Speed development tools, ranging from traditional stopwatches to timing gates to GPS devices, have evolved and proliferated over time. Recently, the next evolution of toolsets has emerged—phone-based 2D kinematics powered by computer vision. These tools initially sprang up in areas such as baseball and golf (swing and throwing analysis) but are now making their way into the speed ecosystem to help provide kinematic analysis.
The next evolution of toolsets has emerged—phone-based 2D kinematics powered by computer vision—making their way into the speed ecosystem to help provide kinematic analysis, says @hi_c88. Share on XToday, if a coach is interested in the underpinnings of how their athlete generates their speed or time, there are a few options. The first and most timeless option is the simple combination of video and a coach’s eye/intuition. While this approach is enduring for a reason, it relies on years of experience and does not generate objective data.
The second option is for a coach to invest significant time and money in a multi-camera 3D-markered or markerless solution and find people/companies to help process that data. While this option does provide the highest quality of data (it is essentially “lab-grade”), the cost for equipment alone can be more than $50,000, and the barriers to processing that data are high—requiring at least one person with a graduate-level degree in biomechanics and a person with significant computer programming experience. Additionally, the setup time and portability of this type of system are quite cumbersome, limiting its practicality.
The third option, and probably the most prevalent for those trying to generate objective data, is to use “enhanced video” capture tools like Kinovea or Dartfish. These tools allow coaches to capture video and generate data by manually tagging events throughout the video, providing the building blocks to calculate the desired metrics. The downside of this option is that it’s a manual, drudgery-filled process—meaning that for most coaches, it truly can’t be scaled for use with an entire population, rendering it more of a “one-off” type of activity than something that can be routinely incorporated into a practice, like timing gates or GPS devices can.
With the recent advancement of computer vision, there is now a fourth option for coaches that provides its own set of advantages and disadvantages—single-camera 2D kinematics tools. This past September, I helped build and launch one of the first of these tools: BreakAway Speed, an iPhone-based, 2D, markerless biomechanics capture system. The general setup concept is table-stakes among the various solutions—videos of athletes are captured at 240 fps to ensure the highest fidelity, and distances are calibrated using either cones or football field lines. Users can perform multiple types of tests—flys, accelerations, or changes of direction—and receive their results back in just a few minutes.
There are multiple benefits to incorporating this type of solution into your coaching toolkit (or paired with existing technology, i.e., timing gates) that I want to share here, chief among them being:
- Automated kinematic analysis.
- Ease of data-sharing with athletes.
- Ease of use with a software solution.
Automated Kinematic Analysis
As discussed, for most coaches, kinematic analysis has simply not been accessible due to significant barriers to entry—namely, the expensive financial costs of hardware or untenable time investment needed to get data at scale. Coaches can invest in expensive 3D capture systems such as Vicon, Qualisys, and Theia, which provide gold-standard data but are typically impractical cost-wise—not to mention they present significant constraints operationally. I’ve used fantastic software like Kinovea to create kinematic data manually; however, from experience, doing that at any level of scale becomes a significant time investment.
A new third choice is beginning to appear—headlined by products like VueMotion, Ochy, and my product, BreakAway Speed. These products offer a simpler version of the expensive 3D setup mentioned above.
With one camera view, AI tracks each point on the body (heel, toe, knee, elbow, etc.) for each video frame. These points are then turned into sprint-specific kinematic metrics (like stride length) using biomechanical calculations. While this single-camera setup will never be as accurate as eight-camera marker-based pose detection, proponents of these solutions feel that the slight decrease in kinematic accuracy is outweighed by the dramatic decrease (more than 100x less) in price and accessibility. Detractors of these solutions may argue differently.
However, coaches now have access to an entirely new set of ways to quantify how their coaching is impacting a runner. By leveraging this new, objective information, they can optimize sprint techniques and monitor the effectiveness of coaching interventions. Coaches can receive metrics they currently get from timing gates/GPS devices (things like max speed and time), along with new metrics such as stride length, contact time, thigh angle separation, and more.
We’ve even seen customers utilize both technologies simultaneously; using timing gates for that truly instant feedback (<1 second) in conjunction with BreakAway Speed allows for instantaneous feedback in addition to metrics explaining how that time occurred just a few minutes later. Because tools like BreakAway Speed cannot capture a sprint distance greater than 20 yards due to AI limitations, users can still truly capture an entire 40-yard dash with timing gates while setting up a tool like BreakAway Speed to only maybe capture the 20- to 30-yard split or the 0–10 start, giving a more detailed window into how that 40-yard dash time occurred.
Additionally, asymmetry metrics are included to better help with return to play and rehab scenarios. These metrics generally take a few minutes to generate, so while they don’t have the immediacy you would get from a timing gate solution, they are typically processed quickly enough to still allow for on-field feedback.
Lastly, because these solutions are video-based, coaches can blend what they see with their eye from video analysis with the objective data generated by the AI, fusing the time-honored tradition of video review with more novel, data-driven approaches.
Facilitating Data Delivery to Athletes and Education
A significant benefit of these types of solutions is that a coach only needs their phone to record and receive data. This makes life easier during the “data capture” part of a coach’s workflow. Another benefit is that, by being phone-based, the distribution of this data to their athletes can also be enormously simplified. With these solutions, coaches can simply assign videos to their athletes on the application and enter their phone numbers to have their data automatically delivered to the athlete.
When data is processed, it is then delivered to the coach and the athlete simultaneously! Athletes receive all their data and a social-media-friendly video to show off how fast they ran.
Video 1. Example of BreakAway clip sent to an athlete.
With this type of automated coach-to-athlete connection, data delivery doesn’t just go from coach to athlete but also from athlete to coach, enabling scenarios for remote coaching or digital evaluations to be much more efficient. Because this data is rather complex for anyone to digest fully, these tools can incorporate simple AI-powered explanations of their data to better help athletes (and coaches) with the education process on how to apply this data.
Data delivery doesn’t just go from coach to athlete but also from athlete to coach, enabling scenarios for remote coaching or digital evaluations to be much more efficient, says @hi_c88. Share on XSoftware, Not Hardware
One challenge with solutions such as GPS devices and even timing gates is the simple fact that they are hardware. Especially with GPS devices, they require non-trivial operational work such as tagging, charging, setting up, and distributing devices, in addition to dealing with vests. To be clear, these tools certainly have their place, drive significant value, and can be used in conjunction with 2D kinematic tools.
The simplicity of an iPhone-based app enables easy supplementation of these tools—athletes can be assessed while wearing a GPS device or by using Freelap cones as the markers. Having multiple layers of applied data can then give a more holistic evaluation.
Lastly, because these tools only require your existing iPhone or iPad, they can be offered at a more affordable price, comparable to having a Netflix subscription rather than purchasing an appliance.
While still early, automated 2D kinematic analysis tools for fundamental athleticism are here and ready for utilization by coaches. As AI and computer vision continue to improve, these tools will only get more accurate, improving the quality of data generated and steadily closing the gap between single-camera and multi-camera solutions.
Additionally, as the technology and businesses based on this technology evolve, the scope of movements that can be analyzed will also increase. Currently, most solutions focus on linear running, but in the near future, users will be able to analyze hurdle jumps, triple jumps, broad jumps, and vertical/drop jumps, just to name a few. The template is built—and as it progresses, it will go both deeper (increased accuracy) and wider (more movement types).
Today, though, by unlocking kinematic analysis, facilitating data sharing, and being an easy-to-deploy/purchase software, a tool like this can drive significant value for coaches looking to make their first investment or add on to their existing technology stack.
Since you’re here…
…we have a small favor to ask. More people are reading SimpliFaster than ever, and each week we bring you compelling content from coaches, sport scientists, and physiotherapists who are devoted to building better athletes. Please take a moment to share the articles on social media, engage the authors with questions and comments below, and link to articles when appropriate if you have a blog or participate on forums of related topics. — SF