cyfaust 0.1.0__cp311-cp311-macosx_15_0_arm64.whl

cyfaust/resources/libraries/filters.lib

@@ -0,0 +1,3125 @@
+//##################################### filters.lib ########################################
+// Filters library. Its official prefix is `fi`.
+//
+// The Filters library is organized into 22 sections:
+//
+// * [Basic Filters](#basic-filters)
+// * [Comb Filters](#comb-filters)
+// * [Direct-Form Digital Filter Sections](#direct-form-digital-filter-sections)
+// * [Direct-Form Second-Order Biquad Sections](#direct-form-second-order-biquad-sections)
+// * [Ladder/Lattice Digital Filters](#ladderlattice-digital-filters)
+// * [Useful Special Cases](#useful-special-cases)
+// * [Ladder/Lattice Allpass Filters](#ladderlattice-allpass-filters)
+// * [Digital Filter Sections Specified as Analog Filter Sections](#digital-filter-sections-specified-as-analog-filter-sections)
+// * [Simple Resonator Filters](#simple-resonator-filters)
+// * [Butterworth Lowpass/Highpass Filters](#butterworth-lowpasshighpass-filters)
+// * [Special Filter-Bank Delay-Equalizing Allpass Filters](#special-filter-bank-delay-equalizing-allpass-filters)
+// * [Elliptic (Cauer) Lowpass Filters](#elliptic-cauer-lowpass-filters)
+// * [Elliptic Highpass Filters](#elliptic-highpass-filters)
+// * [Butterworth Bandpass/Bandstop Filters](#butterworth-bandpassbandstop-filters)
+// * [Elliptic Bandpass Filters](#elliptic-bandpass-filters)
+// * [Parametric Equalizers (Shelf, Peaking)](#parametric-equalizers-shelf-peaking)
+// * [Mth-Octave Filter-Banks](#mth-octave-filter-banks)
+// * [Arbitrary-Crossover Filter-Banks and Spectrum Analyzers](#arbitrary-crossover-filter-banks-and-spectrum-analyzers)
+// * [State Variable Filters (SVF)](#state-variable-filters)
+// * [Linkwitz-Riley 4th-order 2-way, 3-way, and 4-way crossovers](#linkwitz-riley-4th-order-2-way-3-way-and-4-way-crossovers)
+// * [Standardized Filters](#standardized-filters)
+// * [Averaging Functions](#averaging-functions)
+//
+// #### References
+// * <https://github.com/grame-cncm/faustlibraries/blob/master/filters.lib>
+//
+//########################################################################################
+
+// NOTE ABOUT LICENSES:
+// Each function in this library has its own license. Licenses are declared
+// before each function. Corresponding license terms can be found at the 
+// bottom of this file or in the Faust libraries documentation.
+
+ma = library("maths.lib");
+ba = library("basics.lib");
+ro = library("routes.lib");
+de = library("delays.lib");
+an = library("analyzers.lib");
+ef = library("misceffects.lib");
+si = library("signals.lib");
+fi = library("filters.lib"); // for compatible copy/paste out of this file
+
+declare name "Faust Filters Library";
+declare version "1.3.0";
+
+//===============================Basic Filters============================================
+//========================================================================================
+
+//----------------------`(fi.)zero`--------------------------
+// One zero filter. Difference equation: \(y(n) = x(n) - zx(n-1)\).
+//
+// #### Usage
+//
+// ```
+// _ : zero(z) : _
+// ```
+//
+// Where:
+//
+// * `z`: location of zero along real axis in z-plane
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/filters/One_Zero.html>
+//----------------------------------------------------------
+declare zero author "Julius O. Smith III";
+declare zero copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare zero license "MIT-style STK-4.3 license";
+zero(z) = _ <: _,mem : _,*(z) : -;
+
+//------------------------`(fi.)pole`---------------------------
+// One pole filter. Could also be called a "leaky integrator".
+// Difference equation: \(y(n) = x(n) + py(n-1)\).
