Canon is dead to me after they forced 3rd parties to stop all RF lenses. I bought the last AF lens by Viltrox and it's fantastic. I regret getting an RF body after they threatened them, and now only allow 3rd parties on their crop sensor bodies. Fuji, Sony, Leica, Nikon, all allow 3rd party lenses. I switched to Fuji and won't go back.
I'm feeling a little bit of buyer's remorse for that same reason. I love my Canon, but having a closed loop lens system is not ideal and is insanely pricey. All of the "adapters" that allow you to use 3rd party lenses suck.
It is a little sad honestly how Canon has now been surpassed by Sony in both sensor technology and lens optical quality and yet they double down on forcing people to use their expensive lenses. I miss the old days when EF mount was the lingua franca of autofocus mounts with a thriving ecosystem of lenses from Sigma, Tamron, and the like. Even now some of the Sigma Art lenses for SLR mounts, like the 40mm and 105mm f/1.4, are peerless in terms of image quality.
Also not pleased that third party lens vendors were excluded. However the way I look at this is perhaps we would have no Canon cameras at all if the revenue they derive from lenses had not been captured.
Wait, did they reverse their stance on this again? I thought they had finally started to open the mount up to 3rd parties. Really disappointing to hear that this is not actually the case.
I'm in the same boat. I wouldn't be as upset if the Canon range was uniformly excellent. But every one of their RF ultra-wide zooms suffers from softness or extreme vignetting. I sent back a $2200 RF 15-35mm, and for landscapes I now just use my iPhone Pro.
They weren't dead to me immediately just because there are still third party RF lenses that rock, but when Sirui did the AF anamorphic lenses and I couldn't get one for my Canon then it really bummed me out. :/
Nikon aren't doing much better on letting third-party lenses into the ecosystem, sadly - speaking as someone with three Nikon DSLRs and Sigma glass I'm very happy with.
Wait, what? That’s horrific! I love my range of lenses for my long-suffering 5d3. I wouldn’t have bought it if the only options were the canon lenses. How disappointing to hear.
I used to have to manually focus to take photos of my dog. The apparent noise from the early Ring USM AF in the 85mm/1.8 lens drove him mad and he would run out of the room.
We did initially wonder if it was some psychological effect of pointing a camera at him, but one night I hid behind a curtain and used the AF motor and he still shot out of the room.
Subsequent USM AF motors were silent to him and he was then content to have his photo taken.
My dog also developed an aversion to cameras, and I wondered if it was the USM focusing. That might have been what started it, but she still doesn't like any cameras at all, USM or not.
In case anyone else was looking for it, this article from the same author covers the more algorithm-y question of how a camera body decides to actually use these motors when you press the "focus" button: https://exclusivearchitecture.com/03-technical-articles-DSLR...
I can't find any articles on how a camera body decides what is and isn't a desired subject. I'm guessing there's some amount of machine learning-type stuff involved in that, seeing as how they can detect human (and bird?) faces?
Nowadays they use pre-trained pattern recognition AI models, yes, which has become much more impressive (and CPU-intensive) with mirrorless cameras where the entire resolution of the main image sensor is available for analyzing the scene. Some higher-end traditional DSLRs have a "high"-resolution (around 0.1 MPix or so) metering sensor that is used to assist the AF system (eg. what Canon calls iSA and iTR [1]).
Traditionally, cameras would just focus using the single focus point the photographer has selected, or if they have selected a larger area focusing mode, the camera would typically pick the closest point of a group of points, assuming that that's usually what the photographer is interested in. (Remember that traditional (D)SLRs have a discrete AF sensor with at most a few dozen focusing points to choose from!)
In tracking AF modes (eg. Canon's Servo AF), depending on settings, the camera tries to avoid sudden shifts in focus even if a foreground object momentarily occludes the original target. Tracking AF also has to predict the subject's motion to prevent the focus from lagging behind a fast-moving subject. Higher-end cameras allow configuring the AF behavior in terms of how reactive vs "sticky" it should be when tracking a subject, and how linear the subject's motion is expected to be.
With mirrorless cameras the focus switched from specialized sensors to on-CMOS contiuous exposure sensors, so movement is easy to detect. At this point the cameras have specialized AI hardware to run the models, and they also accept user input (on R5 MkII you can register up to ten people to prioritize focus on[1]). The focusing options are now very complex[2][3], and combined with lots of customization options on the camera's buttons you can have very specialized/personalized setups for different types of photography.
Sure, as I said in the first paragraph, AF is these days very impressive thanks to the large amount of data available (but of course this would have been too much data back in the day, when there wasn't nearly enough CPU power to process it fast enough). I wanted to give more historical context for how AF worked before fancy AI.
The AF settings, except those related to face/object recognition, haven't actually changed that much since the 7D Mk II days. The preset system is more general now and allows you to store and recall all AF settings rather than just the three tracking-related variables. The high-end DSLRs used to have six cases for different types of sports that you could modify but not rename.
Sony A9 III even has a configuration setting of whether it should focus on the left or right eye of the person :) It also can remember faces and prioritize them if there other faces. Let's say someone shoots their kid on the football field and wants only them to be in focus.
The left eye/right eye option has been around even on lower end cameras for the better part of a decade, which is kinda wild. My fuji from 2017 has that, and Fuji are specifically known for having worse autofocus performance than Canon or Sony or Nikon. Nikon and Sony specifically seem to be top of the pile right now.
The Z8 has a whole separate processor dedicated to autofocus and the viewfinder which, in practice, means it can shoot 20 FPS full quality 45 megapixel RAW files with continuous 120fps autofocus without blacking out the viewfinder for each shot, which is absolutely insane.
I have been aggravated by (and bellyaching about) the ridiculous lack of a way to control the focusing motors in these lenses directly when shooting video, using a follow-focus wheel.
Video shooters are still strapping janky gear-tooth strips onto these lenses, and then bolting bulky mechanical follow-focus mechanisms on the OUTSIDE of lenses that have focusing motors already built in.
I can only imagine that manufacturers refuse to make direct control available in order to protect their "cinema" lens lines, where a manual normal lens sells for thousands of dollars. And yet they sell some falsely-named "hybrid" lenses that are supposedly somehow better for video... despite lacking even geared focusing rings that are compatible with follow-focus units... let alone a control port that could be used for a focusing wheel to use the autofocus motors.
I looked at the Canon camera-control SDK, and sure enough... focus control is omitted from their entire line, except for two PTZ cameras that aren't suitable for cine use.
A side project I've been neglecting is controlling autofocus on Sony Alpha cameras (specifically the A7ii and A7iii which I have) for the purpose of focus-stacking: shooting a few dozen shots with minimal-step focal differences and blending them all in photoshop later.
It took me a while to find the API; Sony made one then pretty much scrubbed it from the internet and has ignored my emails asking for one :-)
But even with the API, achieving results are tricky. Whether or not the focus motor steps properly seems to be based on how busy the camera is at the time... I'm trying various combinations of [shoot] - [sleep] - [step the focus] - [sleep] - [shoot] but the exactly which of these steps actually succeeds just seems so damn random.
BlackMagic does some pretty innovative stuff. They offered an Arduino add-on board that could communicate with and control their cameras over SDI, including access to the Canon EF mount and its commands.
Somebody did in fact create a control board with a knob to control lens focus through this thing. It can be done. The manufacturers just refuse to do it.
I've always wanted to know how the various autofocus systems worked. This page is incredible. I wish there was another one with the Nikon autofocus systems since that's what I actually have.
I still have several lenses with autofocus that don't have an AF motor in it at all, the motor is in the camera body instead there's a tiny screw on the lens mount that transfers the motor rotation to the autofocus parts in the lense. This was very slow and noisy though on my cameras.
Whereas I think "no metering" is never a problem on Canon.
Over the decades as a DSLR shooter (on Canon), I even saw Nikon shoot themselves in the foot not once but twice.
* When Nikon introduced "E" lenses (electronic aperture, like Canon and all modern designs), very few bodies were compatible with it. They didn't have the foresight to introduce compatible bodies before any lenses with E were released. For example, this was released in 2008 ( https://www.kenrockwell.com/nikon/24mm-pc.htm ), and the earliest compatible body is from 2007. It doesn't work with tons of crop (DX) bodies like the D90, as well as any film camera.
* When Nikon introduced "AF-P" (stepper motor) lenses in 2016, there were no compatible bodies before 2013. Again, not enough future-proofing.
Meanwhile, Canon took a different approach. They pissed off users twice - when transition from the FL mount (1964) to FD (1971), and FD to EF (1987). They basically got it all right with EF - fully electronic, no aperture slider, no focus screw, big diameter.
