Review: Canon EF100-400mm f/4.5-5.6 L USM Mk1
This lens is only available used. It was extremely popular as a relatively low-cost way into Canon EF long zoom lenses, and has been replaced by a Mk2 version. It's still a popular buy. I really like mine: it can give really great results. It has one or two idiosyncrasies and limitations so a little understanding will help you get the best from it, especially in more demanding situations. I've also included a very brief and basic introduction to an aspect of optical lens design, which helps to highlight a technical feature of this lens........
......so here's a semi-technical discourse you may or may not be interested in. You can of course skip it. But it does give a tiny bit of insight into how much work went into your camera lens. It also highlights a special feature of the component set used in this one.
This lens actually contains a very valued optical component; a fluorite lens element. A modern camera 'lens' generally contains at least 6 glass parts you would've individually recognised as a lens. Each little lens looks something like a round spectacle glass. In a complex optical system, each will have various diameters, thicknesses and profiles. The complete camera 'lens' is really a 'compound optic'; many individual lenses working together. Each little lens is called an 'element'. The EF100-400 Mk1 has 17 elements all in a row. Just one is made of the aforementioned fluorite, which is actually a term for Calcium Fluoride in crystalline form, or CaF2. This element here has largish diameter, and is located in a triplet group near the front of the barrel. Fluorite is expensive and hard to manufacture. It is no longer used in any but the most expensive Canon super-telephoto lenses, and generally for at most two of the elements. For any further understanding of the relevance of fluorite in the design, you may need my 'Lens Design 101' spiel below. Why not read it, it'll be fun! You too can become a physics nerd when all you wanted was to do was take nice pictures.
Lens Design 101
Lenses work by redirecting incoming beams of light, according to their profile. The intention of a lens or lens group is to bring a cone of incoming light to a sharp focus. A 'focal plane' of focussed light gives the image on the sensor or film. Lenses achieve this because light bends at the boundary between air and glass. We call this 'refraction'. Every transparent material bends light to a different degree, specified by 'refractive index'. Fine so far. A problem arises because glass does not bend light of different colours by the same amount. We call this 'dispersion'. Low dispersion is good, more is bad. This means a single glass lens won't focus a picture with many colours in it properly, the more so if the glass used has high dispersion. The result is artificial colour fringing, most visible on light/dark boundaries. As always, there's a technical name for it. We call this artefact 'chromatic aberration' or CA. CA will also reduce general resolution. It gets to be more of a problem with telephoto lenses, which is why fluorite is used in Canon's top super-telephotos as part of the 'recipe' to mitigate it.
How is colour fringing tackled?
As always, people came up with ideas to remedy this particular problem. There's a design called the achromatic doublet, probably the first successful compound lens. It's still very common. It considerably reduces the colour fringing problem.
This is one quite simple example of the use of compound optics to optimise the overall result. A lens like the 100-400mm is a very much more complex design task, requiring a lot of iterative optimisation, for Chromatic Aberration minimisation and other requirements.
The dispersion, or 'colour spreading' of optical materials is defined by something called the Abbe number. Higher Abbe number means less dispersion and less fringing. Fluorite has about the highest Abbe number available; about 95%, where 100% is no colour spreading. Many common glass types are down at around 60%, which is why they make good prisms. But only if you like to demonstrate a 'rainbow' like the one on the cover of 'Dark Side of the Moon'. Fluorite would be very poor for that. There are a few expensive glass types, such as Ohara FPL53, with similar Abbe number. Fluorite also has very low internal light scattering. You can see this when we point a visible wavelength laser at it. You can't see the light path inside the 'glass'.
In addition to being costly to make in suitable sized slabs, fluorite is fragile and hard to work with. It needs cutting and polishing with extreme care and precision, more so than other glass elements. It also has a high temperature expansion coefficient, one reason L telephoto bodies are painted white. It helps minimise heat absorption, which would eventually find its way into the optical elements.
So if you get one of these, you are in slightly exotic territory. There aren't too many lenses with large fluorite elements, and most cost much more than this one. The Mk2 version, released in 2014, uses a smaller such element, nested further back inside the lens. The nearest equivalent RF lens, the 100-500mm, has no fluorite. It attempts to equal or better the overall result using other, less exotic, high-performance glass types.