+//
+// #### Usage
+//
+// ```
+// _ : pole(p) : _
+// ```
+//
+// Where:
+//
+// * `p`: pole location = feedback coefficient
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/filters/One_Pole.html>
+//------------------------------------------------------------
+declare pole author "Julius O. Smith III";
+declare pole copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare pole license "MIT-style STK-4.3 license";
+pole(p) = + ~ *(p);
+
+//----------------------`(fi.)integrator`--------------------------
+// Same as `pole(1)` [implemented separately for block-diagram clarity].
+//------------------------------------------------------------
+declare integrator author "Julius O. Smith III";
+declare integrator copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare integrator license "MIT-style STK-4.3 license";
+integrator = + ~ _;
+
+//-------------------`(fi.)dcblockerat`-----------------------
+// DC blocker with configurable break frequency.
+// The amplitude response is substantially flat above \(fb\),
+// and sloped at about +6 dB/octave below \(fb\).
+// Derived from the analog transfer function:
+// $$H(s) = \frac{s}{(s + 2 \pi fb)}$$
+// (which can be seen as a 1st-order Butterworth highpass filter)
+// by the low-frequency-matching bilinear transform method
+// (i.e., the standard frequency-scaling constant 2*SR).
+//
+// #### Usage
+//
+// ```
+// _ : dcblockerat(fb) : _
+// ```
+//
+// Where:
+//
+// * `fb`: "break frequency" in Hz, i.e., -3 dB gain frequency.
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/pasp/Bilinear_Transformation.html>
+//------------------------------------------------------------
+declare dcblockerat author "Julius O. Smith III";
+declare dcblockerat copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare dcblockerat license "MIT-style STK-4.3 license";
+dcblockerat(fb) = *(b0) : zero(1) : pole(p)
+with {
+  wn = ma.PI*fb/ma.SR;
+  b0 = 1.0 / (1 + wn);
+  p = (1 - wn) * b0;
+};
+
+//----------------------`(fi.)dcblocker`--------------------------
+// DC blocker. Default dc blocker has -3dB point near 35 Hz (at 44.1 kHz)
+// and high-frequency gain near 1.0025 (due to no scaling).
+// `dcblocker` is as standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : dcblocker : _
+// ```
+//------------------------------------------------------------
+declare dcblocker author "Julius O. Smith III";
+declare dcblocker copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare dcblocker license "MIT-style STK-4.3 license";
+dcblocker = zero(1) : pole(0.995);
+
+//----------------------------`(fi.)lptN`--------------------------------------
+// One-pole lowpass filter with arbitrary dis/charging factors set in dB and
+// times set in seconds.
+//
+// #### Usage
+//
+// ```
+// _ : lptN(N, tN) : _
+// ```
+//
+// Where:
+//
+// * `N`: is the attenuation factor in dB
+// * `tN`: is the filter period in seconds, that is, the time for the
+// impulse response to decay by `N` dB
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/mdft/Exponentials.html>
+//----------------------------------------------------------
+declare lptN author "Julius O. Smith III";
+declare lptN copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare lptN license "MIT-style STK-4.3 license";
+lptN(N, tN, x) = x : si.smooth(ba.tau2pole(tN / log(10.0^(float(N)/20.0))));
+// Special cases of lptN
+lptau(tN, x) = lptN(8.6858896381, tN, x); // Tau time constant, i.e., 1/e atten. after tN secs
+lpt60(tN, x) = lptN(60, tN, x); // T60 constant, i.e., 1/1000 atten. after tN secs
+lpt19(tN, x) = lptN(19, tN, x); // T19 constant, i.e., 1/e^2.2 atten. after tN secs
+
+//=======================================Comb Filters=====================================
+//========================================================================================
+
+//------`(fi.)ff_comb`--------
+// Feed-Forward Comb Filter. Note that `ff_comb` requires integer delays
+// (uses `delay`  internally).