I have even successfully used a year-2017 Canon lens with Ring USM AF and IS on an original EOS 650 film body (released in 1987), and both AF and IS work perfectly. Mind you, both of those features did not even exist at the birth of EOS. Presumably the AF electronic protocol is agnostic of what motor technology is in the lens, and IS can be a lens-only thing without the body knowing.
Nikon's compatibility gotchas don't hold a candle to Canon's excellent (albeit imperfect) compatibility. Nikon keeps making the same mistakes over and over again, such as not having a screw-drive motor in the F to Z mount adapter; the company really seems to hate people with old lenses and bodies. Oh and this is not new either; Nikon's F to 1 adapter (does anyone remember that highly cropped mirrorless camera?) also lacked certain features.
This third party company announced a Nikon F to Z lens adapter that will autofocus with older screwdriver autofocus lenses. They already had a Nikon F to Sony adapter and now they have a Nikon F to Nikon Z adapter.
What annoys me is that the AF-P could probably be supported on a lot more cameras, they did add it to the D3300 with a firmware update. They just didn't do it for more cameras...
Newer Nikon lenses have a few motor types but if I'm not mistaken they support old mechanical (body driven) aperture and focus drive as well (the D6 and somewhat recently F6 film camera were still for sale)
Nikon has a complex ecosystem keeping generally the same mount for many decades and general support for it, where others developed a specific autofocus mount when they moved towards those systems.
Their mirrorless Z mount and lenses don‘t have mechanical drive. Neither do their official F-to-Z adapters - much to the chagrin of people using their older glass.
Yeah the in-body motor thing is only for the older Nikon lenses, I don't think AF for that is still supported if you're using Z-mount adapters. Nikon seems to have a few similar AF motor technologies like the canon ones in this article.
They ended up with all electronic contacts when they switched lens mounts for the “eos” series cameras though at the expense of not having backward compatable lenses. (I had a bunch of manual focus lenses at the time…) They had one “auto focus” camera for the old lenses the t80. (https://global.canon/en/c-museum/product/film115.html) but almost no lenses supported autofocus.
I think most of settings the camera sent to the older lenses (aperture primarily) was done mechanically.
The way these old cameras autofocused was pretty interesting. With some light passing through a semi transparent mirror onto different sensors.
Edit: this site has another page about autofocus. Quite extensive.
I have one of those mirrorless slrs now. The focus is amazingly good. Lots of software (eye detect etc). The focus sensors are integrated into the image sensor I believe.
All my nikon cameras focus with a semi transparent mirror on a specialized autofocus sensor. I even have an nikon film camera that has that cicuitry. You can't autofocus like mirrorless cameras with PDAF pixels in the main sensor with a mirror in the way.
Cool videos, thanks! The driver shown in the first video is indeed a piezoelectric actuator, but not the type Canon uses in their lenses. The driver shown in the second video is probably related to Canon's Nano USM.
As far as I know, Canon controls a large part of the supply chain and has numerous factories to manufacture quality-critical parts (especially their lenses, zoom mechanisms, and barrels) themselves. But there are numerous components that are more efficient to obtain from third parties, such as microchips. I am not sure about the motors and actuators, though.
A note of possible interest for this crowd: all these motor types are open loop, so you can't actually command a specific state with any accuracy / repeatability. In other words, if you send a signal to the lens to focus at, say, 3m, it won't necessarily be in exactly the same state as the next time you focus at 3m.
This is why camera calibration can be tricky, and you often don't want to touch the cameras after you've done a calibration pass.
It’s open loop in that the measurement of it being in focus is reliant on the subject matter, and a different measurement.
I’m not making this up, camera manufacturers have told me to my face that focus is open loop, period. They can’t guarantee repeatable focus.
Notably the measurement isn’t of the state of the motor/gearing. Furthermore, being “in focus” means the subject matter’s out of focus blur is below some threshold; there is a range of focus states that qualifies — but those seemingly small differences can affect camera calibration, with >pixel-level differences in effective focal lengths.
I can back you up on that. I work on a highly custom system that uses EF lenses through an EF-to-serial adapter. The lenses don't even know what position they're at when you power them on. If you want to move them to a specific focus step you have to go through a homing procedure that drives the focus motor to a stop (I don't recall if it's the near or far stop).
Lens-to-lens it's definitely not consistent. Step 142 on two different lenses will have a different focal depth. We calibrate each lens ourselves (thankfully you can read the serial number through the EF mount) and then still have to do closed-loop image-based sharpness estimation to guarantee that things are as good as we can get.
How could they then tell to what distance they're focused?
I used to have two Ultrasonic lenses, the 17-40/4L and the 17-55/2.8. They both had distance scales which would move around as the lens focused.
My current Olympus lenses have a focus-by-wire manual mode, with a distance scale on the barrel. The camera also reports the focus distance in the EXIF. Are these just vague ballparks?
Sorry for the late reply. What we’ve found is that they’re all close but not identical. Our system has a point LIDAR pointed at the ground and then uses that to drive the focal distance for the lens (as well as input for some other sensor fusion stuff). We generally run the lenses wide open (f/2.0 or f/2.8). Mapping step counts to distance would get us close, but unless we did calibration for each individual lens it wasn’t close enough to get the focus tack sharp. When you’re looking at the focal distance in your EXIF metadata you’re probably not validating whether that 2.4m distance to the subject was actually 2.3m and for the vast majority of cases the absolute accuracy of it doesn’t matter. Same with the distance markers on the focus ring… you’re not setting the focus within a few cm based on those marks, just getting it close enough that you can do closed-loop focusing with an eyeball in the loop.
Once we did the per-lens calibration curve we didn’t have to do it again. I’m assuming the small differences are just manufacturing tolerance and they’re relying on the focus/phase sensors to get it the rest of the way.
Some - not all - of the EF lenses have a distance encoder as well. It's fairly approximate; I think it exists mostly for the benefit of the flash system, which needs a rough starting point for the distance in some of the open-loop modes.
For the current Olympus ones, I think there's a broadly similar encoder on the ones with proper manual focus scales, and the pure fly-by-wire ones reset focus at connection so the camera can work it out.
Correct. The AF sensor continually takes more samples even while the motor is moving. You can tell because sometimes the motor overshoots and then you can see it come back.
Yup… but it’s not necessarily in the same state if you do it twice. It’s close, but not identical. Even if the image metadata claims the focus point is identical, the lens in all likelihood is not in the same state and will have some deviation in its intrinsics.
I have an EF 400mm f2.8 (2nd gen) lens. This thing focuses significantly faster with my old 7Dmk2 than my very new R6Mk2. With all other lenses it’s the opposite. Why does this happen?
I have been told that this might be due to the old 7D and 1D bodies sending more power to the AF motors, but haven’t seen this corroborated anywhere.
Thanks for your feedback. Until now I have never blogged about the creation of these illustrations, but it's a nice idea, there is tons of "behind-the-scenes" material. With regards to the viewfinder image, the actual path of light was drawn with the help of a Japanese optical engineer who was able to provide professional raytracing simulations.
turnsout Thank you very much. Trying to understand complex things has always fascinated me. Indeed it takes a lot of time (rd. 1.600 hours for the complete series of Canon lens chapters) but its people like you who drive me forward. I don't have a donation page yet, I never thought it would be used :-)
Amazing website. If I didn't know better, I thought this content came from Canon itself. A couple of comments from me:
> The following comparison chart shows all of the autofocus drive types that have ever been used in Canon's EF, EF-S, RF, and RF-S lenses.
They missed mentioning EF-M, Canon's foray into mirrorless interchangeable-lens cameras.
> Although the ultrasonic noise cannot be perceived by the human ear, it can be picked up by sensitive microphones which can be problematic during video shootings.
I used several Ring USM lenses and can hear the noise when focusing. It's a soft swishy kind of white noise, not harmonic. The Ring USM definitely generates noise in the audible range, not only ultrasonic.
> The latest types of autofocus drives, including the Stepper Motor, Nano USM, and Voice Coil Motor, offer focus-by-wire.
Though I have nothing against focus-by-wire on principle, the implementations have not been good. Having used several EF-M (all STM) and RF (some STM and some Nano USM) lenses that have focus-by-wire, I find that the motor actuation significantly lags behind the user turning the focus ring (which is a digital encoder), and the granularity of the focus steps are visible.
Meanwhile, the old EF Ring USM's full-time manual (FTM) was really good by comparison; it didn't take much force, was fully responsive with no lag, and was fully analog when turned by hand and had no discrete steps. I miss that, as new lenses don't use this design.
