We’ve all had great shots spoiled by blur. One of the first jobs most of us do with any new set of images is work through them to discard the technical failures – and poor focus is one of the main reasons for failure.

Some images are blurred because of bad focusing or insufficient depth of field. Many more are blurred, though, because the camera moved during the exposure. The popular theory is that there are ‘safe’ and ‘unsafe’ shutter speeds for handheld shooting. In practice, the boundaries aren’t so clear. Our boxout on Safe Shooting on the next page explains this in more depth and gives a few rules of thumb for working out ‘safe’ speeds.

Camera shake occurs mainly in two key situations; low light and long lenses. Low light means longer exposures and increased risk of shake; longer lenses magnify the effect of any camera movement. You also risk shake when taking night shots with slow flash – the flash-lit foreground will be sharp, but the background may not be. Sharp handheld macro shots can be difficult to achieve, too.

There are other circumstances where camera shake can creep in. If you’re working in a nice, warm interior with all the time in the world to shoot in, it’s possible to get sharp shots even at extremely ‘marginal’ shutter speeds. But outside, in the teeth of a bitter gale, it’s another story. You might struggle to get sharp shots at shutter speeds two or even three times faster than normal.

Sometimes you’re working on a platform that is itself moving. Helicopters, for example, produce a great deal of vibration, and if you’re standing in the back of a Land Rover bouncing over the veldt at 40 miles an hour, you’ll discover just how hard it is to keep your camera steady. Image stabilisation systems are designed to counter this camera movement – but how do they do this?

How they work
Image stabilisers have three main components. The first is a pair of tiny gyroscopes spinning on two mutually perpendicular axes. It’s one of the extraordinary properties of gyroscopes that they are able to ‘detect’ movement. One gyroscope detects ‘pitching’, or vertical movement, while another detects ‘yaw’, which is horizontal movement. Diagonal movements are measured by combining data from both gyroscopes.

A processor then constantly checks, or ‘samples’, the movement of these gyroscopic sensors – Nikon’s SLR lens VR systems check for movement 1,000 times per second. The processor controls tiny motors that counter the detected camera movement by shifting either purpose-built lens elements or the camera sensor itself (depending on the system) to compensate for this movement. In Nikon’s VR system, two VCMs (Voice Coil Motors) are used to produce horizontal and vertical movement respectively, and the two are combined for diagonal movements.

Clearly, this has to happen with exceptional speed. For the system to work, the gyroscopes must pass movement information to the sensor, which then has to drive the motors, which themselves have to respond near-instantaneously – and all this during the exposure itself. There are limits to the amount of movement that image stabilisers can cope with. Manufacturers will typically quote their effectiveness in ‘shutter speeds’. For example, they might say that their system allows you to shoot at shutter speeds two or three times slower than usual.

Image-stabilisation technology continues to advance, though. Nikon reckons that its latest stabilisation systems can produce sharp shots at shutter speeds up to four stops slower than usual. It’s not the same as using a tripod. A tripod will give you sharp shots at any shutter speed, while an image stabiliser simply improves your chances when shooting handheld at ‘marginal’ speeds.
    
Lens or sensor shift?
Image stabilisation systems come in two main types: lens-based and sensor-based. Lens-based systems are the most common and have been available in professional-quality telephotos for some time now. Increasingly, though, they’re appearing on ‘standard’ zooms too, like Canon’s 24-105mm IS USM optic for full-frame DSLRs like the Eos 5D, which is a real step up for enthusiasts.

It’s not the whole lens that moves during the stabilisation process, but a single element or group within the lens. When this element moves, it shifts the position of the image on the sensor. Externally, you won’t see any sign of movement, but if you put your ear to the lens you will hear the gyroscopes rotating while the image stabilisation is active.

Image stabilisation (or Vibration Reduction in Nikon cameras) is a very delicate science. The on-board processor has to be able to distinguish between ‘deliberate’ movement, when the photographer is simply reframing the shot or adjusting position, for example, and ‘accidental’ movement during exposure.

