ostrich mentality in senior management.
ReplyDeleteThis issue is real and tends to be lens related, in my opinion. The only two lenses that I failed to recreate this issue with were the new 33/1.4 and 23/1.4 (LM series) from Fuji. Each time I focused on a stationary subject I got a perfect focus (I'm using X-S10 with updated firmware). If I take the 35/1.4 (An old Fuji lens) and half press the shutter for a few times on the same subject from the same distance, most of the time the focus will be "all over the place". A very frustrating issue which makes Fuji's AF super unreliable.
ReplyDeleteThanks for your comment and insight. I was a huge critic of Fujifilm’s AF system due to its issues with AF inconsistency. That said, a lot of things have occurred since I originally wrote these series of posts.
DeleteI’ve now had an X-H2S and an X-T5 since their introductions and, after thorough and repeated testing over the past year or more, I’m not having the inconsistency issue with either. That’s not to say that in some copies of these camera bodies it doesn’t occur as I still get comments and private emails from disgruntled owners. But there are an awful lot of pro and enthusiast photographers using these bodies and I’m not seeing or reading widespread comments and complaints about the AF issue. I’ve listened, watched and read literally hundreds of blog posts, articles and YouTube videos specifically looking for comments on the inconsistency issue. I’ve come to the conclusion that the issue probably isn’t as widespread as I first thought. Also, over the past year or so, Fujifilm has made a few firmware updates which might have silently addressed the issue without acknowledging it publicly. Again, I’m having no issues at all nor are photographer friends who have X-T5s reporting issues to me with AF inconsistency.
Finally, I’m still a little paranoid about focusing accuracy based upon my experience with my X-T3 and X-T4 and find myself actively checking focus after making exposures. As President Reagan used to say about the old Soviet Union, “Trust by verify.” ~Dennis
Thanks for bringing this topic up. Rather late input but I've recently purchased an X-S10 coupled with the 35 1.4 R out of interest as Olympus user mainly. The kit misses focus indeed frequently sometimes by a mile despite the focus confirmation hitting home. Occasionally, I wouldn't know where the camera focused actually meaning either infinity for a close subject or absolute near focus for a landscape. I could be a coincidence but I noticed this happens when the subject under the focus point is relatively dark vs background but as I said, it could be just a coincidence - this would never happen with my Olympuses at least to my knowledge. Not to mention a minor shuttershock issue but that's another topic. Not being my main kit, I set it now permanently to manual focus (and ES not to worry about shutershock too much). I just "pre-focus" with the back AF button and double check by rotating the focus ring.
ReplyDeleteThank you for commenting. The most accurate AF focus with Fujifilm cameras is when you set the selector to Manual and then use Back-Button AF (BBF). That is configured in the menu for either AF-S or AF-C. I use Manual Focus and AF-C in this manner. I don’t think there is an AF-On button on your model, but as you seem to have done, reprogram the AF/AE Lock button for BBF. The reason for this is that the camera focuses with the aperture wide open and only stops down when you press the shutter. Using normal AF-S or AF-C causes the camera to AF at the stopped down aperture, which in my opinion, allows the camera to AF anywhere within the depth of field and vary in accuracy. All this said, I’m now using both an X-T5 and X-H2S and have no AF issues whatsoever. I’ve test both thoroughly and am satisfied that whatever Fujifilm has done with firmware and the new image processor, it seems to have cured the issue, at least in my two cameras. ~Dennis
DeleteOh no theres an AFON button next to the AEL on the X-S10 which is quite a nice touch actually considering the "entry-level" bucket_I wouldve used manual focus fully but focus by wire is painfully slow. In either case am enjoying the setup particularly on forest walks. Also thank you for your further advice.
DeleteInteresting coming across this! I've noticed this issue with my XH2S so I think I'll reach out to Fujfilm... I just tested with my XT3 to see if I was just imagining this with my XH2S but my XT3 focuses much more accurately.
ReplyDeleteWondering if anyone has had any luck approaching Fujifilm about this issue?
I called and spoke with a Fujifilm technical representative on three occasions and also sent my X-T4 in for a complete check highlighting the AF system and the AF inconsistency issue was never acknowledged. I believe there is a fundamental issue with how they have engineered the AF system that causes the inconsistency. I recommend switching the focusing option to Manual, the setting the menu item to allow the AE/AF Lock button or the AF-On button (if your camera has one) to activate either AF-S or AF-C while in manual focus mode. Using AF in the normal fashion causes the lens to stop down to its set aperture when autofocusing but with the camera set to manual focusing and using back-button AF, the lens’ aperture remains wide open, thus reducing the depth of field and increasing the AF accuracy. ~Dennis
DeleteRecently i began suspecting something could be wrong with my 50-140. So tested it with my X-T5 and X-T3 (both on the latest firmware), and added the 90mm to the mix for comparison.
ReplyDeleteCameras on tripod, ibis disabled, AF-S and BBF.
Used f2.8 and f4 on the 50-140, and f2 f2.8 and f4 on the 90mm.
Tested the 50-140 at 50/90/140mm.
Re-focused each time i changed aperture and focal length.
What i got was totally inconsistent results. One would expect sharper pictures from smaller apertures, and the same results whichever body used.
But that´s not what i got. For example, on the 50-140 at 50 the f4 shot would be the softer one, but at 90 that was the f2.8, as expected. Switching cameras the results would be different. Same thing with the 90mm and both cameras. Inconsistency.
What´s going on?
I might be recalling wrong, but i think that my (now gone) X-T2 and X-Pro2 didn´t have this issues. Been taking a look at shots taken with them many years ago and i see really tack sharp results. Much better acutance on the 50-140mm back then than the results i was getting now that prompted me to do these tests.
Admittedly, back then i mainly used manual focus mode with BBF set to AF-C, but the fact that focus tracking stops the moment the shutter is half-pressed made me stop using it.
Hi all, i just bought a XS-10 in December 2023 with a XF 18-55 and a Tamron 18-300 and end up in despair with lacking the sharpness. Did hunderets of shots but have still the issues.
ReplyDeleteI thought I had this issue, and felt silly realized the auto focus manual dial on the front of my camera was accidentally switched.
ReplyDeleteIt’s May 2024 and I have been experiencing this issue most noticeably with my XH2 using the 10-24, 150-600. I do a lot of nature photography. I have had MANY scenarios where I am lucky to creep up on a subject- autofocus box is green and looks to be focused on the EVF- I chimp a few images and notice entire sequence of image is soft- the subject isn’t moving - I’m on a tripod and shutter speed is 1/2500 of a sec. and using subject detection, AF box finds the subjects eye. Images totally soft. Started doing some research and see many people saying the continuous focus will not lock onto static subjects - This is beyond frustrating as I have missed quite a few precious moments and many missed shots that if sharp- would be amazing shots. But I’ve come back with entire card full of soft garbage. When the system works - you can create some incredible images. I don’t like that I can be doing everything right- and the system fails me. A LOT. That’s not acceptable for us who do this as a way of life and rely on quality work to pay the bills.
ReplyDeleteMy X-T10 with 18-55 f2.8-4 has the same hunting in AF-S using holding the shutter button - but not always which is confusing. If I switch to AF-M and back button focus it nails the same focus every time. Since I back button focus on my Sony it's not a big deal.
ReplyDelete