Struct Lowpass

pub struct Lowpass<const N: usize>(/* private fields */);

Arbitrary order, high dynamic range, wide coefficient range, lowpass filter implementation. DC gain is 1.

Type argument N is the filter order. N must be 1 or 2.

The filter will cleanly saturate towards the i32 range.

Both filters have been optimized for accuracy, dynamic range, and speed on Cortex-M7.

Trait Implementations

impl<const N: usize> Clone for Lowpass<N>

fn clone(&self) -> Lowpass<N>

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<const N: usize> Default for Lowpass<N>

fn default() -> Self

Returns the “default value” for a type.

impl<const N: usize> Filter for Lowpass<N>

type Config = [i32; N]

The filter configuration Config contains the filter gains.

For the first-order lowpass this is a single element array [k] with the corner frequency in scaled Q31: k = pi*(1 << 31)*f0/fn where f0 is the 3dB corner frequency and fn is the Nyquist frequency. The corner frequency is warped in the usual way.

For the second-order lowpass this is [k**2/(1 << 32), -k/q] with q = 1/sqrt(2) for a Butterworth response.

In addition to the poles at the corner frequency the filters have zeros at Nyquist.

The first-order lowpass works fine and accurate for any positive gain 1 <= k <= (1 << 31) - 1. The second-order lowpass works and is accurate for 1 << 16 <= k <= q*(1 << 31).

fn update(&mut self, x: i32, k: &Self::Config) -> i32

Update the filter with a new sample.

fn get(&self) -> i32

Return the current filter output

fn set(&mut self, x: i32)

Update the filter so that it outputs the provided value. This does not completely define the state of the filter.

impl<const N: usize> Copy for Lowpass<N>