In the Nikon system, half-pressure on the shutter release activates the VR system, but allowance is made for deliberate camera movement. When the shutter release is pressed in fully, though, the system will then counter all movement until the exposure has been completed.

Panasonic offers an optical (Mega OIS) system in many of its compact cameras with two operating modes: Mode 1 and Mode 2. In Mode 1, the stabilisation component in the lens is moving constantly, which gives a steadier image on the LCD, but less effective shake reduction during the exposure. In Mode 2, the image stabilisation is activated only when you take the shot – the viewfinder image can jitter and shake, but the image stabilisation during the exposure is more effective. This is because in Mode 2 the stabilisation element is kept at the centre of the optical path during composition, giving it a greater range of movement when you take the shot. Nikon uses this principle in the design of its own VR systems, resetting the correction element to a central position directly before the exposure. Konica Minolta pioneered a different system, which has now been adopted into Sony’s DSLRs. Instead of moving lens elements to shift the position of the image on the CCD, it worked out a way to shift the CCD itself. This has some advantages, chief among which is the fact that the system should work with almost any lens. Both the Alpha 100 and Alpha 700 have adopted this technology, adding an on-body switch to control whether or not the stabilisation is applied. Olympus is using a similar, in-body system called the Supersonic Wave Drive. The user has control over more than simply whether the stabilisation is applied or not, as Olympus has added a second mode. This insures only the vertical axis is affected, so when a subject is being tracked along the horizontal the CCD isn’t needlessly trying to compensate for deliberate motion.

Compact cameras offer a diverse array of techniques to minimise the amount of shake apparent in images. For the most part the technology is utilised in superzoom cameras, as the larger optical magnification makes shake virtually unavoidable. Fuji, for example, uses a high ISO to keep shake down, where as the likes of Panasonic and Sony operate on an in-lens optical stabilisation system. As the lenses aren’t interchangeable, compacts don’t offer image stabilisation across the board, with it mostly only appearing on higher-spec models.

It’s worth mentioning that there is such a thing as ‘digital image stabilisation’. This is most often found in the camcorder market, and it’s quite a different process. The camcorder checks successive frames to detect unwanted movement between them and cancels it out by shifting individual frames. This doesn’t apply in still photography, where you’re only shooting one frame in the first place. The digital image stabilisation systems in camcorders are designed to reduce movement between frames, and not movement blur during each individual exposure.

When to switch off
Image stabilisers aren’t always the answer, however. What’s more, they’re not always desirable. If you’ve ever owned a lens or camera with image stabilisation, you might wonder why you’ve got the option of switching the stabilisation off.

The first, and probably most important, reason is that the gyroscopes themselves introduce a background level of vibration. This probably won’t be noticeable in handheld shots where the ‘stabilised’ image will usually be sharper than a non-stabilised one regardless. However, if you are using a tripod then it’s possible that the gyroscopic vibration may very slightly reduce the image quality. This is a matter for some debate in the photographic community, though Canon’s website does say that there’s a possibility of ‘feedback loops’ when using image stabilisation on a tripod. Nikon has taken the trouble to introduce sophisticated ‘automatic tripod detection’ in its latest products, to reduce the chances of this problem occurring.

Rather less importantly, image stabilisation uses quite a lot of power. If you’re shooting in good light and camera shake is unlikely to be a problem, you’ll get more shots per charge if you leave it switched off. And what if you’re panning with moving subjects? An image stabiliser could cause problems here as the mechanism attempts to counteract the deliberate movement of the camera.

Know your limits
All this discussion about moving objects introduces a very important limitation of image-stabiliser technology. They can stop (or reduce) camera movement, but they can’t stop subject movement, which will also introduce blur.

