Keeping it steady
Andy Westlake explains everything you need to know about image stabilisation and how to get the best results from your camera
andy Westlake explains everything you need to know about image stabilisation and how to get the best out of your camera
Image stabilisation has become so ubiquitous and universally accepted as a useful tool that it’s easy to forget it’s become a mainstream feature only recently. But, now, while no camera maker would dream of selling a camera kit without some form of stabilisation except at bargainbasement prices, just a decade ago Canon and Nikon were happily shifting shelf-loads of DSLRs with unstabilised 18-55mm kit zooms.
Perhaps because of this, there are still a number of myths and misunderstandings regarding stabilisation, its benefits and its pitfalls. In this article I’m going to take a close look at all aspects of the technology, so you can better understand how it works, what it can and cannot do, and how you can exploit it to get better pictures.
In 1994 Nikon became the first company to include optical stabilisation in a camera, in a 38-105mm lens that was built into the Zoom 700VR 35mm film compact. However the following year, Canon introduced the technology to a wider audience with the EF 75-300mm f/4.5-5.6 IS USM: the first image-stabilised lens for SLR cameras. At
around £400 it was expensive for a consumer telezoom, and its stabilisation system was considerably less effective than modern ones, bringing perhaps two stops of benefit.
Like most transformative technologies, image stabilisation was rather derided when it first appeared; it was seen as a crutch for poor handholding technique that offered little practical advantage for the price. But it expanded the range of light conditions in which photographers could shoot handheld, especially given the low- ISO film most of us were using at the time. The Canon lens also introduced photographers to the benefits of a stabilised viewfinder, allowing more accurate composition.
Initially, relatively few lenses included image stabilisation, and these were mostly large, expensive telephoto primes. More than a decade passed before the technology appeared in the mass-produced, inexpensive kit lenses routinely sold with SLRs.
Canon and Nikon embraced in-lens optical stabilisation simply because with film, it was the only practical solution. Consequently, thirdparty lens makers such as Sigma, Tamron and Tokina also adopted in-lens stabilisation to keep their offerings competitive. However, once digital cameras started to become mainstream an alternative approach emerged, with Minolta debuting in-body stabilisation in its Dimage A1 bridge camera of 2003. Similar technology has since found its way into models from Pentax, Olympus, Sony and Panasonic.
While photographers have spent a lot of time arguing over the relative merits of these two approaches, it turns out that their strengths are actually complementary. So most recently we’ve seen Panasonic, Sony and Olympus evolve hybrid systems that combine the best of in-lens and in-body stabilisation. This approach gives astonishingly effective results, with Panasonic and Olympus providing the current state- of-the-art technology.
What does IS do?
Whenever you’re handholding a camera or lens, you can never keep it completely steady. Instead, the precise angle the camera is pointing will be continually changing, as your muscles work to counteract its weight. If this movement is sufficiently large during the course of the exposure, then your images will turn out blurred.
Image-stabilisation systems use gyro sensors to measure these camera movements that cause image blurring. Optical systems then move a group of elements within the lens up-and- down and side-to-side to compensate, so that the image is projected stably onto the film or sensor. In-body systems move the image sensor itself to track the shaking image; as well as moving up-anddown or side-to-side, the sensor can be rotated around the lens’s optical axis.
It’s crucial to understand that IS doesn’t affect image blurring from subject motion at slow shutter speeds. So it’s of little benefit if you need to keep moving objects sharp; on the other hand it’s genuinely useful when you want to exploit motion blur for artistic effect.
Axes of evil
You’ll often see image-stabilisation systems described in terms of the number of axes of shake they can counteract. In-lens systems are almost invariably 2-axis: they correct up- down and left-right angular movements (pitch and yaw). In most situations these are by far the biggest cause of blur, so 2-axis stabilisation works perfectly well for most photography.
However during long- exposure shooting, another type of camera movement becomes important, in the form of rotation around the lens axis (or roll). There’s nothing
in-lens systems can do to combat this, but in-body systems can correct it by rotating the sensor in the opposite direction. This means 3-axis in-body systems can be more effective with slow shutter speeds, particularly when you’re shooting with wideangle lenses.
For close- up photography, there’s another problem. At high subject magnifications, movement of the lens’s entrance pupil a fraction of a millimetre upwards or sideways relative to the subject can result in visible blurring. This can be addressed by both in-lens and in-body systems, but relatively few lenses attempt to deal with it. However a few of Canon’s macro lenses include the firm’s ‘Hybrid IS’ that provides 4-axis correction. In-body 5-axis systems can also correct for this motion.
Some readers may have realised that logically, there’s a sixth ‘axis’, which is a back-and-forth movement of the camera relative to the subject. However this can only be compensated by continuous autofocus. Camera and lens manufacturers generally like to tell you how effective their stabilisation systems are, and the CIPA standard of testing now used should ensure that these quoted numbers are comparable between different firms. You’ll see wording along the lines of ‘the optical stabilisation delivers four stops benefit’, which essentially means that you’ll be able to use shutter speeds four stops slower than would be otherwise possible. However, this means you first need to know what shutter speed you could safely use without IS.