Why? Because f is actually a lowercase italic variable that denotes the focal length (15 mm in this case), and "/" (slash) really means division. The size of the aperture is (15 mm / 2.8) = 5.4 mm; that is the real diameter of the hole that light passes through. The absolute aperture is 5.4 mm and the relative aperture is f/2.8. https://en.wikipedia.org/wiki/F-number
Thanks nayuki for the detailed comment and positive feedback.
Thanks for pointing out the missing EF-M, I am going to add that during the next polishing round. With regards to USM noise, the ultrasonic whine can be picked up by very good ears. I think the Nano USM is excited in the range of 60+ kHz, so at least that type of USM motor is absolutely inaudible.
Syntax-wise with F2.8, you got me. I know it's not 100% correct, but it was a deliberate choice that I made. Historically, there is either 1:2.8 or f/2.8 which will always be correct representments of the aperture (exit pupil) diameter. However, pursuing a cleaner look, I avoided the division or slash characters because the relation between the focal length and the f-number is known by every photographer. This is why Canon has also made the transition to the syntax "F2.8" in all their RF and RF-S-lenses - at least their product names and labels printed on the lens barrels. (The same is true with the focal length where Canon doesn't print "mm" on the barrel - to achieve a cleaner less mathematical look). So I believe what I did here was just go with the trend :-)
> the relation between the focal length and the f-number is known by every photographer
I'm going to disagree with this. I think very few photographers understand what an absolute aperture size is. Here's a good test: If you take a "constant (relative) aperture" zoom lens like a 24-70mm f/2.8, what happens if you keep the aperture "constant" at f/2.8 and zoom the lens in and out? I think the answer will surprise many people.
I think most people don't even know that the f (or incorrectly "F") refers to the focal length; I think the vast majority of photography education just treats "F" or "f/" as some piece of fixed notation or incantation without explaining what it means. (Kind of like if I asked you, what does the "degree" in "degree Celsius" mean?)
> Canon has also made the transition to the syntax "F2.8" in all their RF and RF-S-lenses - at least their product names and labels printed on the lens barrels
Oh wow, I didn't notice this one. Your observation is correct; Canon did change their official naming of RF lenses to use the "F2.8" syntax instead of the previous "f/2.8". Oh well, I still disagree with it because it is bad mathematical syntax and spits in the face of tradition.
> The same is true with the focal length where Canon doesn't print "mm" on the barrel - to achieve a cleaner less mathematical look
True, and I have subconsciously noticed this.
> So I believe what I did here was just go with the trend :-)
Please don't change the naming of old lenses. Also, I would prefer new lenses to be hammered back into the old naming scheme, but that's more open to debate. Curiously, the camera shop Vistek sometimes uses old naming for RF lenses: https://www.vistek.ca/store/434924/canon-rf-1535mm-f28l-usm-... "Canon RF 15-35mm f/2.8L IS USM Lens"
I have checked the video you've linked. My suggestion would be that this whooshing noise is not generated by the piezoelectric part but rather from the connected focusing mechanism (especially the helical focusing barrel which turns, and the focusing lens cell which slides back and forth. These parts sit inside the stationary lens barrel, and so a bit of friction is unavoidable. Just my guess.)
Syntax-wise, I will have to think about it. As with all my other articles, I like being very precise with physical formulas and mathematical expressions. But here is my point: These values in photography are not pinpoint precise anyway. When optical engineers come up with lens designs, the precisely calculated values for the aperture are something like f/4.63 which is then simply rounded to the next possible f-value specified on the lens barrel, and calculated focal length is often 25-49mm and Canon simply names that lens 24-50mm on the barrel (other manufacturers do that in a similar way). So, this is just my personal opinion, but it appears to me that in the context of photography, these values are just halfway decent guide values, almost like approximate descriptions of the lens behaviour. For that reason, I do understand why lens manufacturers move away from the precise syntax, and treat these values more like an abstract product description. Of course, if used in a strictly scientific context, I would rather prefer to use f/4.5. But I might be pretty alone with this view.
PS: I have planned a new chapter explaining all this, including the interesting fact that you've mentioned, where the diameter of the exit pupil actually changes according to the focal length, even with a constant aperture lens. But it will probably be end of summer when this is going online.
> this whooshing noise is not generated by the piezoelectric part but rather from the connected focusing mechanism
Indeed. I never implied that the piezoelectric part makes an audible whooshing noise. I only meant to say that the entire Ring USM system makes that noise, and so it will affect audio recorded on the camera's on-board microphone. Either way, that noise explains the motivation for the development of STM AF and I guess Nano USM.
Agreed with everything else you said. Thanks again for your fantastic articles and replies!
It’s an amazing article. And it’s only the tip of the iceberg - there are many more. This has to be the work of a madman (meant in the best possible way).
Oh neat! I have a Newscale micro motor demo kit from years ago, and I wondered if they were ever successful in the market. But that hip-gyrating Micro USM action sure looks familiar.
I always wanted to build a controller for Canon Telephoto lenses to use them with c-mount cameras and control the focus from a PC. Might be helpful for that.
A couple of universities built a telescope called Dragonfly, that utilizes 48 Canon 400mm lenses to photograph the skies of New Mexico. The kicker is they don't use camera bodies. Instead, they designed and built mounts that directly connect the lens to a sensor. A PC controls each lens through some custom electronics using commands they reverse-engineered from the Canon control software. These commands control focus, aperture, and triggering a photograph. Images from the 48 lenses are then combined in the computer into a highly detailed final image. This was done back in 2013 so its possible Canon has released an API by now but I suspect they still keep it proprietary.
I visited some similar telescope in Australia at https://www.sidingspringobservatory.com.au/ run by Macquarie University ... the students explained the lens system could not be too new and could not be too old. Basically it sounded like someone had partly reverse engineered one generation of the lens interface only.
I guess the camera and lens manufacturer wants their lenses to be used with their cameras and to have better results than other manufacturers, whereas random companies want to clone the interface and sell cheaper lenses that also work with the cameras. Realising how awesome the lenses are, the students want an array of them, however they don't have the budget to buy a similar array of top end manufacturers' cameras.
A number of machine vision camera vendors have higher-end variants that support active EF-mount lenses. They tend to be quite expensive models with larger sensor, since you tend to lose a lot adapting full-frame lenses to tiny C-mount sensors.
If you want a computer-controlled lens, one of the cheapest options is to use Blackmagic cameras that support SDI and/or USB control. You can then control either native M4/3 active lenses (focus, aperture, zoom) or supported EF-mount lenses if you use one of the EF->M4/3 active adapters.
I was a big fan of USM lenses when I had a Canon, but I had one go bad with fungus or something and for a while had only a 35mm prime which I even used to take pictures of birds. Then I lost my Canon and decided to get a Sony circa 2019 or so because all the reviews I saw for Nikon and Canon said the autofocus sucked on entry-level full frame cameras.
I’ve recently been shooting film with an old EOS camera from the 90s I bought used and it was really nice being able to use the EOS lenses I bought for my DSLR in the 2000s and 2010s. It’s a dying standard now but it’s really impressive it lasted as long as it has, with significant technological innovation on both sides of the lens mount while retaining full compatibility. A brand new EOS EF lens still works with an 80s camera and a new 80D from 2017 can still use the lenses from the 1980s without any adapter. 30 years ain’t bad for a standard!
I must be old -- I vividly remember the Nikon crowd crowing loudly online forums (fredmiranda.com I'm looking at you) about how Canon broke backwards compatibility when they moved from FD to EF. Whereas you could slap an F-mount lens from the 1950s on any Nikon DSLR ever made, no problemo. (Remarkably, this continues to be true!)
Except if you mount an old, so called pre AI lens, on certain Nikons, you will have to disassemble the camera to get it off. Many modern Nikons can't autofocus older AF lenses due to not having a motor. And many AF film SLRs couldnt meter with manual lenses. So far from perfect compatibility.
Nikon F mount was introduced in 1959. The oldest mount used today is probably Leica M from 1954
Oh yeah, that’s even better, totally agree, but it doesn’t negate my point. I don’t think Canon’s 30 or 40 years would be matched by a hypothetical present-day camera upstart, let alone Nikon’s 70 years.
As a Canon owner, Nikon has a much longer back compatibility range. Having autofocus motors and IS in body rather than in lens seems to be a part of their trick.
The traditional Nikon mount has a small screw that is turned. The camera autofocus speed is limited by the amount of torque that can be applied to the screw - which can make focusing some of the heavier lenses slower.
The "AF-D" lenses have contacts back to the camera body that communicate distance information (that is in turn used by the camera body to calculate flash power).
The G mount lenses remove the manual coupling for the f/stop which means that only bodies that can control the aperture from the body can use them. My FM3A has no aperture control on the body and so with that camera, I unlock the aperture ring.
The AF-S camera lenses have the focusing motor in the body.
The impressive part of Nikon's compatibility isn't only the "you can use an AI-S lens on a modern (professional) body, but also "you can use any of the F mount lenses on an old body" (the G lenses don't have the f/stop ring and the E lenses have the focus motor in the lens).
While it appears that Nikon has mostly shifted to E and G mount, third party lenses are still being manufactured for the F mount.
Difference between G and E-type lenses is that 'E' stands for Electromagnetic Diaphrapham.
The G type lenses have an aperture tab for diaphrapham control as dictated by the camera body, the E-type lenses leave this to being controlled by the camera electronically.
> you can't jiggle the film around to keep it in the same place.
Contax had an AF solution (Contax AX) that actually moved the entire film plane to focus. I suppose that the same thing could be done in other axises, but I also suppose that there is a reason that only one manufacturer tried it.
> Optical Image Stabilisation is effective with movement across a range of frequencies, so it can cope not only with simple camera shake (0.5Hz to 3Hz), but also with the engine vibrations encountered when shooting from a moving vehicle or helicopter (10Hz to 20Hz).
I had this problem in Chrome. I turned off JavaScript and the page worked nicely. (Props for that!) I'm assuming my adblocker or auto-accept-cookies plugin did something weird.
Canon's lenses are great, but they're really expensive. The camara body itself has the same issue.
Sony seems to be the first choice for indie filmmakers and youtubers right now.
L-mount and M43 also seem to have great potential as Panasonic supports Phase Detection Autofocus.
edit:
There is no need to be cynical here. Market figures are cruel. I hope they are all good. Competition is the best. I should also mention the return of Nikon. The patent threshold of RAW such as RED is what we need to oppose.
As the old saying goes, buy the glass, rent the body. Camera bodies change much more quickly than the lenses, so as an indie your money is better spent on the glass. However, shit is so cheap now that you can almost upgrade your gear like your phone. IF you’re a cut ally making money with it, you can pay off your gear in one or two gigs. Unless you’re the “nephew with a 5D asking $500” to shoot that project
Most shops that offered still photo rentals in my city have gone out of business, and it's a major city. I used to be able to count on renting a big telephoto that I didn't need to own, for a few hundred per weekend, but these days I'd have to take a long drive to get to the one remaining shop.
Yeah, online rental places have pretty much made the locals fade away. The one in my local area requires a "deposit" worth the full sales price which makes it out of the reach of anyone but those well off.
We used to joke about the B&H rental by taking full advantage of the 30-day return policy. Even Amazon's return policy qualifies now. You just need to the up front funds available and hope that wear&tear is not noticeable to make the return acceptable. It's hard for local shops to keep up with that.
I like how the messenger is getting shot. It's as if my reporting of what actually happens is being interpreted as that's what I do. Just because I know of how the system gets played does not make me a player.
I get upset when people defraud my neighbors and friends in my community. that's what a local shop is. Just today i went to my local pharmacy to look for l-theanine based on some commentary on HN yesterday. I could have gone on amazon, or to walmart, or to CVS. People interfering with that business aren't cool, regardless of the motivation.
let me put it this way - i never even thought of using amazon as a rental service. even though amazon loves me, and i could get away with it, my brain doesn't spend time thinking about unethical things to do.
read this as: "Someone should have gotten upset before the deposit went up from jerks either breaking or stealing the rental equipment" - as my sibling says, this is why we can't have nice things.
I chose early on to invest in Canon's L lenses even though I'm an amateur. Yes, they are considerably more expensive than consumer lenses but for the price, you get a more rugged, weatherproof body, sharper and faster glass, and a greater selection of both fixed and zoom lens. The down side is they are quite heavy making them more difficult to carry when shooting outdoors and they can require a stronger tripod to support the weight of the lens and camera. But, the images they can produce can be simply stunning at times and that, after all, is why we do photography.
Considering my 70-200 can still hang after 10 years and will last another 20 easily with some TLC I was fine with my $1500 purchase - which I made as an indie filmmaker!
Jokes aside Sony glass really isn’t all that much cheaper.
Canon is dead to me after they forced 3rd parties to stop all RF lenses. I bought the last AF lens by Viltrox and it's fantastic. I regret getting an RF body after they threatened them, and now only allow 3rd parties on their crop sensor bodies. Fuji, Sony, Leica, Nikon, all allow 3rd party lenses. I switched to Fuji and won't go back.
I'm feeling a little bit of buyer's remorse for that same reason. I love my Canon, but having a closed loop lens system is not ideal and is insanely pricey. All of the "adapters" that allow you to use 3rd party lenses suck.
It is a little sad honestly how Canon has now been surpassed by Sony in both sensor technology and lens optical quality and yet they double down on forcing people to use their expensive lenses. I miss the old days when EF mount was the lingua franca of autofocus mounts with a thriving ecosystem of lenses from Sigma, Tamron, and the like. Even now some of the Sigma Art lenses for SLR mounts, like the 40mm and 105mm f/1.4, are peerless in terms of image quality.
Yes, Sigma ART lenses are insane for that price. Also, they look very elegant compare to plastic Canon lenses.
I was hoping for the new Sigma BF camera, simple and elegant. Unfortunately no Bluetooth, which is a deal breaker for me.
Well, they had some very bad financial years behind them. They will let third parties, but it will take time to recuperate losses.
Also not pleased that third party lens vendors were excluded. However the way I look at this is perhaps we would have no Canon cameras at all if the revenue they derive from lenses had not been captured.
Wait, did they reverse their stance on this again? I thought they had finally started to open the mount up to 3rd parties. Really disappointing to hear that this is not actually the case.
Only for their crop bodies, not full frame. Which basically means it's still locked.
I'm in the same boat. I wouldn't be as upset if the Canon range was uniformly excellent. But every one of their RF ultra-wide zooms suffers from softness or extreme vignetting. I sent back a $2200 RF 15-35mm, and for landscapes I now just use my iPhone Pro.
They weren't dead to me immediately just because there are still third party RF lenses that rock, but when Sirui did the AF anamorphic lenses and I couldn't get one for my Canon then it really bummed me out. :/
Exactly why I just bought a Nikon.
Nikon aren't doing much better on letting third-party lenses into the ecosystem, sadly - speaking as someone with three Nikon DSLRs and Sigma glass I'm very happy with.
Wait, what? That’s horrific! I love my range of lenses for my long-suffering 5d3. I wouldn’t have bought it if the only options were the canon lenses. How disappointing to hear.
I used to have to manually focus to take photos of my dog. The apparent noise from the early Ring USM AF in the 85mm/1.8 lens drove him mad and he would run out of the room.
We did initially wonder if it was some psychological effect of pointing a camera at him, but one night I hid behind a curtain and used the AF motor and he still shot out of the room.
Subsequent USM AF motors were silent to him and he was then content to have his photo taken.
My dog also developed an aversion to cameras, and I wondered if it was the USM focusing. That might have been what started it, but she still doesn't like any cameras at all, USM or not.
Maybe the mechanical/EFCS shutter?
Some animals just know and some like being the center of such attention. Others just don’t.
Such a cool website.
In case anyone else was looking for it, this article from the same author covers the more algorithm-y question of how a camera body decides to actually use these motors when you press the "focus" button: https://exclusivearchitecture.com/03-technical-articles-DSLR...
I can't find any articles on how a camera body decides what is and isn't a desired subject. I'm guessing there's some amount of machine learning-type stuff involved in that, seeing as how they can detect human (and bird?) faces?
Nowadays they use pre-trained pattern recognition AI models, yes, which has become much more impressive (and CPU-intensive) with mirrorless cameras where the entire resolution of the main image sensor is available for analyzing the scene. Some higher-end traditional DSLRs have a "high"-resolution (around 0.1 MPix or so) metering sensor that is used to assist the AF system (eg. what Canon calls iSA and iTR [1]).
Traditionally, cameras would just focus using the single focus point the photographer has selected, or if they have selected a larger area focusing mode, the camera would typically pick the closest point of a group of points, assuming that that's usually what the photographer is interested in. (Remember that traditional (D)SLRs have a discrete AF sensor with at most a few dozen focusing points to choose from!)
In tracking AF modes (eg. Canon's Servo AF), depending on settings, the camera tries to avoid sudden shifts in focus even if a foreground object momentarily occludes the original target. Tracking AF also has to predict the subject's motion to prevent the focus from lagging behind a fast-moving subject. Higher-end cameras allow configuring the AF behavior in terms of how reactive vs "sticky" it should be when tracking a subject, and how linear the subject's motion is expected to be.
[1] https://www.canon.com.hk/cpx/en/technical/pa_Overview_of_65-...
