Wild things

For natural history film-making, look to Britain. The UK’s wildlife and nature TV programmes, spearheaded by the BBC’s renowned Natural History Unit in Bristol and fuelled by the indefatigable David Attenborough, are seen and admired around the world and each year brings a new batch of breathtaking imagery.

While the expertise, patience and daring of the camera operators plays a large part in this, none of it would be possible without the input of technology.

The development of cameras and techniques for specialist filming, particularly in wildlife and sports broadcasting, is a story of ingenuity and adaptation. In most cases, the cameras are off-the-shelf products adapted and combined in ways that have, in some cases, been developed over several years and in others, cobbled together on the fly.

They represent a prime attribute of cunning engineering: the ability to take technologies designed for one application and bend them to produce a result its developers would not have dreamed of.

‘A lot of the things we use, such as super-high-speed cameras, are developed for something completely different,’ said Ed Charles, a BBC Natural History Unit researcher, who works with camera operators and tracks down the technologies for assignments. ‘As another example, we use thermal imaging cameras, which were designed for industry, mainly for security firms. But then we have to adapt them to work on a remote basis.’

Charles is working on a series due to be shown early next year, called Nature’s Great Events. Each of its six parts will focus on a change in natural conditions that has affected local wildlife: the annual melt in the Arctic; the salmon run in British Columbia; the plankton bloom off the coast of the Pacific Northwest; the greening of the Serengeti; the sardine run off the coast of South Africa; and the flooding of the Okavango Delta.

The last of these, said series producer Karen Bass, was a good example of the use of technology in filming.

‘Our big mission on this series is to show epic transformations of the planet, but through the intimate stories of the animals that live there,’ she said. ‘The Okavango Delta is in the middle of the Kalahari Desert, but it floods once a year. Ten million cubic kilometres of water comes down from Angola and seeps into the Kalahari. It’s not a fast flood — it trickles. The water moves at a metre an hour.’

Charles, who filmed the Okavango with camera operator Martin Dohrn, said: ‘The water was the key thing. We decided we’d always travel with the water and show it at all its different levels.’

Martin Dohrn, who runs independent production house Ammonite Films in Bristol, has a reputation for developing camera technology. Ammonite is known for two products: starlight cameras, which use image intensifiers to enable filming at night without using infrared or any additional lighting; and motion control camera systems, which allow cameras to make precisely-controlled movements to follow events on a small scale.

It was the latter that attracted Charles. ‘It’s not that common for a camera operator to develop their own cameras, but Martin is unique,’ he said. ‘They will generally just hire some cameras and go and shoot with them, but Martin is a different breed.’

Dohrn says developing cameras is a throwback to an earlier era of film-making. ‘There’s such a lot of kit now available, and there’s a perception from young people entering the field that you buy your camera then the important thing is your own ability in framing and getting the exposure with these tools. Twenty years ago film offered plenty of possibilities, but there were many more possibilities if you were to meet an engineer and have lenses or mounts built for you.’

The technology that took Dohrn and Charles to the Okavango was the Frankencam — a macro camera capable of filming extreme close-ups of tiny subjects, which is mounted on a head linked to a complex system of motors that allows it to move in any direction. This system, known as motion control, was developed for special-effects-heavy filming, where many elements have to be combined into a single shot and each filmed with identical camera moves. Dohrn has adapted this system for filming insects.

‘It really has evolved,’ he said. ‘It started off as a one-axis motor to do time-lapse tracking, looking into a fishtank. All it could do was go backwards and forwards at a set speed. Then we added two axis movement, then three; and we realised that even three axes was impossible for a normal human to control; the effect of the rotating axis completely melted your brain. So we invented electronics that allowed the three axis motors to talk to each other, so the operator didn’t have to work out which way the axes was pointing; you could just drive the camera by looking at the monitor.’

At this point, the system was known as Ant-cam, because Dohrn was using it to film the behaviour of ants within a hive. ‘That took about five years to develop from the first version, and it was a great device; you could put a miniature camera on it and drive it anywhere within a 2m cube. But because of the motors on it, it wasn’t very good at precision. It was good at speed, but not so good at stopping.’

This led to the further development into Frankencam, a three-axis system using a camera mount stabilised by counterweights and moved by DC drives from Swiss firm Maxon Motor. ‘The controller has a double-joystick set and a focus controller. You make sure all the motors are set to zero, turn on the power, then it’s much like driving — move the right lever to go forward or backward, the left lever to steer. You can follow a woodlouse with it — it’ll look like it’s a couple of feet long — go around to look at its head, start off behind it, overtake and track back so it walks towards the lens. You can do all the sort of shots you’d take on a large scale with a helicopter, but just six inches above the ground and looking at things a millimetre across.’

Dohrn added: ‘We always refer to Frankencam as ground-breaking equipment. That’s because, quite often, it hits the ground and breaks.’

The Frankencam at work

Unlike conventional motion control systems, Frankencam never has to repeat a movement, but its control system gives its operator the ability to follow the movements of an unpredictable element such as water flowing through sand, or the erratic movements of insects, precisely. ‘We see the water as a life-giving force, and we have a wonderful sequence of a dragonfly coming out of the sand,’ said Charles.

According to Dohrn, the driving force for developing new camera techniques is usually an ‘application push’ — a desire to film animals in a new way. Once a system is developed, there tends to be a ‘technology push’ — a need to find other things that can be filmed with that system. This, in turn, leads to further development.

Ammonite’s Starlight cameras are a case in point. ‘We wanted to be able to film without lights of any sort, because it would allow us to film behaviour that we wouldn’t be able to see otherwise,’ he said. ‘So we bought camera modules that were designed to work with a particular imaging chip, and persuaded an image intensifier company to bond its intensifier to that chip, connected it to the camera module, and build a body and control system and an output system around that.’

