Even the most controversial umpire could not call this one out © Hawk-Eye Innovations

"At least about one in five lbws that are given by the on-field umpires are incorrect," says Dr Paul Hawkins, the 30-year-old managing director of Hawk-Eye, and the brains behind the technological device which has proved to be such a boon to television coverage and such an eyesore - generally - to the umpires. I am sitting in Sky TV's van, from where Hawkins and Gary, his colleague, operate and give inputs to the director of the telecast team.

It's an impressive set-up. There are 19 miniature TV screens showing various angles of the actual game and Hawk-Eye's simulation. (Actually, only 18 are dedicated to the cricket: one of the screens is tuned to Great Britain's Davis Cup tennis match against Austria.) There, between feeding the Sky team with data, Hawkins explains how the system works: there are six cameras, at long-on, long-off, fine leg, third man and two square of the wicket.

Only three of the cameras are used to capture the motion of the ball from each end - those at long-off, long-on and one of the square ones - but these are special cameras, capturing data at the rate of 120 frames per second, compared to the 25 frames per second by regular TV cameras. The camera square of the pitch only captures the last third of the ball's journey towards the batsman, ensuring that any deviation in the air or off the pitch is captured accurately.

Then, all this is fed into a central computer, which extrapolates this information to predict the path of the ball. There is also the wicket-to-wicket pitch mat which, combined with Hawk-Eye's analysis, gives the viewer a pretty good idea of whether a batsman was lbw or not.

Most of the time, at least. Soon after I am in the room, Wavell Hinds bowls a gentle in-dipper which Paul Collingwood plays forward to. However, the ball clearly strikes pad first. "That's plumb," both Hawkins and I say in unison. Sure enough, technology proves us right - the ball strikes pad between middle and off, and is probably clipping middle. Hawkins immediately feeds this information in to the director of the telecast who, however, chooses to ignore the data - probably because there was no appeal from the West Indians - leaving only the three of us privy to the information that Collingwood got a life in his brief innings.

To further demonstrate the accuracy of the system, Hawkins shows me the Hawk-Eye image of a ball which just misses Geraint Jones's outside edge. The simulation is almost exactly like the real thing - the ball pitches, shapes away, and goes past the bat. I am impressed.

The Sky van is only the front end of Hawk-Eye's operations, though. The actual processing of data happens in Hawk-Eye's own van, where Matt and John monitor the data after every ball, and ensure that nothing's amiss. "Paul gets all the credit," jokes Matt, "but we're the guys who do the groundwork." He then proceeds to explain to me why the cameras are placed where they are, how they identify and track the ball, and how data from all cameras are processed to get the actual mapping of the ball's trajectory.

"Today's weather is just ideal for us," says Matt. If that comment had been broadcast over the public-address system at The Oval, the chances are that he would have been lynched by the spectators freezing to death in the cold, but Matt soon explains what he meant: "Sometimes on a sunny day the shadow of the ball may hinder the tracking of its flight, but on a day when the lighting is diffused, as it is today, there are no shadows to worry about."

What happens on a sunny day, then? "We need to instruct the cameras to track the ball, not its shadow." That, he says, is something they can overcome by tuning the system between the real ball and the shadow. On a regular day, there may be one or two deliveries which are tracked incorrectly and need to be rectified, but so far today, the team has a 100% accuracy record.

How have the umpires reacted to this technological innovation? "Some of them have come to me and tried to find out more about it, but most of them have their doubts," says Hawkins. The system is not totally foolproof - on a rare occasion, one of the cameras may not work, says Paul - but most of the time, the results are extremely accurate, he assures. It has certainly added to the TV coverage; how long before the umpires benefit from it as well?

S Rajesh is assistant editor of Wisden Cricinfo.