With mirrorless cameras the focus switched from specialized sensors to on-CMOS contiuous exposure sensors, so movement is easy to detect. At this point the cameras have specialized AI hardware to run the models, and they also accept user input (on R5 MkII you can register up to ten people to prioritize focus on[1]). The focusing options are now very complex[2][3], and combined with lots of customization options on the camera's buttons you can have very specialized/personalized setups for different types of photography.
[1] https://cam.start.canon/en/C017/manual/html/UG-04_AF-Drive_0... [2] https://cam.start.canon/en/C017/manual/html/UG-04_AF-Drive_0... [3] https://cam.start.canon/en/C017/manual/html/UG-04_AF-Drive_0...
Sure, as I said in the first paragraph, AF is these days very impressive thanks to the large amount of data available (but of course this would have been too much data back in the day, when there wasn't nearly enough CPU power to process it fast enough). I wanted to give more historical context for how AF worked before fancy AI.
The AF settings, except those related to face/object recognition, haven't actually changed that much since the 7D Mk II days. The preset system is more general now and allows you to store and recall all AF settings rather than just the three tracking-related variables. The high-end DSLRs used to have six cases for different types of sports that you could modify but not rename.
Sony A9 III even has a configuration setting of whether it should focus on the left or right eye of the person :) It also can remember faces and prioritize them if there other faces. Let's say someone shoots their kid on the football field and wants only them to be in focus.
The left eye/right eye option has been around even on lower end cameras for the better part of a decade, which is kinda wild. My fuji from 2017 has that, and Fuji are specifically known for having worse autofocus performance than Canon or Sony or Nikon. Nikon and Sony specifically seem to be top of the pile right now.
The Z8 has a whole separate processor dedicated to autofocus and the viewfinder which, in practice, means it can shoot 20 FPS full quality 45 megapixel RAW files with continuous 120fps autofocus without blacking out the viewfinder for each shot, which is absolutely insane.
The a9 iii shoots 120FPS 24.6MP raw files with continuous autofocus and without blacking the viewfinder. Top-end cameras are insane.
I sort of miss my Canon EOS 1 at 10 fps (film) when shooting sports. "If you saw it, you missed it."
The $$$ per second of shooting that is hilarious. I guess that’s what shooting IMAX feels like, but only for like 3 and a half seconds at a time.
The technical term for the "stickiness" you're referring to, where a system is resistant to change or has a memory, is hysteresis.
Similarly, promo materials for the Elan 7 talked about how the camera was able to determine exposure based on a database of hundreds of photos.
I have no idea how this worked, but would have loved to see the photos they used for “training” this system 25 years ago.
As a Canon user since the mid-80's, I found this fascinating reading.
Edit: Wow - there's a whole collection of Canon lens technology articles there: https://exclusivearchitecture.com/03-technical-articles-CLT-...
Thank you for your feedback! And great that you've found the other chapters - most of them are already complete. Enjoy!
Well, there goes the rest of today…
Really interesting work.
I have been aggravated by (and bellyaching about) the ridiculous lack of a way to control the focusing motors in these lenses directly when shooting video, using a follow-focus wheel.
Video shooters are still strapping janky gear-tooth strips onto these lenses, and then bolting bulky mechanical follow-focus mechanisms on the OUTSIDE of lenses that have focusing motors already built in.
I can only imagine that manufacturers refuse to make direct control available in order to protect their "cinema" lens lines, where a manual normal lens sells for thousands of dollars. And yet they sell some falsely-named "hybrid" lenses that are supposedly somehow better for video... despite lacking even geared focusing rings that are compatible with follow-focus units... let alone a control port that could be used for a focusing wheel to use the autofocus motors.
I looked at the Canon camera-control SDK, and sure enough... focus control is omitted from their entire line, except for two PTZ cameras that aren't suitable for cine use.
Nikon has just released a powered zoom lens. So hopefully there will be more of these in the future.
https://www.nikon.com/company/news/2025/0213_imaging_02.html
A side project I've been neglecting is controlling autofocus on Sony Alpha cameras (specifically the A7ii and A7iii which I have) for the purpose of focus-stacking: shooting a few dozen shots with minimal-step focal differences and blending them all in photoshop later.
It took me a while to find the API; Sony made one then pretty much scrubbed it from the internet and has ignored my emails asking for one :-)
But even with the API, achieving results are tricky. Whether or not the focus motor steps properly seems to be based on how busy the camera is at the time... I'm trying various combinations of [shoot] - [sleep] - [step the focus] - [sleep] - [shoot] but the exactly which of these steps actually succeeds just seems so damn random.
BlackMagic does some pretty innovative stuff. They offered an Arduino add-on board that could communicate with and control their cameras over SDI, including access to the Canon EF mount and its commands.
Somebody did in fact create a control board with a knob to control lens focus through this thing. It can be done. The manufacturers just refuse to do it.
Focus-stacking is a feature on (iirc) the A9 level cameras, which are 5x the price, so it's hardly surprising that they want to lock it down.
I've always wanted to know how the various autofocus systems worked. This page is incredible. I wish there was another one with the Nikon autofocus systems since that's what I actually have.
I still have several lenses with autofocus that don't have an AF motor in it at all, the motor is in the camera body instead there's a tiny screw on the lens mount that transfers the motor rotation to the autofocus parts in the lense. This was very slow and noisy though on my cameras.
Nikon's lens/body compatibility matrix is horrendously complicated. https://www.kenrockwell.com/nikon/compatibility-lens.htm
Whereas I think "no metering" is never a problem on Canon.
Over the decades as a DSLR shooter (on Canon), I even saw Nikon shoot themselves in the foot not once but twice.
* When Nikon introduced "E" lenses (electronic aperture, like Canon and all modern designs), very few bodies were compatible with it. They didn't have the foresight to introduce compatible bodies before any lenses with E were released. For example, this was released in 2008 ( https://www.kenrockwell.com/nikon/24mm-pc.htm ), and the earliest compatible body is from 2007. It doesn't work with tons of crop (DX) bodies like the D90, as well as any film camera.
* When Nikon introduced "AF-P" (stepper motor) lenses in 2016, there were no compatible bodies before 2013. Again, not enough future-proofing.
Meanwhile, Canon took a different approach. They pissed off users twice - when transition from the FL mount (1964) to FD (1971), and FD to EF (1987). They basically got it all right with EF - fully electronic, no aperture slider, no focus screw, big diameter.
I have even successfully used a year-2017 Canon lens with Ring USM AF and IS on an original EOS 650 film body (released in 1987), and both AF and IS work perfectly. Mind you, both of those features did not even exist at the birth of EOS. Presumably the AF electronic protocol is agnostic of what motor technology is in the lens, and IS can be a lens-only thing without the body knowing.
Nikon's compatibility gotchas don't hold a candle to Canon's excellent (albeit imperfect) compatibility. Nikon keeps making the same mistakes over and over again, such as not having a screw-drive motor in the F to Z mount adapter; the company really seems to hate people with old lenses and bodies. Oh and this is not new either; Nikon's F to 1 adapter (does anyone remember that highly cropped mirrorless camera?) also lacked certain features.
This third party company announced a Nikon F to Z lens adapter that will autofocus with older screwdriver autofocus lenses. They already had a Nikon F to Sony adapter and now they have a Nikon F to Nikon Z adapter.
https://nikonrumors.com/2025/02/28/monsteradapter-la-fz1-off...
I've been using my old 85mm f1.4 "D screwdriver" lens and others in manual focus mode on my Z8 and Z5.
What annoys me is that the AF-P could probably be supported on a lot more cameras, they did add it to the D3300 with a firmware update. They just didn't do it for more cameras...
Nikon's lenses seem to follow a similar timeline, with different names:
Wave motor = AF-S (1998)
Stepper motor = AF-P (2016)
Voice coil motor = SSVCM (2022)
https://imaging.nikon.com/imaging/lineup/lens/glossary/
Tolerably modern Nikon lenses have in-lens AF I believe.
Newer Nikon lenses have a few motor types but if I'm not mistaken they support old mechanical (body driven) aperture and focus drive as well (the D6 and somewhat recently F6 film camera were still for sale)
Nikon has a complex ecosystem keeping generally the same mount for many decades and general support for it, where others developed a specific autofocus mount when they moved towards those systems.
Their mirrorless Z mount and lenses don‘t have mechanical drive. Neither do their official F-to-Z adapters - much to the chagrin of people using their older glass.
Yeah the in-body motor thing is only for the older Nikon lenses, I don't think AF for that is still supported if you're using Z-mount adapters. Nikon seems to have a few similar AF motor technologies like the canon ones in this article.