Other Aspects of Camera Lens Design
Avoiding chromatic aberration is only one aspect of total lens design. Camera lenses often need to zoom. This one goes back to 100mm, making it extremely versatile. Distortion, resolution, vignetting and overall light transmission are some other areas to optimise. Designing to avoid complex and expensive-to-manufacture element cross-section profiles is usually a further constraint. Then there's autofocus, a major design task, performed in association with the camera body. There's a lot of very skilled and painstaking work involved overall. The design and production of lenses like this was and remains a highly specialised business.
Does it work?
Yes, on the whole, it does work, rather well. Personally I think the fluorite front element results in a richness to colour tone and light rendition not entirely available otherwise. You can't really put that back in with software in quite the same way. Not with any I've got, anyway. I used to own a non-Canon EF fit lens of a similar general nature, and the images never had the same vividness, whatever you did with them in post processing software. This other lens was as sharp as the Canon, perhaps even a little more so. However, there's definitely more to a lens than resolution performance figures and other straightforward measurables. I'm speculating, but freedom from internal element light scatter and good suppression of flare and internal reflections can make a lens capable of special results. And fluorite has been called the 'magic dust' in Canon's top-end super-telephotos. Whether you believe that or not, different materials do give quite different results on the many things we do measure. A look at an optical glass catalogue will illustrate this.
And what about the bane of telephoto lenses, Chromatic Aberration? It's there, if you look in at the pixel level, but it's only a pixel or two wide on an 18Mp
APS-C. The in-camera lens correction profile in my 5D III removes it pretty much entirely.
Wide open, at f/5.6, this lens is not perfectly sharp, as some people have pointed out. There's another issue called 'spherical aberration' which is likely responsible for that. Things here improve progressively when stopped down and by f/7.1 it's rather impressive. It may not bother you in the first place, and centre sharpness is decent even wide open. Usually I don't stop down and think this problem has been exaggerated. It's not exactly soft left wide, and there's a certain 'body' to the result which can work well with animal subjects.
It looks slightly intimidating, but is fairly simple to use. Unless you want to focus close, leave the focus selector at 6.5m-infinity. Mode 2 of the stabiliser is useful for side-to-side tracking, where the stabiliser will only attempt to correct in the vertical axis. If you're taking big subjects, taking up a lot of your frame, and if there's reasonable light, just snap away and focus should be fine. Otherwise, you may sometimes want to tweak with manual focus to get the very best out of this lens. There is full-time manual focus override available even if you are set to AF. You need to set autofocus in the camera to single-shot, and normally I'd put the shutter to high-speed continuous. A small bird often moves its head really fast, and even with a DSLR you won't be quick enough on the shutter, so take a burst and select later for optimum composition and sharpness. The focus ring is very smooth, although it does seem to drag the adjacent lock ring easily. Autofocus is fast and precise in good light with a good sized target. R-series mirrorless cameras will probably improve autofocus consistency if not speed. It's not a combination I've tried. I've found autofocus performance in moderately low light to be mediocre, although this will depend on the DSLR body to some degree. My 7D benefitted a little from micro-focus calibration of the lens-camera combination, whereas my 5D3 was spot on as supplied.
When photographing birds or small animals, DSLR autofocus can struggle with accuracy, particularly in low light. Small aperture lenses exacerbate this as they give the phase detect autofocus less light path difference to work with. Telephoto photography like this is probably a learning curve unless you anticipate the issues. Depth of field at maximum aperture can also be very small, particularly with a close subject. Additionally, both camera shake and subject motion result in large position errors within the frame. In general, set the camera body to single point focus. It's still often best to verify or adjust manually, and a lens like this makes that easy to do. The large bulk of the focusing adjustment takes place automatically, leaving potential for a minor refinement if necessary. Any autofocus system is likely to hunt if your subject is partially obscured by foliage etc. Prioritise short exposures, and experiment with higher ISO settings in the 1000-5000 range. I rarely shoot birds at less than ISO 800, to help freeze movement, and a modern DSLR is very clean at that setting and usually considerably higher.
Image stabilisation of an early form is fitted to this lens: it really was a bit of a trailblazer, when launched in 1998, for a lower-cost super-telephoto. This is certainly effective, but you will get slightly better results if you have a tripod, or enough light, so it can be switched off. The Mk2 version of the lens provides usefully more stabilisation. Remember that it is of no help with subject motion.
