Skip to main content

New Update, New Standards: Ochy’s Speed Estimation Now More Precise Than Ever

Khaldon avatar
Written by Khaldon
Updated over a week ago

A New Dimension in Running Analysis

With this update, Ochy takes performance analysis to the next level: the app can now estimate your running speed , directly from your video. It’s one more step toward a complete understanding of your technique — no sensors, no lab required.

Spatio-Temporal Metrics: Scientifically Precise

Thanks to extensive research, Ochy now delivers scientific-grade precision on key metrics:

  • Step frequency : 98.85% accuracy

  • Ground contact time : 88.71%

  • Flight time (air time): 93.04%

In short: the data you receive is reliable enough to truly support your progress, with impressive results achieved from simple smartphone footage.

Joint Angles: The Knee Leads the Way

When it comes to kinematics, the knee joint shows the highest level of accuracy , comparable to results from advanced 3D motion studies. Hip joint analysis also performs well, while ankle and elbow readings are more impacted by the limits of 2D pose estimation — a known area for future improvement.

What’s remarkable is that all of this precision comes from casual video recordings — not high-speed cameras or treadmill labs , proving Ochy’s ability to make elite-level tools available to everyone.

What You’ll Notice in the App 📲

This update isn’t just about numbers — it also brings noticeable improvements to your everyday experience:

  • 🏃‍♂️ Your running speed is now displayed directly in your analysis

  • 🧠 New exercises appear when your scores are strong (side view)

  • 📋 Exercise recommendations are now embedded directly in your report

  • 🏪 In-store analyses are automatically shared via email with athletes (exclusive to running store accounts)

Download Ochy and Discover Your Stride

Whether you’re just getting started or chasing a new PR, this update gives you powerful new tools to better understand your running, improve smartly, and reduce injury risk.

Did this answer your question?