Canon being late to the AF party (last century).
They ended up with all electronic contacts when they switched lens mounts for the “eos” series cameras though at the expense of not having backward compatable lenses. (I had a bunch of manual focus lenses at the time…) They had one “auto focus” camera for the old lenses the t80. (https://global.canon/en/c-museum/product/film115.html) but almost no lenses supported autofocus.
I think most of settings the camera sent to the older lenses (aperture primarily) was done mechanically.
The way these old cameras autofocused was pretty interesting. With some light passing through a semi transparent mirror onto different sensors.
Edit: this site has another page about autofocus. Quite extensive.
https://exclusivearchitecture.com/03-technical-articles-DSLR...
Shorter canon article: https://support.usa.canon.com/kb/s/article/ART170280
I have one of those mirrorless slrs now. The focus is amazingly good. Lots of software (eye detect etc). The focus sensors are integrated into the image sensor I believe.
All my nikon cameras focus with a semi transparent mirror on a specialized autofocus sensor. I even have an nikon film camera that has that cicuitry. You can't autofocus like mirrorless cameras with PDAF pixels in the main sensor with a mirror in the way.
There's a new third-party F to Z adapter that does have the screw drive: https://nikonrumors.com/2025/02/28/monsteradapter-la-fz1-off...
These piezoelectric drivers are very cool.
Here's a video showing the mechanism:
https://www.youtube.com/watch?v=7iHL4ZCkCKc
And here's a video showing such a motor at full speed:
https://www.youtube.com/watch?v=JtttNnmCVmU
Cool videos, thanks! The driver shown in the first video is indeed a piezoelectric actuator, but not the type Canon uses in their lenses. The driver shown in the second video is probably related to Canon's Nano USM.
Do companies like Canon manufacture these motors or actuators themselves? Or is there some other company that specializes in this sort of thing?
As far as I know, Canon controls a large part of the supply chain and has numerous factories to manufacture quality-critical parts (especially their lenses, zoom mechanisms, and barrels) themselves. But there are numerous components that are more efficient to obtain from third parties, such as microchips. I am not sure about the motors and actuators, though.
A note of possible interest for this crowd: all these motor types are open loop, so you can't actually command a specific state with any accuracy / repeatability. In other words, if you send a signal to the lens to focus at, say, 3m, it won't necessarily be in exactly the same state as the next time you focus at 3m.
This is why camera calibration can be tricky, and you often don't want to touch the cameras after you've done a calibration pass.
The system stops the motors once it reaches focus. Therefore is a closed loop system.
It’s open loop in that the measurement of it being in focus is reliant on the subject matter, and a different measurement.
I’m not making this up, camera manufacturers have told me to my face that focus is open loop, period. They can’t guarantee repeatable focus.
Notably the measurement isn’t of the state of the motor/gearing. Furthermore, being “in focus” means the subject matter’s out of focus blur is below some threshold; there is a range of focus states that qualifies — but those seemingly small differences can affect camera calibration, with >pixel-level differences in effective focal lengths.
It’s open loop.
I can back you up on that. I work on a highly custom system that uses EF lenses through an EF-to-serial adapter. The lenses don't even know what position they're at when you power them on. If you want to move them to a specific focus step you have to go through a homing procedure that drives the focus motor to a stop (I don't recall if it's the near or far stop).
Lens-to-lens it's definitely not consistent. Step 142 on two different lenses will have a different focal depth. We calibrate each lens ourselves (thankfully you can read the serial number through the EF mount) and then still have to do closed-loop image-based sharpness estimation to guarantee that things are as good as we can get.
How could they then tell to what distance they're focused?
I used to have two Ultrasonic lenses, the 17-40/4L and the 17-55/2.8. They both had distance scales which would move around as the lens focused.
My current Olympus lenses have a focus-by-wire manual mode, with a distance scale on the barrel. The camera also reports the focus distance in the EXIF. Are these just vague ballparks?
Sorry for the late reply. What we’ve found is that they’re all close but not identical. Our system has a point LIDAR pointed at the ground and then uses that to drive the focal distance for the lens (as well as input for some other sensor fusion stuff). We generally run the lenses wide open (f/2.0 or f/2.8). Mapping step counts to distance would get us close, but unless we did calibration for each individual lens it wasn’t close enough to get the focus tack sharp. When you’re looking at the focal distance in your EXIF metadata you’re probably not validating whether that 2.4m distance to the subject was actually 2.3m and for the vast majority of cases the absolute accuracy of it doesn’t matter. Same with the distance markers on the focus ring… you’re not setting the focus within a few cm based on those marks, just getting it close enough that you can do closed-loop focusing with an eyeball in the loop.
Once we did the per-lens calibration curve we didn’t have to do it again. I’m assuming the small differences are just manufacturing tolerance and they’re relying on the focus/phase sensors to get it the rest of the way.
Some - not all - of the EF lenses have a distance encoder as well. It's fairly approximate; I think it exists mostly for the benefit of the flash system, which needs a rough starting point for the distance in some of the open-loop modes.
For the current Olympus ones, I think there's a broadly similar encoder on the ones with proper manual focus scales, and the pure fly-by-wire ones reset focus at connection so the camera can work it out.
There's a list of the EF lenses and which forms of distance information they provide here: https://web.archive.org/web/20130425064359/http://www.lenspl...
Correct. The AF sensor continually takes more samples even while the motor is moving. You can tell because sometimes the motor overshoots and then you can see it come back.
Yup… but it’s not necessarily in the same state if you do it twice. It’s close, but not identical. Even if the image metadata claims the focus point is identical, the lens in all likelihood is not in the same state and will have some deviation in its intrinsics.
Great write up, thanks.
I have a question I’m hoping you can answer.
I have an EF 400mm f2.8 (2nd gen) lens. This thing focuses significantly faster with my old 7Dmk2 than my very new R6Mk2. With all other lenses it’s the opposite. Why does this happen?
I have been told that this might be due to the old 7D and 1D bodies sending more power to the AF motors, but haven’t seen this corroborated anywhere.
Hi fooker, thanks for your feedback.
I haven't heard of this issue before, but a quick research showed me that other people have (similar) AF problems with longer lenses on an R6 II. See here https://community.usa.canon.com/t5/EOS-DSLR-Mirrorless-Camer...
Sorry for not providing an explanation, perhaps I can find some time to look deeper into this - but I can't promise.
@ExAr - A treasure trove of information, thanks for putting in the effort!
Do you blog about how you go about creating these graphics? For example, this view finder image: https://exclusivearchitecture.com/images/technical-articles/... looks very impressive!
How do you simulate the rays+eye?
Thanks for your feedback. Until now I have never blogged about the creation of these illustrations, but it's a nice idea, there is tons of "behind-the-scenes" material. With regards to the viewfinder image, the actual path of light was drawn with the help of a Japanese optical engineer who was able to provide professional raytracing simulations.
This is seriously impressive! Just curious, how do you have the time to do all these deep-dives, and how can people give you money??
turnsout Thank you very much. Trying to understand complex things has always fascinated me. Indeed it takes a lot of time (rd. 1.600 hours for the complete series of Canon lens chapters) but its people like you who drive me forward. I don't have a donation page yet, I never thought it would be used :-)
Amazing website. If I didn't know better, I thought this content came from Canon itself. A couple of comments from me:
> The following comparison chart shows all of the autofocus drive types that have ever been used in Canon's EF, EF-S, RF, and RF-S lenses.
They missed mentioning EF-M, Canon's foray into mirrorless interchangeable-lens cameras.
> Although the ultrasonic noise cannot be perceived by the human ear, it can be picked up by sensitive microphones which can be problematic during video shootings.
I used several Ring USM lenses and can hear the noise when focusing. It's a soft swishy kind of white noise, not harmonic. The Ring USM definitely generates noise in the audible range, not only ultrasonic.
> The latest types of autofocus drives, including the Stepper Motor, Nano USM, and Voice Coil Motor, offer focus-by-wire.
Though I have nothing against focus-by-wire on principle, the implementations have not been good. Having used several EF-M (all STM) and RF (some STM and some Nano USM) lenses that have focus-by-wire, I find that the motor actuation significantly lags behind the user turning the focus ring (which is a digital encoder), and the granularity of the focus steps are visible.
Meanwhile, the old EF Ring USM's full-time manual (FTM) was really good by comparison; it didn't take much force, was fully responsive with no lag, and was fully analog when turned by hand and had no discrete steps. I miss that, as new lenses don't use this design.
> Canon EF 15mm F2.8 Fisheye (et cetera)
Please change the syntax to "f/2.8"; this is even in Canon's official pages. https://global.canon/en/c-museum/product/ef263.html
Why? Because f is actually a lowercase italic variable that denotes the focal length (15 mm in this case), and "/" (slash) really means division. The size of the aperture is (15 mm / 2.8) = 5.4 mm; that is the real diameter of the hole that light passes through. The absolute aperture is 5.4 mm and the relative aperture is f/2.8. https://en.wikipedia.org/wiki/F-number
Thanks nayuki for the detailed comment and positive feedback.
Thanks for pointing out the missing EF-M, I am going to add that during the next polishing round. With regards to USM noise, the ultrasonic whine can be picked up by very good ears. I think the Nano USM is excited in the range of 60+ kHz, so at least that type of USM motor is absolutely inaudible.
Syntax-wise with F2.8, you got me. I know it's not 100% correct, but it was a deliberate choice that I made. Historically, there is either 1:2.8 or f/2.8 which will always be correct representments of the aperture (exit pupil) diameter. However, pursuing a cleaner look, I avoided the division or slash characters because the relation between the focal length and the f-number is known by every photographer. This is why Canon has also made the transition to the syntax "F2.8" in all their RF and RF-S-lenses - at least their product names and labels printed on the lens barrels. (The same is true with the focal length where Canon doesn't print "mm" on the barrel - to achieve a cleaner less mathematical look). So I believe what I did here was just go with the trend :-)
> With regards to USM noise
Here is a video showing the "whooshing" noise of Ring USM, and it is clearly in the audible range (not ultrasonic): https://imgur.com/a/canon-ring-usm-autofocus-sound-YCrV1CR
> the relation between the focal length and the f-number is known by every photographer
I'm going to disagree with this. I think very few photographers understand what an absolute aperture size is. Here's a good test: If you take a "constant (relative) aperture" zoom lens like a 24-70mm f/2.8, what happens if you keep the aperture "constant" at f/2.8 and zoom the lens in and out? I think the answer will surprise many people.
I think most people don't even know that the f (or incorrectly "F") refers to the focal length; I think the vast majority of photography education just treats "F" or "f/" as some piece of fixed notation or incantation without explaining what it means. (Kind of like if I asked you, what does the "degree" in "degree Celsius" mean?)
> Canon has also made the transition to the syntax "F2.8" in all their RF and RF-S-lenses - at least their product names and labels printed on the lens barrels
Oh wow, I didn't notice this one. Your observation is correct; Canon did change their official naming of RF lenses to use the "F2.8" syntax instead of the previous "f/2.8". Oh well, I still disagree with it because it is bad mathematical syntax and spits in the face of tradition.
> The same is true with the focal length where Canon doesn't print "mm" on the barrel - to achieve a cleaner less mathematical look
True, and I have subconsciously noticed this.
> So I believe what I did here was just go with the trend :-)
Please don't change the naming of old lenses. Also, I would prefer new lenses to be hammered back into the old naming scheme, but that's more open to debate. Curiously, the camera shop Vistek sometimes uses old naming for RF lenses: https://www.vistek.ca/store/434924/canon-rf-1535mm-f28l-usm-... "Canon RF 15-35mm f/2.8L IS USM Lens"
Thanks for following this up.
I have checked the video you've linked. My suggestion would be that this whooshing noise is not generated by the piezoelectric part but rather from the connected focusing mechanism (especially the helical focusing barrel which turns, and the focusing lens cell which slides back and forth. These parts sit inside the stationary lens barrel, and so a bit of friction is unavoidable. Just my guess.)
Syntax-wise, I will have to think about it. As with all my other articles, I like being very precise with physical formulas and mathematical expressions. But here is my point: These values in photography are not pinpoint precise anyway. When optical engineers come up with lens designs, the precisely calculated values for the aperture are something like f/4.63 which is then simply rounded to the next possible f-value specified on the lens barrel, and calculated focal length is often 25-49mm and Canon simply names that lens 24-50mm on the barrel (other manufacturers do that in a similar way). So, this is just my personal opinion, but it appears to me that in the context of photography, these values are just halfway decent guide values, almost like approximate descriptions of the lens behaviour. For that reason, I do understand why lens manufacturers move away from the precise syntax, and treat these values more like an abstract product description. Of course, if used in a strictly scientific context, I would rather prefer to use f/4.5. But I might be pretty alone with this view.
PS: I have planned a new chapter explaining all this, including the interesting fact that you've mentioned, where the diameter of the exit pupil actually changes according to the focal length, even with a constant aperture lens. But it will probably be end of summer when this is going online.
> this whooshing noise is not generated by the piezoelectric part but rather from the connected focusing mechanism
Indeed. I never implied that the piezoelectric part makes an audible whooshing noise. I only meant to say that the entire Ring USM system makes that noise, and so it will affect audio recorded on the camera's on-board microphone. Either way, that noise explains the motivation for the development of STM AF and I guess Nano USM.
Agreed with everything else you said. Thanks again for your fantastic articles and replies!
> EF-M, Canon's foray into mirrorless interchangeable-lens cameras
Canon's foray into SMALL mirrorless interchangeable-lens cameras.
The R series are mirrorless, but not that small. too bad.
EF-M is fairly dead at this point isn't it?
I still use mine occasionally since there is no direct replacement in the RF product line (they're all much larger).
It’s an amazing article. And it’s only the tip of the iceberg - there are many more. This has to be the work of a madman (meant in the best possible way).
Thanks so much for putting so much effort into this, loved reading it: the diagrams and explanations are top-tier. Inspirational.
Oh neat! I have a Newscale micro motor demo kit from years ago, and I wondered if they were ever successful in the market. But that hip-gyrating Micro USM action sure looks familiar.
This sounds like a fun gadget. Just out of curiosity, do you still have that demo kit, and would you consider selling it? :-)
I always wanted to build a controller for Canon Telephoto lenses to use them with c-mount cameras and control the focus from a PC. Might be helpful for that.
A couple of universities built a telescope called Dragonfly, that utilizes 48 Canon 400mm lenses to photograph the skies of New Mexico. The kicker is they don't use camera bodies. Instead, they designed and built mounts that directly connect the lens to a sensor. A PC controls each lens through some custom electronics using commands they reverse-engineered from the Canon control software. These commands control focus, aperture, and triggering a photograph. Images from the 48 lenses are then combined in the computer into a highly detailed final image. This was done back in 2013 so its possible Canon has released an API by now but I suspect they still keep it proprietary.
https://en.wikipedia.org/wiki/Dragonfly_Telephoto_Array
I visited some similar telescope in Australia at https://www.sidingspringobservatory.com.au/ run by Macquarie University ... the students explained the lens system could not be too new and could not be too old. Basically it sounded like someone had partly reverse engineered one generation of the lens interface only.
I guess the camera and lens manufacturer wants their lenses to be used with their cameras and to have better results than other manufacturers, whereas random companies want to clone the interface and sell cheaper lenses that also work with the cameras. Realising how awesome the lenses are, the students want an array of them, however they don't have the budget to buy a similar array of top end manufacturers' cameras.
In Sydney they have interesting viewing nights at https://www.mq.edu.au/faculty-of-science-and-engineering/dep...
A number of machine vision camera vendors have higher-end variants that support active EF-mount lenses. They tend to be quite expensive models with larger sensor, since you tend to lose a lot adapting full-frame lenses to tiny C-mount sensors.
If you want a computer-controlled lens, one of the cheapest options is to use Blackmagic cameras that support SDI and/or USB control. You can then control either native M4/3 active lenses (focus, aperture, zoom) or supported EF-mount lenses if you use one of the EF->M4/3 active adapters.
I was a big fan of USM lenses when I had a Canon, but I had one go bad with fungus or something and for a while had only a 35mm prime which I even used to take pictures of birds. Then I lost my Canon and decided to get a Sony circa 2019 or so because all the reviews I saw for Nikon and Canon said the autofocus sucked on entry-level full frame cameras.
You should be able to get fungus CLA’ed. Leica shooters using true vintage lenses sometimes need to do this when picking up a find from a flea market.
To save anyone else looking it up, CLA == Clean, Lubricate, Adjust
https://photo.stackexchange.com/questions/15613/what-is-cla-...
https://markorussa.wordpress.com/2015/08/31/general-instruct...
Amazing size with such attention to detail, congratulations.
Finally I will find out why RF50.8, 35.8 focusing is so noisy.(2870 is not)
The graphics are amazing. Congrats for the work!
Meanwhile, Pentax screw drives go brrrrr
Woah, this is so well made. Kudos to the author. Upvoted.
