Adaptive Filter
Block Catalogue / Signal Processing
First-order IIR low-pass whose cutoff is recomputed every sample as fc = exp(Offset + Deriviate·|dx_filt| + Magnitude·|x|). Built for heteroscedastic signals such as EMG envelopes: it smooths hard when the signal is quiet and opens up on fast transients.
Input and output
| Property | Value |
|---|---|
| Type | adaptivefilterblock, adaptivefilter, mosaic.models.adaptivefilterblock |
| Inputs | exactly 1 |
| Consumes | Vector<double> (one sample, one element per channel) or Matrix<double> [timesteps × channels] |
| Publishes | same type and shape as the input (Vector<double> → Vector<double>, Matrix<double> → Matrix<double>) |
Parameters
| # | Name | Type | Default | Meaning |
|---|---|---|---|---|
| 0 | Alpha | double | 5.0 | Derivative pre-filter smoothing on a UI scale 0–10; internally α = Alpha/10 |
| 1 | Offset | double | 0.0 | Base exponent |
| 2 | Magnitude | double | -20.0 | \|x\| contribution to the exponent |
| 3 | Deriviate | double | 10.0 | \|dx_filt\| contribution to the exponent |
All optional and cumulative; a shorter list takes the defaults. No param sets the sample rate — fs comes from DesiredRate, defaulting to 200 Hz.
Requirements and use
Tune offset, derivative and magnitude terms for the units of your signal. Check the output against a known input before applying the filter to live control.
For parameter positions and connection rules, see Pipeline Configuration. For capturing output, see Recording.
Implementation
API reference. Model defaults above describe configuration loading; palette hints may expose a subset or select different defaults.