Since the start of the 2025-26 season, the Deutsche Fußball Liga (DFL) has been using 3D tracking technology to capture detailed movement data from every Bundesliga match. 

Tracking cameras record player activity from kick-off, generating a precise digital model of each footballer on the pitch.

The system interprets the body of every player as a digital “skeleton”, mapping data points to key body parts including the head, shoulders, hips, joints and feet. The result is a comprehensive dataset that tracks positioning and movement in real time while enabling accurate visualisation and virtualisation of player actions.

Unlike conventional player tracking, which only records a single body point for each field player and referee, 3D data records at least 21 specific body points, and the ball position 50 times per second, generating a total of 140 million data points per match.

According to Dr Hendrik Weber (pictured above), DFL SVP of Sports Technology and Innovation, the technology opens up a wide range of potential applications, including for broadcast analysis. 

“That is the beauty of 3D data. You can fly within the virtual room, and then go into different perspectives and discuss them.”

Speaking to SVG Europe at SportsInnovation 2026 in Düsseldorf, he outlines how the data is captured and manipulated, and how it could be used to support presentation and analysis during and after matches. 

“To collect the 3D data, we are agnostic of the final use case,” he says. “There is a process in place where this is being done for every match, live in every stadium, and the data pipeline is standardised.”

The DFL started by using it for offside decisions for VAR. 

“We really wanted to make sure that this was working because it is a sensitive topic,” continues Weber. “And there’s already some output there for broadcast, in terms of replay. So this [visualisation] already goes into the broadcast product. 

“The data is collected by EA-owned TRACAB Technologies. We have had TRACAB in our stadiums for 12 years, so we have a long track record with TRACAB. They do the collection piece, and for off-site, the replay piece. 

“For visualisation, we used Unreal Engine during POCs with Sky Germany on a couple of games. What we will do now is use the Frostbite Engine from EA, and this is happening as we speak. They used SportsInnovation to showcase it to the world, and now we are in talks with them. 

Find out more about the science behind 3D tracking:

• https://www.dfl.de/en/innovation/automated-event-detection-becomes-a-reality/
• https://www.svgeurope.org/blog/headlines/121921/

In terms of how the data and visualisations might be utilised by broadcasters or by the league itself, things are very much at an early stage.  

“Obviously, the first questions we asked our broadcast partners were: are we seeing this as an additional thing? Is it complementary to what we have? Or is it cannibalising the media product? We are currently trying out a couple of things to maybe get a more concrete idea.

Weber can certainly see how it might be used for analysis, though, and trials are currently taking place.  

“For the halftime show, there’s an editorial team during the first half, picking scenes, preparing them, and then giving this to the [analysis team] on the halftime break. They pick a specific scene, show the [live action] footage for real, then go to the virtual recreation, and then play around with it a little bit. That is the beauty of 3D data. You can fly within the virtual room, and then go into different perspectives and discuss them. 

“It could be near-live or delayed for the halftime show, or used after the game to go into a deeper talk,” he adds. 

There is also the possibility of using the data to create alternative live feeds. “You could use it as an additional broadcast, on a different channel. In Germany, for example, there’s already a separate ‘Tactical Feed’ where fans can see [the game overlaid with graphics on tactics and movement] and there’s also a Twitch feed. 

“So, we could use the 3D tracking data for a Fantasy Football feed, for example. We would use the 3D visualisation and augment this with Fantasy Football statistics, and then it becomes very real for the audience that is focused on that. It is still the POC phase, and we try out and see what works. But that would be a use case which is closest to live.” 

The output from this data could also be used for shoulder content on digital platforms, adding supplementary understanding for VAR decisions, for example. 

“On the digital side, as a league, since we have our own app, we could push out certain sequences, which are of interest. In the beginning, this could be around VAR because we already use visualisations, and this comes quite naturally, if you have a certain handball incident, for example, for more transparency. You could add additional information to help viewers better understand why the VAR changed a decision. I can see that this brings added value to the fan.”

“We jointly need to find out what the format is that brings added value. It’s more the editorial side, like how to pick the scene and how to prepare it so that it actually fits into the media format.”

Like all new technology innovations, the proof of the pudding is in the eating – if it’s not practical or affordable, well, it’s just tech for tech’s sake. 

3D tracking, however, is more straightforward than one might imagine – and the speed at which the visualisations can be rendered creates opportunities.   

“We have been using tracking cameras for 20 years, so the tracking data has been there for a long time. 3D is just a little bit additional, just a couple of more cameras, which is not really a big issue. And then you have 4K or 8K cameras, which we have put in place now – we changed them to a higher resolution, but that is actually not that much of a big issue. The data collection itself for 3D data is two seconds, and for the whole workflow into the rendering piece, it’s six seconds. I’m pretty sure we can cut this down further.

“Hardware-wise, it’s not a problem. We have our fibre network to all the stadiums. So, for 3D data, we push the data to the cloud and then back again. But the computing is done a little bit on the edge, and edge computing on the cameras themselves. And then on the cloud. And then we push it to the use cases.

“There are a couple of stadiums where the stands are too low, and there’s some need for calibration. And if, for example, stadiums have a roof, you open the roof back and forth, the cameras need to be calibrated because of the movement. The workflow, actually, is pretty lightweight. It’s not that complicated.”

The next step for the DFL is to find out the best way to use the data and the visualisations. 

“We jointly need to find out what the format is that brings added value. It’s more the editorial side, like how to pick the scene and how to prepare it so that it actually fits into the media format. 

“We will do more POCs this season, but really we are thinking about preparing for the next right cycle, to see how this can be used as a totally new piece of our rights package we are going to offer the market.”

The DFL is currently talking to various parties, including commentators, presenters and editorial teams, as well as data storytellers and data analysts from Sportec Solutions, to get feedback and ideas for how this tech might be used. 

SportsInnovation 2026 took place 4-5 March in Düsseldorf