I’ve recently been shooting film with an old EOS camera from the 90s I bought used and it was really nice being able to use the EOS lenses I bought for my DSLR in the 2000s and 2010s. It’s a dying standard now but it’s really impressive it lasted as long as it has, with significant technological innovation on both sides of the lens mount while retaining full compatibility. A brand new EOS EF lens still works with an 80s camera and a new 80D from 2017 can still use the lenses from the 1980s without any adapter. 30 years ain’t bad for a standard!
I must be old -- I vividly remember the Nikon crowd crowing loudly online forums (fredmiranda.com I'm looking at you) about how Canon broke backwards compatibility when they moved from FD to EF. Whereas you could slap an F-mount lens from the 1950s on any Nikon DSLR ever made, no problemo. (Remarkably, this continues to be true!)
Except if you mount an old, so called pre AI lens, on certain Nikons, you will have to disassemble the camera to get it off. Many modern Nikons can't autofocus older AF lenses due to not having a motor. And many AF film SLRs couldnt meter with manual lenses. So far from perfect compatibility. Nikon F mount was introduced in 1959. The oldest mount used today is probably Leica M from 1954
You could always physically mount them but there were compatibility issues as Nikon added functionality over the years. It was a tradeoff.
Oh yeah, that’s even better, totally agree, but it doesn’t negate my point. I don’t think Canon’s 30 or 40 years would be matched by a hypothetical present-day camera upstart, let alone Nikon’s 70 years.
Too bad Nikon practically stopped making DSLRs
To be fair, so did everyone else. Except Pentax.
As a Canon owner, Nikon has a much longer back compatibility range. Having autofocus motors and IS in body rather than in lens seems to be a part of their trick.
The traditional Nikon mount has a small screw that is turned. The camera autofocus speed is limited by the amount of torque that can be applied to the screw - which can make focusing some of the heavier lenses slower.
https://www.discoverdigitalphotography.com/2012/lens-mounts-...
The "AF-D" lenses have contacts back to the camera body that communicate distance information (that is in turn used by the camera body to calculate flash power).
The G mount lenses remove the manual coupling for the f/stop which means that only bodies that can control the aperture from the body can use them. My FM3A has no aperture control on the body and so with that camera, I unlock the aperture ring.
The AF-S camera lenses have the focusing motor in the body.
VR in Nikon is done in lens. https://kenrockwell.com/nikon/80400vr.htm It needs to - you can't jiggle the film around to keep it in the same place.
---
The impressive part of Nikon's compatibility isn't only the "you can use an AI-S lens on a modern (professional) body, but also "you can use any of the F mount lenses on an old body" (the G lenses don't have the f/stop ring and the E lenses have the focus motor in the lens).
While it appears that Nikon has mostly shifted to E and G mount, third party lenses are still being manufactured for the F mount.
https://www.bhphotovideo.com/c/product/1519140-REG/tokina_at... - and you can see all the parts of the F mount, manual aperture ring control, 5 pins for distance, '-' slotted screw for focus.
Difference between G and E-type lenses is that 'E' stands for Electromagnetic Diaphrapham.
The G type lenses have an aperture tab for diaphrapham control as dictated by the camera body, the E-type lenses leave this to being controlled by the camera electronically.
> you can't jiggle the film around to keep it in the same place.
Contax had an AF solution (Contax AX) that actually moved the entire film plane to focus. I suppose that the same thing could be done in other axises, but I also suppose that there is a reason that only one manufacturer tried it.
The challenge would be to move the film around multiple axis at a range that would be useful.
https://www.canon.com.cy/pro/infobank/image-stabilisation-le...
> Optical Image Stabilisation is effective with movement across a range of frequencies, so it can cope not only with simple camera shake (0.5Hz to 3Hz), but also with the engine vibrations encountered when shooting from a moving vehicle or helicopter (10Hz to 20Hz).
Nikon his shifted almost exclusively to their Z mount lenses. And some of these have in-lens and/or on-camera VR.
The SLR lenses are still in wide production. https://www.nikonusa.com/c/lenses/dslr-lenses/overview
The Z lenses are for the mirrorless bodies and only work on the mirrorless bodies. https://www.nikonusa.com/c/lenses/mirrorless-lenses/overview
As long as SLR bodies are produced, Nikon will continue to make SLR lenses (and probably for a while afterwards)
https://en.wikipedia.org/wiki/Nikon_D6 was released in 2020.
Nikon lenses will mount on any Nikon body, but that’s sort of where the compatibility ends.
Their f mount autofocus lenses are a variety of standards that are not at all backwards compatible across eras.
I have an EOS 35mm lying around that can use the same lenses too.
Am I spoiled by expecting these images to be dynamic? I remember some wonderful posts with graphs you could interact with.
You're probably thinking of https://ciechanow.ski/mechanical-watch/
I can't imagine how long this post took to make, let alone trying to do animations.
What an absolute masterpiece this website is! Thanks for sharing.
On the ring-type USM motor you can click on some images to see animations running. But there is no additional interactivity at this point.
Cannot scroll pages in Firefox.
Sorry to hear that. I tried on my Firefox and it worked. Not sure what might cause that issue.
depending on how the UA's anti-tracking is set up, your cookie pop-up (non-GDPR compliant btw) might prevent the scroll. It's this snippet
Which aspect of it is non-GDPR compliant?
the accept and reject buttons should not have differentiated colours. See point 3.4, 3.5 (p. 27 onwards) in https://noyb.eu/sites/default/files/2024-07/noyb_Cookie_Repo...
You are right, I can change that.
Hey, nitpicks aside your site is great! Coming from a Fuji buyer…
Much appreciated :)
I had this problem in Chrome. I turned off JavaScript and the page worked nicely. (Props for that!) I'm assuming my adblocker or auto-accept-cookies plugin did something weird.
Canon's lenses are great, but they're really expensive. The camara body itself has the same issue.
Sony seems to be the first choice for indie filmmakers and youtubers right now.
L-mount and M43 also seem to have great potential as Panasonic supports Phase Detection Autofocus.
edit:
There is no need to be cynical here. Market figures are cruel. I hope they are all good. Competition is the best. I should also mention the return of Nikon. The patent threshold of RAW such as RED is what we need to oppose.
As the old saying goes, buy the glass, rent the body. Camera bodies change much more quickly than the lenses, so as an indie your money is better spent on the glass. However, shit is so cheap now that you can almost upgrade your gear like your phone. IF you’re a cut ally making money with it, you can pay off your gear in one or two gigs. Unless you’re the “nephew with a 5D asking $500” to shoot that project
Most shops that offered still photo rentals in my city have gone out of business, and it's a major city. I used to be able to count on renting a big telephoto that I didn't need to own, for a few hundred per weekend, but these days I'd have to take a long drive to get to the one remaining shop.
Yeah, online rental places have pretty much made the locals fade away. The one in my local area requires a "deposit" worth the full sales price which makes it out of the reach of anyone but those well off.
We used to joke about the B&H rental by taking full advantage of the 30-day return policy. Even Amazon's return policy qualifies now. You just need to the up front funds available and hope that wear&tear is not noticeable to make the return acceptable. It's hard for local shops to keep up with that.
Hard for local shops to keep up with fraud? You don't say.
I like how the messenger is getting shot. It's as if my reporting of what actually happens is being interpreted as that's what I do. Just because I know of how the system gets played does not make me a player.
I get upset when people defraud my neighbors and friends in my community. that's what a local shop is. Just today i went to my local pharmacy to look for l-theanine based on some commentary on HN yesterday. I could have gone on amazon, or to walmart, or to CVS. People interfering with that business aren't cool, regardless of the motivation.
let me put it this way - i never even thought of using amazon as a rental service. even though amazon loves me, and i could get away with it, my brain doesn't spend time thinking about unethical things to do.
read this as: "Someone should have gotten upset before the deposit went up from jerks either breaking or stealing the rental equipment" - as my sibling says, this is why we can't have nice things.
aka "This is why we can't have nice things"
I chose early on to invest in Canon's L lenses even though I'm an amateur. Yes, they are considerably more expensive than consumer lenses but for the price, you get a more rugged, weatherproof body, sharper and faster glass, and a greater selection of both fixed and zoom lens. The down side is they are quite heavy making them more difficult to carry when shooting outdoors and they can require a stronger tripod to support the weight of the lens and camera. But, the images they can produce can be simply stunning at times and that, after all, is why we do photography.
Considering my 70-200 can still hang after 10 years and will last another 20 easily with some TLC I was fine with my $1500 purchase - which I made as an indie filmmaker!
Jokes aside Sony glass really isn’t all that much cheaper.
Agreed. EF's short flange focal distance can also be easily connected to Sony to utilize dual ISO. Indie uses manual focus anyway.
Yeah I have a lot of beef with Canon but there's no denying they make universal donor lenses