Answers are grounded in this paper and cite their sources.
A causal smoother that removes its own lag pays for it in transient shape: the step response overshoots, and the usual state-space route to zero lag, the constant-velocity Kalman filter, rings. We construct a fixed-weight family of zero-lag smoothers whose step response overshoots exactly once and never rings. The construction is a single ladder of identical exponential low-pass poles, read at its cumulative depths and combined with the unique affine weighting that fixes unit gain and zero time lag: the Linear Pole Moving Average (LPMA), whose two-pole member at first order is DEMA (Mulloy's 2 EMA - EMA^[2]). We prove that this weighting is the minimum-weight-energy (equivalently smoothest, minimum-roughness) zero-lag weighting, exactly and independently of period and smoothness order. Its central property is a transient-shape theorem: at every integer smoothness order, and (down to one sign-count step verified on the parameter grid) at every order below one, for every pole count and period, the unit-step error is weakly unimodal, so the response overshoots the target exactly once and thereafter approaches it from one side without ringing. The proof rests on the single-sign-change and total-positivity structure of the pole tail sums. At non-integer orders above one the same reduction holds and no-ringing becomes the unimodality of an explicit non-negative auxiliary kernel, a clean conjecture corroborated at scale. The family spans a shallow design space in period and pole count, with an exact second-moment law, a compound-Poisson pole-count limit and a fractional-pole completion; and the guarantee itself extends, at integer smoothness, to the affine weighting on any integer-depth support of the ladder of at least two depths, so no-ringing is the property of a whole schema of zero-lag filters rather than of one construction. On physiological signals under heavy broadband noise it is competitive with parameter-optimised classical smoothers.
No takes yet. Share an insight, caveat, or question.
Marcus John Hendry Don (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: