1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600
// Copyright (c) 2021-2022 Weird Constructor <weirdconstructor@gmail.com>
// This file is a part of synfx-dsp. Released under GPL-3.0-or-later.
// See README.md and COPYING for details.
//! A collection of filters, ranging from simple one poles to more interesting ones.
use crate::{f, Flt};
use std::simd::f32x4;
// one pole lp from valley rack free:
// https://github.com/ValleyAudio/ValleyRackFree/blob/v1.0/src/Common/DSP/OnePoleFilters.cpp
#[inline]
/// Process a very simple one pole 6dB low pass filter.
/// Useful in various applications, from usage in a synthesizer to
/// damping stuff in a reverb/delay.
///
/// * `input` - Input sample
/// * `freq` - Frequency between 1.0 and 22000.0Hz
/// * `israte` - 1.0 / samplerate
/// * `z` - The internal one sample buffer of the filter.
///
///```
/// use synfx_dsp::*;
///
/// let samples = vec![0.0; 44100];
/// let mut z = 0.0;
/// let mut freq = 1000.0;
///
/// for s in samples.iter() {
/// let s_out =
/// process_1pole_lowpass(*s, freq, 1.0 / 44100.0, &mut z);
/// // ... do something with the result here.
/// }
///```
pub fn process_1pole_lowpass(input: f32, freq: f32, israte: f32, z: &mut f32) -> f32 {
let b = (-std::f32::consts::TAU * freq * israte).exp();
let a = 1.0 - b;
*z = a * input + *z * b;
*z
}
#[derive(Debug, Clone, Copy, Default)]
pub struct OnePoleLPF<F: Flt> {
israte: F,
a: F,
b: F,
freq: F,
z: F,
}
impl<F: Flt> OnePoleLPF<F> {
pub fn new() -> Self {
Self {
israte: f::<F>(1.0) / f(44100.0),
a: f::<F>(0.0),
b: f::<F>(0.0),
freq: f::<F>(1000.0),
z: f::<F>(0.0),
}
}
pub fn reset(&mut self) {
self.z = f(0.0);
}
#[inline]
fn recalc(&mut self) {
self.b = (f::<F>(-1.0) * F::TAU() * self.freq * self.israte).exp();
self.a = f::<F>(1.0) - self.b;
}
#[inline]
pub fn set_sample_rate(&mut self, srate: F) {
self.israte = f::<F>(1.0) / srate;
self.recalc();
}
#[inline]
pub fn set_freq(&mut self, freq: F) {
if freq != self.freq {
self.freq = freq;
self.recalc();
}
}
#[inline]
pub fn process(&mut self, input: F) -> F {
self.z = self.a * input + self.z * self.b;
self.z
}
}
// Fixed one pole with setable pole and gain.
// Implementation taken from tubonitaub / alec-deason
// from https://github.com/alec-deason/virtual_modular/blob/4025f1ef343c2eb9cd74eac07b5350c1e7ec9c09/src/simd_graph.rs#L4292
// under MIT License
#[derive(Debug, Copy, Clone, Default)]
pub struct FixedOnePole {
b0: f32,
a1: f32,
y1: f32,
gain: f32,
}
impl FixedOnePole {
pub fn new(pole: f32, gain: f32) -> Self {
let b0 = if pole > 0.0 { 1.0 - pole } else { 1.0 + pole };
Self { b0, a1: -pole, y1: 0.0, gain }
}
pub fn reset(&mut self) {
self.y1 = 0.0;
}
pub fn set_gain(&mut self, gain: f32) {
self.gain = gain;
}
pub fn process(&mut self, input: f32) -> f32 {
let output = self.b0 * self.gain * input - self.a1 * self.y1;
self.y1 = output;
output
}
}
// one pole hp from valley rack free:
// https://github.com/ValleyAudio/ValleyRackFree/blob/v1.0/src/Common/DSP/OnePoleFilters.cpp
#[inline]
/// Process a very simple one pole 6dB high pass filter.
/// Useful in various applications.
///
/// * `input` - Input sample
/// * `freq` - Frequency between 1.0 and 22000.0Hz
/// * `israte` - 1.0 / samplerate
/// * `z` - The first internal buffer of the filter.
/// * `y` - The second internal buffer of the filter.
///
///```
/// use synfx_dsp::*;
///
/// let samples = vec![0.0; 44100];
/// let mut z = 0.0;
/// let mut y = 0.0;
/// let mut freq = 1000.0;
///
/// for s in samples.iter() {
/// let s_out =
/// process_1pole_highpass(*s, freq, 1.0 / 44100.0, &mut z, &mut y);
/// // ... do something with the result here.
/// }
///```
pub fn process_1pole_highpass(input: f32, freq: f32, israte: f32, z: &mut f32, y: &mut f32) -> f32 {
let b = (-std::f32::consts::TAU * freq * israte).exp();
let a = (1.0 + b) / 2.0;
let v = a * input - a * *z + b * *y;
*y = v;
*z = input;
v
}
#[derive(Debug, Clone, Copy, Default)]
pub struct OnePoleHPF<F: Flt> {
israte: F,
a: F,
b: F,
freq: F,
z: F,
y: F,
}
impl<F: Flt> OnePoleHPF<F> {
pub fn new() -> Self {
Self {
israte: f(1.0 / 44100.0),
a: f(0.0),
b: f(0.0),
freq: f(1000.0),
z: f(0.0),
y: f(0.0),
}
}
pub fn reset(&mut self) {
self.z = f(0.0);
self.y = f(0.0);
}
#[inline]
fn recalc(&mut self) {
self.b = (f::<F>(-1.0) * F::TAU() * self.freq * self.israte).exp();
self.a = (f::<F>(1.0) + self.b) / f(2.0);
}
pub fn set_sample_rate(&mut self, srate: F) {
self.israte = f::<F>(1.0) / srate;
self.recalc();
}
#[inline]
pub fn set_freq(&mut self, freq: F) {
if freq != self.freq {
self.freq = freq;
self.recalc();
}
}
#[inline]
pub fn process(&mut self, input: F) -> F {
let v = self.a * input - self.a * self.z + self.b * self.y;
self.y = v;
self.z = input;
v
}
}
// one pole from:
// http://www.willpirkle.com/Downloads/AN-4VirtualAnalogFilters.pdf
// (page 5)
#[inline]
/// Process a very simple one pole 6dB low pass filter in TPT form.
/// Useful in various applications, from usage in a synthesizer to
/// damping stuff in a reverb/delay.
///
/// * `input` - Input sample
/// * `freq` - Frequency between 1.0 and 22000.0Hz
/// * `israte` - 1.0 / samplerate
/// * `z` - The internal one sample buffer of the filter.
///
///```
/// use synfx_dsp::*;
///
/// let samples = vec![0.0; 44100];
/// let mut z = 0.0;
/// let mut freq = 1000.0;
///
/// for s in samples.iter() {
/// let s_out =
/// process_1pole_tpt_highpass(*s, freq, 1.0 / 44100.0, &mut z);
/// // ... do something with the result here.
/// }
///```
pub fn process_1pole_tpt_lowpass(input: f32, freq: f32, israte: f32, z: &mut f32) -> f32 {
let g = (std::f32::consts::PI * freq * israte).tan();
let a = g / (1.0 + g);
let v1 = a * (input - *z);
let v2 = v1 + *z;
*z = v2 + v1;
// let (m0, m1) = (0.0, 1.0);
// (m0 * input + m1 * v2) as f32);
v2
}
// one pole from:
// http://www.willpirkle.com/Downloads/AN-4VirtualAnalogFilters.pdf
// (page 5)
#[inline]
/// Process a very simple one pole 6dB high pass filter in TPT form.
