Filters¶

Filters apply signal processing to telemetry channels. Use them to smooth noisy signals, compute derivatives, or transform data.

Filter Methods¶

Savitzky-Golay¶

Polynomial smoothing filter that preserves signal shape better than simple averaging.

Window — number of samples in the smoothing window (must be odd)
Poly Order — polynomial degree (lower = smoother, higher = more detail)

Best for: general smoothing while preserving peaks and valleys.

Butterworth¶

Classic frequency-domain filter with flat passband response.

Tau — time constant controlling the cutoff frequency
dt — sample interval

Best for: removing high-frequency noise with a clean cutoff.

EMA (Exponential Moving Average)¶

Simple recursive smoothing where recent samples have more weight.

Tau — time constant (higher = smoother)
dt — sample interval

Best for: fast, lightweight smoothing.

One Euro¶

Adaptive filter that smooths slow movements but preserves fast ones.

Min Cutoff — minimum cutoff frequency
Beta — speed coefficient (higher = less smoothing during fast changes)
D Cutoff — derivative cutoff frequency

Best for: signals that alternate between steady-state and rapid changes.

Integration¶

Computes the running integral (cumulative sum) of a channel.

Initial — starting value for the integral

Differentiation¶

Computes the rate of change (derivative) of a channel.

Passthrough¶

No filtering — passes the signal through unchanged. Useful as a placeholder.

Configuration¶

In the workflow panel:

Method — select the filter type
Branch Mode — apply to all branches or a specific one
Channel Types — toggle which channel types to filter (Raw, Pred, Val, AuxMath)
Parameters — adjust filter-specific settings

Tip

The Denoise toggle in the workflow panel is separate from the filter node. Denoise corrects timing errors in IBT files before any other processing — it is not a signal smoother.