+// `ff_comb` is a standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : ff_comb(maxdel,intdel,b0,bM) : _
+// ```
+//
+// Where:
+//
+// * `maxdel`: maximum delay (a power of 2)
+// * `intdel`: current (integer) comb-filter delay between 0 and maxdel
+// * `del`: current (float) comb-filter delay between 0 and maxdel
+// * `b0`: gain applied to delay-line input
+// * `bM`: gain applied to delay-line output and then summed with input
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/pasp/Feedforward_Comb_Filters.html>
+//------------------------------------------------------------
+declare ff_comb author "Julius O. Smith III";
+declare ff_comb copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare ff_comb license "MIT-style STK-4.3 license";
+ff_comb(maxdel,M,b0,bM) = _ <: *(b0), bM * de.delay(maxdel,M) : +;
+
+//------`(fi.)ff_fcomb`--------
+// Feed-Forward Comb Filter. Note that `ff_fcomb` takes floating-point delays
+// (uses `fdelay` internally).
+// `ff_fcomb` is a standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : ff_fcomb(maxdel,del,b0,bM) : _
+// ```
+//
+// Where:
+//
+// * `maxdel`: maximum delay (a power of 2)
+// * `intdel`: current (integer) comb-filter delay between 0 and maxdel
+// * `del`: current (float) comb-filter delay between 0 and maxdel
+// * `b0`: gain applied to delay-line input
+// * `bM`: gain applied to delay-line output and then summed with input
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/pasp/Feedforward_Comb_Filters.html>
+//------------------------------------------------------------
+declare ff_fcomb author "Julius O. Smith III";
+declare ff_fcomb copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare ff_fcomb license "MIT-style STK-4.3 license";
+ff_fcomb(maxdel,M,b0,bM) = _ <: *(b0), bM * de.fdelay(maxdel,M) : +;
+
+//-----------`(fi.)ffcombfilter`-------------------
+// Typical special case of `ff_comb()` where: `b0 = 1`.
+//------------------------------------------------------------
+declare ff_combfilter author "Julius O. Smith III";
+declare ff_combfilter copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare ff_combfilter license "MIT-style STK-4.3 license";
+ffcombfilter(maxdel,del,g) = ff_comb(maxdel,del,1,g);
+
+
+//-----------------------`(fi.)fb_comb`-----------------------
+// Feed-Back Comb Filter (integer delay).
+//
+// #### Usage
+//
+// ```
+// _ : fb_comb(maxdel,intdel,b0,aN) : _
+// ```
+//
+// Where:
+//
+// * `maxdel`: maximum delay (a power of 2)
+// * `intdel`: current (integer) comb-filter delay between 0 and maxdel
+// * `del`: current (float) comb-filter delay between 0 and maxdel
+// * `b0`: gain applied to delay-line input and forwarded to output
+// * `aN`: minus the gain applied to delay-line output before summing with the input
+// 	and feeding to the delay line
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/pasp/Feedback_Comb_Filters.html>
+//------------------------------------------------------------
+declare fb_comb author "Julius O. Smith III";
+declare fb_comb copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare fb_comb license "MIT-style STK-4.3 license";
+fb_comb(maxdel,N,b0,aN) = (+ <: de.delay(maxdel,N-1),_) ~ *(-aN) : !,*(b0) : mem;
+
+
+//-----------------------`(fi.)fb_fcomb`-----------------------
+// Feed-Back Comb Filter (floating point delay).
+//
+// #### Usage
+//
+// ```
+// _ : fb_fcomb(maxdel,del,b0,aN) : _
+// ```
+//
+// Where:
+//
+// * `maxdel`: maximum delay (a power of 2)
+// * `intdel`: current (integer) comb-filter delay between 0 and maxdel
+// * `del`: current (float) comb-filter delay between 0 and maxdel
+// * `b0`: gain applied to delay-line input and forwarded to output
+// * `aN`: minus the gain applied to delay-line output before summing with the input
+// 	and feeding to the delay line
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/pasp/Feedback_Comb_Filters.html>
+//------------------------------------------------------------
+declare fb_fcomb author "Julius O. Smith III";
+declare fb_fcomb copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare fb_fcomb license "MIT-style STK-4.3 license";
+fb_fcomb(maxdel,N,b0,aN) = (+ <: de.fdelay(maxdel,float(N)-1.0),_) ~ *(-aN) : !,*(b0) : mem;
+
+//-----------------------`(fi.)rev1`-----------------------
+// Special case of `fb_comb` (`rev1(maxdel,N,g)`).
+// The "rev1 section" dates back to the 1960s in computer-music reverberation.