To define a ‘safe’ shutter speed at which images shouldn’t be blurred due to handshake, photographers have traditionally used the ‘1/focal length’ rule of thumb. This suggests that in order to get sharp images, you need to shoot at speeds of at least 1/50sec with a 50mm lens, or 1/25sec with a 24mm lens, or 1/100sec with a 100mm lens, and so on. This assumes you’re using a full-frame camera; otherwise, you’ll need to take the crop factor into account. So with that 50mm lens, you would need at least 1/80sec on an APS- C camera or 1/100sec on a Micro Four Thirds model.
In reality, this is an overly simple way of looking at things. Camera shake is random, meaning it doesn’t affect every picture in the same way even when they’re shot identically. So if you take five shots at 1/50sec with a 50mm (equivalent) lens, a couple might be pin-sharp, one might be distinctly blurred, and the others somewhere in between. Drop the shutter speed to 1/25sec and you could still get one or two usable shots out of five. This random nature means that there’s actually no hard cut- off point for a safe shutter speed where pictures are ‘sharp’ not ‘blurred’.
What’s more, the rule assumes that you always get the same degree of shake, regardless of the camera or lens you’re using. But there’s a world of difference between handholding an 18-55mm kit zoom that weighs 200g and a 1.5kg 70-200mm f/2.8. In general, as your lenses get larger and heavier, you’ll need to err towards selecting increasingly higher shutter speeds than the rule might suggest. Other factors can also affect shake; for example it’s easier to keep a camera steady if it has a decent handgrip. Equally, low temperatures, high winds or even how much coffee you’ve been drinking can all increase shake. Even the angle at which you’re pointing the camera will have an effect; you’ll get less-sharp shots pointing your camera up or down, compared to holding it level.
What’s more, if you want to get pixel-sharp images from a modern sensor of 24MP or
‘If you exploit image stabilisation fully, it’s possible to get sharp images handheld at very slow shutter speeds’
more, you’ll need to use faster shutter speeds again. Personally I’d err towards choosing shutter speeds twice as fast to be sure of getting genuinely sharp shots, meaning I’d use 1/100sec with a 50mm equivalent lens.
What this all means is that if a system promises four stops of stabilisation with a 50mm equivalent focal length, I wouldn’t take that as meaning I should always get sharp shots at 1/3sec (i.e. four stops slower than 1/50sec). Instead I’d expect to be able to get most of my shots decently sharp at 1/6sec, with a reasonable chance of getting some sharp pictures at slower shutter speeds, especially if I shoot a few duplicates. Likewise, with a 70-200mm f/2.8 giving four stops of stabilisation, at the 200mm end I’d expect to need 1/30sec at least to be confident of getting sharp results.
That said, if you exploit image stabilisation fully, it’s possible to get sharp images handheld at very slow shutter speeds. This is particularly the case when you’re able to support the camera and lens by leaning on a wall or railing, or using a monopod. You’ll get best results when you also practise good handholding technique: adopt a comfortable shooting stance, hold the camera to your eye, and
support the lens with your free hand. In such cases you might be surprised by just how slow you can go; with the latest cameras and lenses I regularly attempt handheld shots at shutter speeds of up to 1 second with wider lenses, and can often get a sharp image out of four or five attempts.
As usual, the best thing to do is to carry out some tests to see what works for you. Shoot a detailed test scene at a range of shutter speeds, taking five shots at each. Examine the resultant files on your computer and determine which meet your standards for sharpness. For optically stabilised zooms, you might like to test at the two extremes of the zoom range; however I’ve found that in-lens systems tend to be equally effective at all focal lengths. The same certainly isn’t true for in-body IS, though, which tends to be most effective with wider lenses. So for these you need to test with a range of different lenses. In-lens versus in-body IS With in-lens stabilisation systems being used by the big two camera manufacturers, Canon and Nikon, it might be tempting to assume that it’s a technically superior approach to solving the problem of hand-shake. But as I’ve already explained, it’s not that simple – the firms’ adoption of optical stabilisation was initially dictated by the requirements of film SLRs.
However, optical stabilisation does have some very significant advantages. It’s more power efficient and effective with telephoto lenses, because a relatively small movement of the corrective optical group is needed to combat blur. By contrast, in-body systems simply can’t move the image sensor far enough, quickly enough to compensate for shake with lenses of around 300mm or longer. Yet it’s precisely these optics that absolutely need stabilisation to be usable handheld – it’s barely possible to compose pictures accurately with ultra-telephotos otherwise.
Optical stabilisation can also be incorporated relatively easily into small, cheap zooms, whereas in-body IS generally adds substantial cost to a camera, along with a little extra bulk. With SLRs, optical IS also stabilises the image that’s projected onto the separate autofocus sensor, theoretically resulting in more accurate and reliable AF compared to using in-body IS. With mirrorless cameras, of course, in-body IS stabilises both the viewfinder image and the AF system.