Sports action is a good example. The only way to ‘freeze’ fast-moving subjects is with a short exposure. You can get away with longer exposures with objects that have a fixed outline (like cars) but not with the flailing limbs of athletes. Sometimes you may want to introduce movement blur, deliberately blurring a sprinter’s legs and arms, for example. But subject blur can only be controlled with shutter speed – image stabilisation is an irrelevance.

You can struggle with subject movement in low light too, of course. While an image stabiliser may enable you to shoot an available-light portrait at an eighth of a second, for example, not all subjects can stay perfectly still during that time. It may work if you’ve arranged a formal session and a fixed pose, but candid shots are another matter. Here, again, there’s no substitute for shutter speed, and this is where you simply have to use fast lenses or high ISOs.
   
Alternative strategies
This brings us on to strategies for coping without image stabilisers. We don’t all have image-stabilised lenses or cameras, but we’ll still want to shoot in the same conditions without camera shake.

The first line of defence is the use of high ISOs as we've already mentioned. Yes, images will be noisier, but it’s better to have a sharp, noisy image than a blurred one. Your high ISO shot will have reduced image quality, but the blurred low ISO version would be headed straight for the Recycle Bin.

Next, make sure you’re shooting at the widest aperture available. Generally, if you’re in program AE mode, the camera will automatically do this in low-light conditions, but it doesn’t hurt to make sure, and you can do this by switching to AP/AV (aperture priority) mode.

Finally, find something to steady the camera against. Bean bags can be useful because they mould themselves to the outline of the camera and whatever surface you place it on, but they are a bit bulky to carry round and offer only one height (the height of whatever surface you can find).

But you can just use a bit of ingenuity. By pressing your camera against door frames, tables or chair backs you can achieve just as much extra stability as an image stabiliser would provide, and maybe more. Alternatively, you can rest the camera on top of a table or any other horizontal surface, then use the self-timer to get sharp shots even with exposures
of many seconds.


Safe shooting
There’s a rule of thumb that links the minimum ‘safe’ shutter speed to the focal length in use. The speed is one divided by
the focal length, or the reciprocal of the
focal length.

Things have got more complicated, though, with the arrival of the digital camera and different sensor sizes. For example, a digital compact might have a 7-21mm zoom, which suggests you should be able to shoot at a seventh of a second at the wide-angle end of the range and 1/21sec at the long end. This isn’t the case. What you need to do is work out the speeds using the equivalent focal length in 35mm film camera terms. In this case, our digital compact has an equivalent focal range of 35-105mm, which gives us more realistic ‘safe’ speeds of 1/35sec at the wide-angle end and 1/125sec at the long end (rounded to the nearest shutter speed).

Digital SLRs generally have a focal factor of 1.6x the focal length on the lens, so that an 18-70mm zoom lens, for example, is approximately equivalent to a 28-110mm (35mm). You can work out safe shutter speeds from this.

Our chart (below) gives you an idea of where this danger area is for any given equivalent focal length. We’ve done the same for image-stabilised cameras, allowing another two stops of shake-free shooting to demonstrate the kind of advantage they have.

In practice, of course, there’s no sharp cut-off point. Our chart should include broad bands of ‘marginal’ speeds rather than narrow lines indicating single values, though attempting to reproduce that in a diagram would quickly get confusing. You can sometimes get sharp shots at half the safe speed or blurred shots at twice that speed. It all depends on the conditions, the camera design and how good you are at holding a camera steady. It’s better to think of a ‘danger’ area rather than a single shutter speed limit.

		You can work out ‘marginal’ shutter speeds using the equivalent focal length of your lens. Image-stabilised cameras give at least a two-stop advantage and sometimes more. The orange line on the graph illustrates 'safe' shutter speeds without image stabilisation. The green line illustrates 'safe' shutter speeds with image stabilisation
		Head along to www.olympus.com for an in-depth explanation of the two-step image-stabilisation system, which can be found on such models as the E-510, and the benefits it has on your photography.
		Nikon’s Japanese website contains technical background information about the company’s Vibration Reduction (VR) technology. This is a simplified version of a diagram demonstrating how the VR works.