This is not a 'fast' (large aperture, low f-number) lens, even before stopping down, and it definitely gives its best in good light. Rapidly-moving animals at long focal lengths in low light are not its forte. But to improve matters significantly you'll have to spend much, much more.
Zooming requires a linear push-pull movement, not the usual twisting action. This quickly becomes intuitive, and actually encourages good hand holding technique and stability. There's a black rubber ridge near the front element for support. There's a ring to adjust sliding friction which can also be used to lock the lens at a particular focal length.
Behind this front rubber support ring is the manual focus ring, also black, which should rotate very smoothly. Both are visible in the picture below left. Behind it is another, slightly smaller, white, ring. This is the variable friction clutch. Turning it clockwise increases friction grip on the extending barrel. Turn it anti-clockwise if you want to zoom more freely. On my example, when you adjust the manual focuser, it drags the lock ring with it. I've heard other people report this too. There's not much clearance between the two rings. This is not usually a problem in practice, because the autofocus system doesn't turn the manual ring, even though turning the ring subsequently adjusts focus. And manual focus is likely to be a minor 'tweak', so the friction won't change much. However, I feel it is best not to over-tighten the lock-ring. It's tempting to do so in order to keep everything solidly locked for storage, or at maximum zoom. But this seems to force the lock ring hard toward the focus ring, possibly risking damage.
Some copies of the lens have had mechanical issues with these focusing and zoom locking rings. Having said that, the lens generally still feels beautifully made, and very rugged and professional. I feel it's best to buy with a good warranty and the possibility of return unless you know exactly what you are looking for. There will also always be a degree of optical sample variability with complex lens systems, so check yours out. Personally I've found Canon products to have good consistency, if not always total perfection. The latter is close to unattainable, frankly.
This is one of my favourite lenses, even though it has required a repair for the issue mentioned above, which had made manual focus impossible. The lens has character, and rewards some experimentation to learn to best use it. It can be really excellent, but if used outside of its element (sorry for that...), it may disappoint also. Some early samples were reported as having sub-par optics. When everything comes together, and assuming your sample is good, as most seem to be, I think it's very rewarding and capable, and achieves results otherwise very hard to attain at the price point. The only other possibilities likely to be similarly pleasing for the outlay are used examples of the EF300mm f/4 L and EF400mm f/5.6 L primes, also both older designs. The first may lack reach for you and both are less flexible in use. Very good examples of the 100-400 Mk1 now go for about half the last new price.
And having emphasised the idiosyncrasies, for many situations, the EF100-400 Mk1 is not hard to use at all. Most super-zoom photography requires a little skill. But my first outing with this one was to an air show on a bright summer's day, and I left the camera on auto. I did 'point-and-shoot' and was extremely pleased with the results.
With most optical materials, refraction increases with reducing wavelength of light. Here, blue is refracted most, as it has the shortest wavelength. We can see that light composed of many colours is not brought to a sharp focus. Images will appear 'smeared'. No two colours reach the same focus.
Credit: DrBob at English Wikipedia
We've now added another element. The other element is differently shaped; it's concave. It's also made of a different sort of glass. This second element bends the beams the other way. Dispersion operates in the opposite sense to the first element. We can now get the same focus for two wavelengths. This is a considerable improvement. Since red and blue are similarly focused, colour fringing will be greatly reduced, and in this case, mostly green.
Credit: DrBob at English Wikipedia
This little fella (right) was shot at f/6.3, also at the 400mm end.
There's really not a lot wrong with sharpness or contrast. It's a slight crop from a frame taken with a 7D Mk1 APS-C. I often manually tweak focus for shots like this, where the subject is small in the frame.
In general, I find colour and contrast really pleasing on this 1998 design. It was made until 2014 when the Mk2 took over. Some facets of performance have to do with production cost, rather than just ongoing development.
A natural crystal of calcium fluoride. The material used in lenses is synthetic but basically the same chemically.
Credit CarlesMillan, CC BY-SA 3.0 <>, via Wikimedia Commons
Left: Shot wide open, f/5.6, at 400mm. Fairly good light, on an EOS 5D Mk3 full-frame. Single point autofocus. This is a fully professional result in my opinion.
Above: 400mm at f/5.6 again. Using the EOS 5D III in-camera lens profile.