On the other hand, with in-lens IS, you have to pay for the privilege every single time you buy a lens. Optical IS has also proven difficult to include in certain lens types, meaning that there are relatively few image-stabilised wideangle zooms and fast primes on the market. In contrast, if your camera has in-body IS, every lens you can use automatically becomes stabilised.
It’s worth noting that not all in-body IS systems are created equal. Instead, smaller sensors have an inherent advantage over larger ones: not only is the sensor unit lighter, which means it takes less power to move, but it also has to travel a shorter distance to achieve
the same stabilising effect. As a result, Micro Four Thirds cameras tend to offer more effective stabilisation, and continue to work with longer lenses, compared to APS- C or full-frame models.
Dual IS systems
With in-body IS excelling in those areas where in-lens IS falls down, it’s no real surprise that some camera manufacturers have started to use them hand in hand. Sony, Olympus and Panasonic all now offer some kind of Dual IS system in which in-body and in-lens stabilisation systems work together. For example, if you were to mount an optically stabilised lens on a Sony mirrorless model with in-body stabilisation, then the lens’s optical correction is used to combat tilt and yaw, while the sensor reduces blur from roll around the lens axis. Olympus’s Sync IS goes a step further and uses both the in-lens and in-body systems together to correct larger tilt-and-yaw movements.
Hybrid systems are the most effective of all, but are still somewhat in their infancy. Olympus’s Sync IS system is staggeringly effective with a specified 6.5 stops stabilisation, meaning its flagship OM- D E- M1 Mark II mirrorless camera allows handheld shooting at shutter speeds longer than a second with the M. Zuiko Digital ED 12-100mm f/4 IS Pro lens. But the firm only offers one other optically stabilised lens, in the shape of the M. Zuiko Digital ED 300mm f/4 IS Pro. Meanwhile, Panasonic has a much larger range of stabilised lenses, but only its most recent bodies include in-body stabilisation that works as part of its Dual IS system. Frustratingly, Olympus’s and Panasonic’s hybrid systems aren’t cross- compatible, so if you use a Panasonic lens on an Olympus camera (or vice versa) you can use either the optical or in-body IS, but not both together.
Usually you want to correct blur from all possible axes of camera movement. But when you’re panning the camera to track a moving subject, the IS system can fight against what you’re trying to do. So most cameras and lenses have specific modes that turn off stabilisation in the direction you’re moving the camera. Most modern systems also attempt to detect this kind of camera movement and automatically switch to panning mode.
Certain lenses also offer additional stabilisation modes. For example, many Nikon lenses have an ‘Active’ setting that’s designed for use when shooting from a moving vehicle. Essentially, this allows the optics to move further so they can correct larger shaking motions of the lens. However, this increased decentring can bring a sharpness penalty.
Canon and Tamron have also recently introduced another IS mode. While Mode 1 is for normal use and Mode 2 for panning, Mode 3 doesn’t stabilise the viewfinder image and only activates IS at the point of exposure. In principle this gives the greatest stabilisation.
Creative uses of IS
Because IS first appeared on telephoto lenses, it’s tempting to think that it’s not really needed with other lens types. After all you can safely shoot at slower shutter speeds with wideangle lenses, while large-aperture primes let you use higher shutter speeds to avoid hand-shake.
This, however, rather misses the point. Image stabilisation is useful whenever you might want to shoot with a slow shutter speed handheld, regardless of the lens you’re using. Quite simply, it lets you use the lowest possible ISO for the best image quality while keeping shutter speeds high enough to avoid unwanted motion blur. You’ll certainly want stabilisation if you shoot video handheld, too.
You can also exploit stabilisation for more creative uses. Typically, you might wish to use a small aperture for extended depth of field. But you could also use a long shutter speed for deliberate motion blur, without having to set up a tripod. Alternatively in the low-light, high- contrast situations that you’ll come across when shooting about town at night, stabilisation can enable you to expose to maintain highlight details, while also using low ISO settings to exploit their greater dynamic range in order to pull out details in the darker areas of the image.
Image stabilisation can be built into the lens or camera body. Some systems use both at the same time
Here I used a long shutter speed handheld for creative blur Sony Alpha 7 II, FE 24-70mm f/4 OSS ZA at 27mm 1sec at f/10, ISO 50
IS also lets you use smaller apertures Canon EOS 5D Mk IV, Tamron 70-200mm f/2.8 G2 at 200mm, 1/30sec at f/16, ISO 100
Above: Stabilisation is ideal for shooting handheld at night Right: Here I was able to use a low ISO without sacrificing depth-of-field Olympus OM-D E-M10 Mark III, 14-42mm f/3.5-5.6 at 42mm, 1/10sec at f/6.3, ISO 200 Olympus OM-D E-M5 Mark II, 12-40mm f/2.8 at 19mm, 1/5sec at f/2.8, ISO 200
In Olympus’s Sync IS, the in-lens and in-body systems work together to combat pitch and yaw motion