Technical FAQ: Finish-line cameras explained

Published May 11, 2021 07:12AM

Have a question for Lennard? Please email him at veloqna@comcast.net to be included in Technical FAQ.

Dear Lennard,
I am a huge Ruth Winder fan, and I was super excited to see her win Brabantse Pijl, especially since she beat Demi Vollering, who seems to be the next (Dutch again!) superstar after Marianne Vos and Anna van der Breggen. But I have watched the video of the finish over and over, and it really looked like Vollering won. I remember you wrote about how finish-line cameras work a few years ago. Is it a technology thing I’m not understanding for how the judges declared Winder the winner?
— Linda

Dear Linda,
It’s possible we’ll see a super-close sprint during the Giro, so this is a good time to clear this up. Yes, that seeming difference can be explained. Almost a carbon copy happened a few days later at Amstel Gold! Many a fan watching replay after replay was certain that Thomas Pidcock won there, rather than fellow cyclocross racer Wout Van Aert, even though it is clearly visible (perhaps reading glasses are required, it was so close) that Van Aert’s tire touches the line on the finish-camera photo ahead of Pidcock’s wheel.

First of all, it’s best as a watcher of these things to not question the finish-line camera and rather to question whether what you are seeing on video or still images is distorted. That finish-line camera is (or, more accurately, those cameras are) the actual finish, not the lines taped on the road.

Secondly, it’s important to understand how a finish-line camera works, and I explained that as well as I could after a tight Tour finish in 2016. I think it’s worth a review to understand the rest of what I’m about to say.

We are all used to seeing a photo as an instant in time. We understand that the positions of the objects in the photo are their positions relative to each other at that instant. Regardless of whether the objects are separated vertically or horizontally in the photo, it accurately represents their relative positions.

A finish-line camera photo, however, is totally different; it is a snapshot of position, not time. The positions of the objects in the photo that are separated horizontally do not represent their relative positions at an instant in time; only objects lined up vertically with each other are shown at the same instant in time.

I realize that this may be a hard concept to wrap one’s head around because it is so different from how photos normally work, and, indeed, how our eyes work. The entire photo is, literally, the finish line. To determine who got to the line first, a vertical line is drawn on the finish-camera photo, because the finish line that is painted or taped across the road is not visible in it as a line. Rather, the edge of the finish line is that entire smear below the riders’ wheels from the left edge to the right edge of the photo. Since neither the taped line on the road nor the camera is moving, and the “film” in the camera is, the finish line is in the entire photo from edge to edge.

A way to convince your mind that the entire finish-camera photo is of the finish line is to notice that some riders in the photo are throwing their bikes despite being well behind those sprinting for the win. It makes no sense that riders would throw their bikes until just before the finish line, so why would they be doing that? And, indeed, if we watch the video, we can easily see that riders sprinting for minor positions far behind the winner do not throw their bikes until just before the line. The reason those far-back riders are throwing their bikes in the finish-camera photo is that they are at the finish line. The camera is looking through a vertical slit lined up with the finish line, and riders do not appear in the camera’s aperture until they are at the finish line.

A normal camera has an aperture that opens and closes concentrically to allow light in to expose the film for a brief instant in time (it’s easier to think of film being exposed by light than it is to think about digital pixels being exposed by light, so I’ll talk about film rather than pixels). A finish-line camera’s vertical aperture, however, stays open the entire time the finish is being recorded. The film moves across behind the aperture in the same direction as the riders and at roughly the same speed. If the camera operator sets its speed faster than the riders are moving, then each part of their body and their bike goes across the aperture slower than the film does, so they appear stretched out. Similarly, if the camera operator sets its speed slower than the riders are moving, then each part of their body and their bike goes across the aperture faster than the film does, so their image is foreshortened. As I mentioned in a previous column, in the rare instance when a breakaway rider is swarmed at the line by a group of sprinters, that rider’s bike appears to be much longer in the photo than the bikes of those sprinting past him. His or her bike is certainly similar in length, and maybe exactly the same actual length, as their bikes; it only looks longer in the photo because his or her bike moves across the aperture more slowly than theirs do.

The rulebook defines the finish line as being the leading edge of the black line centered between the broader white lines, although the reality is very slightly different. In order to align the finish-line camera with the edge of that black line, it has to actually be lined up very slightly behind it, so that it can “see” a slight bit of white on either side of its alignment. This is such a trivial difference that, even in the case of these two super-tight races, it probably is irrelevant.

There is also a photo doublecheck. For races of this caliber, the UCI mandates two finish-line cameras, one on either side of the road pointing directly at each other. Their outputs also go into separate computers, so that if any one camera or computer has a glitch, all necessary information will still be available to determine the winner. Whenever finish photos are published for viewers to see, only a single one, from one side, is shown. But I imagine that, in the case of a really tight finish, part of the delay in getting the official results out may be that the race jury is consulting the second camera’s image taken from the other side of the road as well.

The people who line up the finish-line cameras and lay down the tape of the finish lines are probably surveying with similar accuracy to those determining the altitude of the top of Mt. Everest. The only way I could imagine that the alignments of the taped line and of the cameras’ lines of sight could be off would be in the case of a road that is narrowing down or flaring out at the finish. Then, a line perpendicular to one side of the road will not be parallel to a line perpendicular to the other side of the road. Still, I imagine that when the construction crew places the barrier-fence panels near the finish, they make sure they are parallel on opposite sides of the road, and the finish line would be laid down perpendicular to them.

This answer thus far has all been to assure you that the finish-camera photo accurately depicts the winner of the race. As for why that may differ from slow-motion or still images taken from a video, the curvature of the wide-angle lens of the video camera on a TV broadcast creates a slight amount of distortion. In all but the closest of finishes, a freeze-frame still image from the TV camera will still show the correct winner. But in both of these race finishes, the TV camera’s distortion was enough to apparently show a different winner from the actual one. Even though it is probably not possible with the layout and security at the finish of a Giro stage, if there happened to be a still image taken from a normal camera, even one looking straight across the line at the moment the winner’s tire hit it, like the ones in the screenshots of tweets at Amstel Gold in this article linked above the same kind of distortion could also lead to misidentifying the winner.