/// Useful in various applications.
///
/// * `input` - Input sample
/// * `freq` - Frequency between 1.0 and 22000.0Hz
/// * `israte` - 1.0 / samplerate
/// * `z` - The internal one sample buffer of the filter.
///
///```
/// use synfx_dsp::*;
///
/// let samples = vec![0.0; 44100];
/// let mut z = 0.0;
/// let mut freq = 1000.0;
///
/// for s in samples.iter() {
/// let s_out =
/// process_1pole_tpt_lowpass(*s, freq, 1.0 / 44100.0, &mut z);
/// // ... do something with the result here.
/// }
///```
pub fn process_1pole_tpt_highpass(input: f32, freq: f32, israte: f32, z: &mut f32) -> f32 {
let g = (std::f32::consts::PI * freq * israte).tan();
let a1 = g / (1.0 + g);
let v1 = a1 * (input - *z);
let v2 = v1 + *z;
*z = v2 + v1;
input - v2
}
/// The internal oversampling factor of [process_hal_chamberlin_svf].
const FILTER_OVERSAMPLE_HAL_CHAMBERLIN: usize = 2;
// Hal Chamberlin's State Variable (12dB/oct) filter
// https://www.earlevel.com/main/2003/03/02/the-digital-state-variable-filter/
// Inspired by SynthV1 by Rui Nuno Capela, under the terms of
// GPLv2 or any later:
/// Process a HAL Chamberlin filter with two delays/state variables that is 12dB.
/// The filter does internal oversampling with very simple decimation to
/// rise the stability for cutoff frequency up to 16kHz.
///
/// * `input` - Input sample.
/// * `freq` - Frequency in Hz. Please keep it inside 0.0 to 16000.0 Hz!
/// otherwise the filter becomes unstable.
/// * `res` - Resonance from 0.0 to 0.99. Resonance of 1.0 is not recommended,
/// as the filter will then oscillate itself out of control.
/// * `israte` - 1.0 divided by the sampling rate (eg. 1.0 / 44100.0).
/// * `band` - First state variable, containing the band pass result
/// after processing.
/// * `low` - Second state variable, containing the low pass result
/// after processing.
///
/// Returned are the results of the high and notch filter.
///
///```
/// use synfx_dsp::*;
///
/// let samples = vec![0.0; 44100];
/// let mut band = 0.0;
/// let mut low = 0.0;
/// let mut freq = 1000.0;
///
/// for s in samples.iter() {
/// let (high, notch) =
/// process_hal_chamberlin_svf(
/// *s, freq, 0.5, 1.0 / 44100.0, &mut band, &mut low);
/// // ... do something with the result here.
/// }
///```
#[inline]
pub fn process_hal_chamberlin_svf(
input: f32,
freq: f32,
res: f32,
israte: f32,
band: &mut f32,
low: &mut f32,
) -> (f32, f32) {
let q = 1.0 - res;
let cutoff = 2.0 * (std::f32::consts::PI * freq * 0.5 * israte).sin();
let mut high = 0.0;
let mut notch = 0.0;
for _ in 0..FILTER_OVERSAMPLE_HAL_CHAMBERLIN {
*low += cutoff * *band;
high = input - *low - q * *band;
*band += cutoff * high;
notch = high + *low;
}
//d// println!("q={:4.2} cut={:8.3} freq={:8.1} LP={:8.3} HP={:8.3} BP={:8.3} N={:8.3}",
//d// q, cutoff, freq, *low, high, *band, notch);
(high, notch)
}
/// This function processes a Simper SVF with 12dB. It's a much newer algorithm
/// for filtering and provides easy to calculate multiple outputs.
///
/// * `input` - Input sample.
/// * `freq` - Frequency in Hz.
/// otherwise the filter becomes unstable.
/// * `res` - Resonance from 0.0 to 0.99. Resonance of 1.0 is not recommended,
/// as the filter will then oscillate itself out of control.
/// * `israte` - 1.0 divided by the sampling rate (eg. 1.0 / 44100.0).
/// * `band` - First state variable, containing the band pass result
/// after processing.
/// * `low` - Second state variable, containing the low pass result
/// after processing.
///
/// This function returns the low pass, band pass and high pass signal.
/// For a notch or peak filter signal, please consult the following example:
///
///```
/// use synfx_dsp::*;
///
/// let samples = vec![0.0; 44100];
/// let mut ic1eq = 0.0;
/// let mut ic2eq = 0.0;
/// let mut freq = 1000.0;
///
/// for s in samples.iter() {
/// let (low, band, high) =
/// process_simper_svf(
/// *s, freq, 0.5, 1.0 / 44100.0, &mut ic1eq, &mut ic2eq);
///
/// // You can easily calculate the notch and peak results too:
/// let notch = low + high;
/// let peak = low - high;
/// // ... do something with the result here.
/// }
///```
// Simper SVF implemented from
// https://cytomic.com/files/dsp/SvfLinearTrapezoidalSin.pdf
// Big thanks go to Andrew Simper @ Cytomic for developing and publishing
// the paper.
#[inline]
pub fn process_simper_svf(
input: f32,
freq: f32,
res: f32,
israte: f32,
ic1eq: &mut f32,
ic2eq: &mut f32,
) -> (f32, f32, f32) {
// XXX: the 1.989 were tuned by hand, so the resonance is more audible.
let k = 2f32 - (1.989f32 * res);
let w = std::f32::consts::PI * freq * israte;
let s1 = w.sin();
let s2 = (2.0 * w).sin();
let nrm = 1.0 / (2.0 + k * s2);
let g0 = s2 * nrm;
let g1 = (-2.0 * s1 * s1 - k * s2) * nrm;
let g2 = (2.0 * s1 * s1) * nrm;
let t0 = input - *ic2eq;
let t1 = g0 * t0 + g1 * *ic1eq;
let t2 = g2 * t0 + g0 * *ic1eq;
let v1 = t1 + *ic1eq;
let v2 = t2 + *ic2eq;
*ic1eq += 2.0 * t1;
*ic2eq += 2.0 * t2;
// low = v2
// band = v1
// high = input - k * v1 - v2
// notch = low + high = input - k * v1
// peak = low - high = 2 * v2 - input + k * v1
// all = low + high - k * band = input - 2 * k * v1
(v2, v1, input - k * v1 - v2)
}
/// This function implements a simple Stilson/Moog low pass filter with 24dB.
/// It provides only a low pass output.
///
/// * `input` - Input sample.
/// * `freq` - Frequency in Hz.
/// otherwise the filter becomes unstable.
/// * `res` - Resonance from 0.0 to 0.99. Resonance of 1.0 is not recommended,
/// as the filter will then oscillate itself out of control.
/// * `israte` - 1.0 divided by the sampling rate (`1.0 / 44100.0`).
/// * `b0` to `b3` - Internal values used for filtering.
/// * `delay` - A buffer holding other delayed samples.
///
///```
/// use synfx_dsp::*;
///
/// let samples = vec![0.0; 44100];
/// let mut b0 = 0.0;
/// let mut b1 = 0.0;
/// let mut b2 = 0.0;
/// let mut b3 = 0.0;
/// let mut delay = [0.0; 4];
/// let mut freq = 1000.0;
///
/// for s in samples.iter() {
/// let low =
/// process_stilson_moog(
/// *s, freq, 0.5, 1.0 / 44100.0,
/// &mut b0, &mut b1, &mut b2, &mut b3,
/// &mut delay);
///
/// // ... do something with the result here.
/// }
///```
// Stilson/Moog implementation partly translated from here:
// https://github.com/ddiakopoulos/MoogLadders/blob/master/src/MusicDSPModel.h
// without any copyright as found on musicdsp.org
// (http://www.musicdsp.org/showone.php?id=24).
//
// It's also found on MusicDSP and has probably no proper license anyways.
// See also: https://github.com/ddiakopoulos/MoogLadders
// and https://github.com/rncbc/synthv1/blob/master/src/synthv1_filter.h#L103
// and https://github.com/ddiakopoulos/MoogLadders/blob/master/src/MusicDSPModel.h
#[inline]
pub fn process_stilson_moog(
input: f32,
freq: f32,
res: f32,
israte: f32,
b0: &mut f32,
b1: &mut f32,
b2: &mut f32,
b3: &mut f32,
delay: &mut [f32; 4],
) -> f32 {
let cutoff = 2.0 * freq * israte;
let p = cutoff * (1.8 - 0.8 * cutoff);
let k = 2.0 * (cutoff * std::f32::consts::PI * 0.5).sin() - 1.0;
let t1 = (1.0 - p) * 1.386249;
let t2 = 12.0 + t1 * t1;
let res = res * (t2 + 6.0 * t1) / (t2 - 6.0 * t1);
let x = input - res * *b3;
// Four cascaded one-pole filters (bilinear transform)
*b0 = x * p + delay[0] * p - k * *b0;
*b1 = *b0 * p + delay[1] * p - k * *b1;
*b2 = *b1 * p + delay[2] * p - k * *b2;
*b3 = *b2 * p + delay[3] * p - k * *b3;
// Clipping band-limited sigmoid
*b3 -= (*b3 * *b3 * *b3) * 0.166667;
delay[0] = x;
delay[1] = *b0;
delay[2] = *b1;
delay[3] = *b2;
*b3
}
// translated from Odin 2 Synthesizer Plugin
// Copyright (C) 2020 TheWaveWarden
// under GPLv3 or any later
#[derive(Debug, Clone, Copy)]
pub struct DCBlockFilter<F: Flt> {
xm1: F,
ym1: F,
r: F,
}
impl<F: Flt> DCBlockFilter<F> {
pub fn new() -> Self {
Self { xm1: f(0.0), ym1: f(0.0), r: f(0.995) }
}
pub fn reset(&mut self) {
self.xm1 = f(0.0);
self.ym1 = f(0.0);
}
pub fn set_sample_rate(&mut self, srate: F) {
self.r = f(0.995);
if srate > f(90000.0) {
self.r = f(0.9965);
} else if srate > f(120000.0) {
self.r = f(0.997);
}
}
pub fn next(&mut self, input: F) -> F {
let y = input - self.xm1 + self.r * self.ym1;
self.xm1 = input;
self.ym1 = y;
y as F
}
}
// Taken from va-filter by Fredemus aka Frederik Halkjær aka RocketPhysician
// https://github.com/Fredemus/va-filter
// Under License GPL-3.0-or-later
/// Basic 4 channel SIMD DC-filter from Understanding Digital Signal Processing by Richard Lyons.
#[derive(Debug, Clone)]
pub struct DCFilterX4 {
y0: f32x4,
x0: f32x4,
/// higher alpha moves the cutoff lower, but also makes it settle slower.
/// See if it's reasonable to make it higher.
alpha: f32x4,
}
impl Default for DCFilterX4 {
fn default() -> Self {
Self {
y0: f32x4::splat(0.),
x0: f32x4::splat(0.),
// alpha: f32x4::splat(0.975),
alpha: f32x4::splat(0.9999),
}
}
}
impl DCFilterX4 {
pub fn process(&mut self, input: f32x4) -> f32x4 {
let y_new = input - self.x0 + self.alpha * self.y0;
self.x0 = input;
self.y0 = y_new;
y_new
}
pub fn reset(&mut self) {
self.y0 = f32x4::splat(0.);
self.x0 = f32x4::splat(0.);
self.alpha = f32x4::splat(0.9999);
}
}