What Is Cursor Smoothing?

Cursor smoothing addresses a visual problem that affects almost every screen recording: the raw cursor path, captured directly from OS events, reflects the imperfect motion of a human hand. Even experienced presenters introduce micro-jitter when moving the cursor — small, rapid oscillations of a few pixels that are invisible during live operation because the human visual system filters them out in real time. In a recorded video, however, these oscillations are preserved frame by frame and become visible as a trembling, unsteady cursor path that makes the recording look unprofessional and distracts from the content.

The technical solution is to apply a low-pass filter or moving average to the cursor position time series before rendering it to the output video. A simple moving average smooths the path by averaging a window of recent positions, reducing high-frequency oscillations at the cost of introducing a slight lag between the actual cursor position and the rendered position. More sophisticated approaches use predictive filtering (such as a Kalman filter or exponential weighted average) to reduce lag while still removing jitter, or apply smoothing only during motion and restore full accuracy when the cursor is stationary — preventing drift away from precise click targets.

Limelight applies cursor path smoothing as part of its rendering pipeline. The smoothing is calibrated to remove visible hand tremor while preserving the intentionality of deliberate cursor movements, so a slow glide across the screen looks graceful and a quick flick to a button still lands accurately. The smoothed path is used by both the cursor spotlight (which moves with the smoothed cursor) and the auto-zoom system (which uses the smoothed position to anchor zoom frames). The result is that every cursor movement in a Limelight recording reads as deliberate and professional, even when recorded casually in a single take.