The Starlight system is now in its third incarnation, with the latest version, which Dohrn is working on now, capable of filming in high definition.

‘Intensifiers have improved massively in the last decade, in terms of definition and in sensitivity, so we can make a much better system than we could before,’ he said.

‘Optically, we make up compound lenses from off-the-shelf units. If we had a lens design partner, we could tell them what end result we wanted and get them to grind lenses for us; I’m sure they could get the results we wanted using just 10 lens elements. Ours tend to have at least 25 — we’d be able to get them much lighter.’

The techniques developed by natural history film-makers can have spin-offs into research. For example, the 2004 BBC series Animal Camera was the inspiration for Oxford University zoologist Christian Rutz to develop a system to observe unusual behaviour in the New Caledonian crow, which lives in rainforests in the South Pacific, and uses tools, such as twigs, to help it forage for the insects it eats. ‘They’re very shy, they avoid humans, and they live in heavily forested areas,’ said Rutz. ‘Then we saw Animal Camera, which used cameras mounted on animals to get their view on the world, and we thought: wouldn’t it be wonderful to attach little cameras to the crows themselves?’

It was not a simple job. Rutz contacted the programme’s camera operator Jonathan Watts, who had developed harnesses to attach cameras to captive animals that had been trained to wear them. Attaching a camera to a wild bird was a different matter. ‘Even his smallest units were too big for our crows, which only weigh about 320g, the size of a small pigeon. The smallest units were 25g, and we needed to get that down to 15g, which is five per cent of the body mass.’

The team’s first task was to research tiny video cameras, and they found the mobile phone market a fruitful source of equipment. ‘There’s a massive commercial interest in very small video technology,’ said Rutz.

Watts’s production company, British Technical Films, then developed trigger chips that would switch on the cameras at a set time after they had been attached to the tail feathers of captive birds.

‘The timing was crucial,’ said Rutz. ‘We don’t film the behaviour of an animal that’s just been trapped and handled; all you’d get is the crow trying to get rid of the camera. So the key to the success of the project was to send the camera to sleep for a while until the crow had habituated to it.’

The system also included a microwave transmitter system, and Watts developed a helical antenna to fit inside the parabolic dish of the receiver to maximise the received signal. ‘Because it is microwave you get a lot of interference with vegetation and you often lose signal. But these birds are so shy and elusive that you’re lucky if you capture them once; it’s very unlikely you’d capture one and get the camera back. Also, we couldn’t include data loggers in the camera tags because of their weight; transmitters are much smaller.’

Rutz, Watts and their colleagues managed to reduce the weight of the camera tags down to 14g, and were the first team to attach video cameras to wild, flying birds. Now the team is working on new camera tags that use lighter memory cards to record longer film segments.

‘We’re making video loggers that store the video on micro-SD cards,’ Rutz said. ‘We’ll attach those to the innermost tail feathers of the crows, because we know that they moult those feathers within a six-week window. The tags will include a radio beacon, which lives for eight to 10 weeks, and that will give us a good chance of finding and recovering the feathers and the tags.’

Meanwhile, the BBC’s research unit is looking for new ways to make images, and again showing how adapting existing technology can open up new ways to make images. A team led by Jigna Chandaria is combining a new type of hardware, a camera with a fisheye lens, with some proprietary software to steady the images generated by a camera mounted on a referee or a jockey in the middle of a match or race.

‘Virtual steadicam’, as the system is called, came from a system that the BBC developed to combine computer graphics with live-action studio or sports broadcasts. ‘The software tracks the position of the camera,’ said Chandaria. In the studio, this allows graphics such as the ‘swingometer’ to be incorporated into General Election broadcasts, and appear as though it is in the studio. ‘One system uses markers on the ceiling of the studio, then the cameras have a little secondary camera pointing straight up. This allows the position and orientation of the camera to be known precisely, and the graphics department can move their graphics around the screen exactly as though it were being filmed by the camera.’

This system was extended to a package called Piero, which is used to analyse images from football and rugby matches by looking for the lines drawn on the pitch. Using this information, graphics can be added to pick out players or highlight the ball as it moves around the pitch. ‘The next step after this is to look for points of texture, which allow you to pick out the points of particular interest in the image,’ said Chandaria. ‘So we combined that with the image from a fisheye lens with a 180° field of view.’

This was mounted onto one of the research team, who ran around, producing a jerky, unwatchable image. ‘We extract a portion from the middle of the field of view, using points of texture to pick out the interesting part of the image, then we can produce an image that appears to have a 60° field of view, compensate for the motion of the camera and remove the distortion.’ What is left is a smoother, gliding motion that is comfortable to watch.

The system is not yet ready for use, however. ‘We just used a Firewire camera,’ said Chandaria. ‘We’d need a broadcast-quality high-definition camera, which is small and robust, and we haven’t found one yet. That would be a next step, if we get funding.’

The team is also trying to stabilise vibrating images that have been recorded by cameras mounted on ski-tips. ‘If we find that we can make a significant improvement, it’ll be worth investing some time in looking for a suitable camera to take this further. But we’ve proven it’s feasible.’

Chandaria has already received some interest in the system. ‘The obvious applications are for sport and news, where you can mount the camera on someone rather than taking in a camera crew, in areas where that isn’t possible. But children’s TV is also interested; you could use them to give a child’s eye view. And Natural History has expressed an interest as well, for animal’s eye viewing.’

Chandaria echoed Martin Dohrn’s comment, that the technology development usually follows a desire by programme makers for a particular sort of filming. ‘Ultimately, it’s the desire to film something that you haven’t seen before, and finding a new way to do that,’ Dohrn said. ‘That’s the driving force for us.’

Stuart Nathan