We’re often asked whether there really is any point to calibrating your camera autofocus. Obviously we’re going to say a big, fat yes! otherwise FoCal wouldn’t exist, but I wanted to show a few charts to explain the sort of performance you’re likely to get “out the box”.
When FoCal is used to calibrate or analyse a lens, results are uploaded to our database which is then used to feed back into FoCal and give you an idea of how your camera and lens performs against other similar equipment. But there’s an awful lot of data now – hundreds of millions of data points collected over more than 5 years – and this data can provide a valuable insight into how cameras and lenses perform.
Canon Cameras/Lenses
First, take a look at a histogram of the valid* calibration values calculated by FoCal across ALL supported Canon cameras and lenses:
(* valid results are those which pass our internal criteria for being “sensible” results. We remove any results from tests which failed under odd circumstances, those who’s result calculation quality is not very good and various other reasons that might mean the data is not really representative).
So what does this chart show? It’s a count of the results at each microadjustment value – the higher the yellow bar, the more results at this microadjustment value. The actual number is unimportant (for these charts the most popular AF Microadjustment value will go up to 1.0) and we can see that the most popular AF Microadjustment result across all Canon cameras and lenses is zero, which is great!
The blue curve is a standard gaussian curve and we can see that the results fairly well mimic this. The vertical blue line is the mean of all the results, and the red shaded area is one standard deviation which is a measure of the spread of the data. About 70% of all the results are within the red area.
What does this mean in the real world? Well, it should mean that if you pick any (FoCal supported) Canon camera and then grab yourself 10 of the same lens type from your (impressive!) shelf, then 7 of those lenses should need a calibration value between about -5 and +5. The other 3 are likely to need a calibration value outside of this range.
We generally say that it’s usually not worth sweating over 1 or 2 AF Microadjustment values, but 5 will make quite a difference to the quality of your captured image and the reliability of your autofocus. Which means even for the “good” 7 Canon mount lenses you took off the shelf, you can still get a fair quality improvement by calibrating, and for the other 3 you’ll definitely see a difference.
What about Nikon?
Here’s the same chart for all Nikon lenses:
This one looks a bit different. We still have a peak at an AF Fine Tune of zero which is good, but the standard deviation is quite a bit wider, and there are some significant numbers of cameras and lenses hanging around at those edge regions.
Again, taking our well stocked shelf analogy we’re now looking at 7 of those 10 Nikon mount lenses you picked up needing a value between about -11 and +10 which is a considerable adjustment.
So for Nikon, calibration is even more critical. 3 out of 10 Nikon mount lenses are likely to be very poor out of the box, and the other 7 will still require a potentially significant adjustment.
Summary
By combining millions of data points into these histograms, hopefully we’ve shown that it’s a really good idea to take advantage of the AF Microadjustment / AF Fine Tune feature of modern Canon and Nikon cameras. Statistically, it’s almost certain that tuning your autofocus will improve your image quality, so fire up FoCal and get calibrating!
18 comments on “Do I need to calibrate autofocus?”
Indeed a convincing argument in favor of calibrating AF. However, how do we know that the results FoCal, or any other product or method, provides are accurate?
I would like to see a follow up article that makes similar comparisons, but not between camera manufacturers. This one would be comparing calibration methods using various software and hardware devices.
Hi Ralph,
Good question, FoCal works by moving the lens to defocus and refocus multiple times and assessing the relative sharpness between captured images at different AF fine tune / microadjustment values. There are other ways to calibrate, Nikon for example has added an in camera calibration feature to it’s newest cameras (we did a direct comparison of the Nikon method and Reikan FoCal on the blog Nikon D500 Automatic AF Fine Tune )
How you might go about assessing the absolute accuracy is more tricky. We believe FoCal provides the most accurate result of any available method, how you can prove that is harder to do! One thing we’ve worked on is checking AF calibration against the dual pixel data available in the Canon 5D4 – that is a fairly rigorous double check and in those cases we can ‘prove’ the result is accurate, information on the blog Bringing Dual Pixel Raw to Reikan FoCal
Best Regards,
Dave
One intersting test would be calibrating you camera and lenses (like me) using FoCal at home.
Then going out into the field (different temperature, different color temperature, etc) and taking a large number of shots of various subjecs at various distances continually switching between the FoCal derived calibration setting and the baseline (eg offset = 0).
