I’ve been using a Freelap setup for short sprint splits in training, and I kept noticing the same pattern: my Freelap 10 m times looked about a tenth of a second slower than times estimated from video. Rather than guessing, I ran a small field test to quantify the difference.
What I tested
I compared two timing methods across 10 block-start trials:
- Freelap timing:
- Start: Freelap e‑Starter at 0 m
- Finish: Tx Junior Pro positioned at 10.80 m, following Freelap’s guidance for timing a nominal 10 m split (the transmitter is placed beyond the split to account for detection-zone geometry).
- Video timing:
- iPhone video at 60 fps
- A simple sound-to-light Arduino trigger near the e‑Starter created a visible “time zero” flash on the start sound.
- Finish was taken as the frame where the torso (chest region) crossed a marker placed at the 10.00 m line.
This is not meant to be a “laboratory-grade” validation—just a practical check using the same tools and definitions I use day-to-day.
The result
Across the 10 trials, Freelap was consistently slower than video.
- Mean difference (Freelap − Video): +0.089 s
- 95% confidence interval for the mean: 0.075 to 0.103 s
- Trial-to-trial spread of the difference (SD): 0.019 s
In other words: in this setup, you can expect Freelap to read roughly 0.09 s slower (or for quick reference simply 0.1s) than this particular video timing method, and it was quite stable across trials.
Interestingly, 0.80 m is a meaningful distance scale: at typical sprinting speeds, the time to cover 0.80 m can be on the order of ~0.09 s, which is very close to the observed offset. So perhaps the remedy is to simply place the Freelap Tx at 10 m instead of at 10.80 m.
A quick note on the start flash and speed of sound
The Arduino trigger was placed about 5 cm from the e‑Starter. Sound travel over 5 cm is about 0.15 ms, which is negligible compared with an ~0.09 s offset. Any meaningful start-side uncertainty is more likely due to sensor thresholding and the fact that video timing is limited by frame resolution.
Practical takeaway
If you’re comparing Freelap 10 m splits to video timing, don’t be surprised if they don’t match perfectly—even if both are consistent. In my setup and definitions, Freelap read about 0.09 s slower than video.
If you care about comparing across methods, the best approach is to:
- stick to one method for tracking progress, or
- apply a consistent offset (with the caveat that event definitions still differ), or
- adjust your setup so the finish definition better matches the reference you care about.
Full report: Validation of Freelap Sprint Timing Against Video Timing for 10 meter Block Starts
Arduino device with sound-to-light converter.
Video snapshot at finish showing torso/chest alignment with marker. This snapshot is from the 9th and last trial which is recorded as 2.18 s (video) and 2.26 s (Freelap). The tablet shows the Freelap time recorded in the previous trial (2.36 s).