Module filtering 

Digital filtering functions (FIR, Biquad IIR Direct Form I & II, LMS Adaptive Filter, Convolution, Correlation).

Structs

BiquadCascadeDf2tInstanceF32

Instance structure for the floating-point Biquad Cascade Transposed Direct Form II filter.

BiquadCascadeDf2tInstanceQ15

Instance structure for the Q15 Biquad Cascade Transposed Direct Form II filter.

BiquadCascadeDf2tInstanceQ31

Instance structure for the Q31 Biquad Cascade Transposed Direct Form II filter.

BiquadCascadeInstanceF32

Instance structure for the floating-point Biquad Cascade Direct Form I filter.

BiquadCascadeInstanceQ15

Instance structure for the Q15 Biquad Cascade Direct Form I filter.

BiquadCascadeInstanceQ31

Instance structure for the Q31 Biquad Cascade Direct Form I filter.

CircularBuffer

Const-generic zero-allocation circular buffer and delay line for real-time DSP sample streams.

DcBlockerQ15

High-pass single-pole used as a DC blocker (Smith Ch. 19).

FirInstanceF32

Instance structure for the floating-point FIR filter.

FirInstanceQ15

Instance structure for the Q15 FIR filter.

FirInstanceQ31

Instance structure for the Q31 FIR filter.

LmsInstanceF32

Instance structure for the floating-point LMS adaptive filter.

LmsInstanceQ15

Q15 LMS adaptive filter.

NlmsInstanceF32

Normalized LMS instance (eps floors the power denominator).

NlmsInstanceQ15

Q15 NLMS instance.

RecursiveMovingAverage

Const-generic N-point moving average filter implemented recursively (Steven W. Smith, Ch. 15, Eq. 15-3): each sample is updated with a single add and subtract, instead of an O(N) convolution sum.

RecursiveMovingAverageQ15

Q15 recursive N-point moving average (same recurrence as RecursiveMovingAverage).

SinglePoleFilter

The cheapest possible IIR filter: a single-pole recursive low-pass or high-pass filter (Steven W. Smith, Ch. 19, Eq. 19-2 / 19-3), needing only one or two multiplies per sample. Coefficients are designed from a decay factor x (see crate::filter_design::single_pole_decay_from_cutoff / crate::filter_design::single_pole_decay_from_time_constant).

SinglePoleFilterQ15

Q15 single-pole recursive low-pass or high-pass filter (same recurrence as SinglePoleFilter). decay is Q1.15 in 0..1 (larger → lower cutoff).

Functions

biquad_cascade_df1_f32

biquad_cascade_df1_q15

biquad_cascade_df1_q31

biquad_cascade_df2t_f32

biquad_cascade_df2t_q15

biquad_cascade_df2t_q31

conv_f32

conv_q7

conv_q15

conv_q31

correlate_f32

correlate_q15

correlate_q31

fast_convolve_f32

Performs fast linear convolution of signal and kernel via FFT multiplication. Output length is signal.len() + kernel.len() - 1.

fir_f32

fir_q15

fir_q31

lms_f32

lms_leaky_f32

Leaky LMS: w ← (1 - leak) w + 2 μ e x. leak = 0 matches lms_f32.

lms_leaky_q15

Leaky LMS in Q15. leak is Q1.15 (0 matches lms_q15).

lms_q15

median_filter_1d_f32

1D Conditional / Thresholded Median Filter for f32.

median_filter_1d_q15

1D Conditional Median Filter for Q15.

median_filter_1d_q31

1D Conditional Median Filter for Q31.

nlms_f32

NLMS: w ← w + μ e x / (eps + ||x||²).

nlms_q15