Motion blur is the smearing or streaking of a moving object within a single video frame. It occurs because each frame is exposed for a period of time, and anything that moves during that exposure is recorded across several positions at once.
The amount of blur depends mainly on how fast the subject moves across the image and how long each frame is exposed. A hand or foot moving quickly near the camera blurs more than the trunk moving slowly in the distance.
Factors That Increase Motion Blur
- Longer exposure times, often chosen automatically in low light
- Fast-moving segments and equipment such as the hands, feet, club, bat, or ball
- Filming close to the subject, which increases movement across the frame
- Camera shake from handheld recording
Frame rate sets an upper limit on exposure time, because a frame cannot be exposed for longer than the interval between frames. Higher frame rates therefore tend to allow shorter exposures, but a phone may still lengthen exposure as far as it can when light is limited. This is why indoor gym footage often looks blurrier than outdoor footage, even at the same frame rate.
Why It Matters for Movement Analysis
Blur makes it harder to place landmarks, mark joint centers, and identify exact events such as foot contact, ball release, or impact. Pose estimation and joint tracking can also become less reliable on blurred frames. Note that camera shake blurs the whole image, while subject blur affects only the moving parts.
Practical Tips
- Add light or move to a brighter area; light is often the single most effective fix
- Use a higher frame-rate mode when the lighting supports it
- Mount the device on a tripod to remove camera-shake blur
- Use a camera app with manual shutter control if a specific short exposure is needed
- Position the camera far enough away that the movement does not sweep across the whole frame
When some blur is unavoidable, choose landmarks consistently, for example near the center of the streak, and avoid drawing firm conclusions from a single blurred frame. Comparing several neighboring frames usually gives a better picture of position.