Then analyze those images by cropping the region where the user or the camera wanted or confirmed focus to be at. What is «sharpest» on average. If FoCal has any value in that it improves focus accuracy in a _general_ sense rather than merely improving the acutance of the specific test that has no other value than calibrating itself, I would expect the calibrated set to feature (on average) slightly more acutance than the non calibrated set.
Hi Knut,
Thank you for your feedback. One of the things we want to do more of are ‘science experiments’ 😉
We’ve played with AF microadjust / fine tune for a good few years and there’s a lot that’s still fairly obscure to most photographers. Some of the details are of no interest (unless you’re designing a tool for AF calibration!) but there’s much information that would benefit photographers that should be talked about and we’re in a great position to do that.
Watch this space! 🙂
Best Regards,
Dave
What are the units for the y axis? It’s not mentioned in the article.
Hi Jon,
The units across the bottom correspond to the available adjustment within the camera. On Nikon it’s called “AF fine tune” and on Canon it’s called “AF microadjustment”. Essentially the camera/lens can be calibrated within a range of -20 adjustment units to +20 adjustment units to allow for front/back focusing. There’s a good overview video linked in the FAQ that explains better Are there any FoCal calibration videos I can watch? 🙂
Best Regards,
Dave
I referred to the y axis, which runs vertically. Without units for both axes, a chart is meaningless.
Hi Jon,
Sorry, yes, the Y axis 🙂
“So what does this chart show? It’s a count of the results at each microadjustment value – the higher the yellow bar, the more results at this microadjustment value. The actual number is unimportant (for these charts the most popular AF Microadjustment value will go up to 1.0) and we can see that the most popular AF Microadjustment result across all Canon cameras and lenses is zero, which is great!”
Best Regards,
Dave
Is the more detailed info available on what are the cameras and lenses that fall into <-20 and +20 categories (essentially those that one can buy seriously defected). And those that come close to zero adjustments (that is those that are best in quality autofocus-wise)?
Hi Dmitry,
Great question, we wanted to present a very broad overview but there are differences between both lenses and camera bodies. I think it will need a more in-depth blog post (coming soon!) 🙂
Best Regards,
Dave
When using longer lenses, 300, 400, 500 & 600 how will this work if you need it tethered? I would assume that I would calibrate my 600 at around normal focus distance @ 20 mrs
Hi Steve.
Yep, FoCal does generally require to be tethered (though it is possible to pre-capture images and FoCal can analyse them without needing to be tethered). With very long lenses it might mean using FoCal on a laptop outside depending. With 600mm lens we recommend a minimum target distance of around 12meters (approx 39 feet). See also What distance should I use for auto focus calibration?
Best Regards,
Dave
Why would FoCal misread the lens I’m trying to calibrate. It sees my Tamron 70-300mm VC as a Tamron SP 35mm
Hi Wayne,
Thanks for your feedback, we’ve made a note and have updated our lens definitions for the next release.
As background, each lens provides identifying information to the camera that is passed to FoCal. FoCal makes a ‘best effort lookup’ to understand the connected lens and display lens information to the user.
Lens information changes (especially new lenses) and we do our best to keep up! Each time there’s a new FoCal release we update the internal lens database.
The important thing to note, even if wrong lens information is shown it does not change the calibration result. FoCal does not take into account the lens name or even it’s focal length during the calibration process.
Best Regards,
Dave
The problem I encounter is with 2 tamron lenses. the 24-70 and the 70-200, both are listed as tamron 24-70. the problem is that the 24-70 might by needed to be +10 and the 70-200 might be -10. So Nikon now just see 24-70 lens on the camera set it to +10, but matter of fact the 70-200 is on the camera.
Soooooo frustrated with this scenario.
Hi Johan,
Yep, it’s an issue we’ve heard of before. It doesn’t come up often as it is just certain Tamron lenses which have the issue and often users may only have one Tamron lens (can’t get confused since there’s only one ever connected to the camera!).
Best bet is to contact Tamron, I haven’t checked recently but I did hear some of the lenses could be ‘fixed’ with an update, worth checking and see what they might be able to do 🙂
Best,
Dave
Hi, I have just tested my Sigma EX 70 – 200mm F2.8 DG HSM lens. On the 70mm length it comes up with -5 on the micro adjustment but on the 200mm end it comes up with -25!! Beyond the capability of the camera to adjust to this amount. Do I need to send the lens in for repair or is there something wrong here?
Thanks
Hi,
Sorry blog comments are not checked very often(!)
If you’re still seeing the issue and you didn’t already, could you raise a support request
Your question is exactly the sort of thing the support team is ready and able to help with, give them a shout! 🙂
Best,
Dave