+// See the `jcrev` and `brassrev` in `reverbs.lib` for usage examples.
+//------------------------------------------------------------
+declare rev1 author "Julius O. Smith III";
+declare rev1 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare rev1 license "MIT-style STK-4.3 license";
+rev1(maxdel,N,g) = fb_comb (maxdel,N,1,-g);
+
+//-----`(fi.)fbcombfilter` and `(fi.)ffbcombfilter`------------
+// Other special cases of Feed-Back Comb Filter.
+//
+// #### Usage
+//
+// ```
+// _ : fbcombfilter(maxdel,intdel,g) : _
+// _ : ffbcombfilter(maxdel,del,g) : _
+// ```
+//
+// Where:
+//
+// * `maxdel`: maximum delay (a power of 2)
+// * `intdel`: current (integer) comb-filter delay between 0 and maxdel
+// * `del`: current (float) comb-filter delay between 0 and maxdel
+// * `g`: feedback gain
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/pasp/Feedback_Comb_Filters.html>
+//------------------------------------------------------------
+declare fbcombfilter author "Julius O. Smith III";
+declare fbcombfilter copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare fbcombfilter license "MIT-style STK-4.3 license";
+fbcombfilter(maxdel,intdel,g) = (+ : de.delay(maxdel,intdel)) ~ *(g);
+
+declare ffbcombfilter author "Julius O. Smith III";
+declare ffbcombfilter copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare ffbcombfilter license "MIT-style STK-4.3 license";
+ffbcombfilter(maxdel,del,g) = (+ : de.fdelay(maxdel,del)) ~ *(g);
+
+
+//-------------------`(fi.)allpass_comb`-----------------
+// Schroeder Allpass Comb Filter. Note that:
+//
+// ```
+// allpass_comb(maxlen,len,aN) = ff_comb(maxlen,len,aN,1) : fb_comb(maxlen,len-1,1,aN);
+// ```
+//
+// which is a direct-form-1 implementation, requiring two delay lines.
+// The implementation here is direct-form-2 requiring only one delay line.
+//
+// #### Usage
+//
+// ```
+// _ : allpass_comb(maxdel,intdel,aN) : _
+// ```
+//
+// Where:
+//
+// * `maxdel`: maximum delay (a power of 2)
+// * `intdel`: current (integer) comb-filter delay between 0 and maxdel
+// * `del`: current (float) comb-filter delay between 0 and maxdel
+// * `aN`: minus the feedback gain
+//
+// #### References
+// * <https://ccrma.stanford.edu/~jos/pasp/Allpass_Two_Combs.html>
+// * <https://ccrma.stanford.edu/~jos/pasp/Schroeder_Allpass_Sections.html>
+// * <https://ccrma.stanford.edu/~jos/filters/Four_Direct_Forms.html>
+//------------------------------------------------------------
+declare allpass_comb author "Julius O. Smith III";
+declare allpass_comb copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare allpass_comb license "MIT-style STK-4.3 license";
+allpass_comb(maxdel,N,aN) = (+ <: de.delay(maxdel,N-1),*(aN)) ~ *(-aN) : mem,_ : +;
+
+
+//-------------------`(fi.)allpass_fcomb`-----------------
+// Schroeder Allpass Comb Filter. Note that:
+//
+// ```
+// allpass_comb(maxlen,len,aN) = ff_comb(maxlen,len,aN,1) : fb_comb(maxlen,len-1,1,aN);
+// ```
+//
+// which is a direct-form-1 implementation, requiring two delay lines.
+// The implementation here is direct-form-2 requiring only one delay line.
+//
+// `allpass_fcomb` is a standard Faust library.
+//
+// #### Usage
+//
+// ```
+// _ : allpass_comb(maxdel,intdel,aN) : _
+// _ : allpass_fcomb(maxdel,del,aN) : _
+// ```
+//
+// Where:
+//
+// * `maxdel`: maximum delay (a power of 2)
+// * `intdel`: current (float) comb-filter delay between 0 and maxdel
+// * `del`: current (float) comb-filter delay between 0 and maxdel
+// * `aN`: minus the feedback gain
+//
+// #### References
+// * <https://ccrma.stanford.edu/~jos/pasp/Allpass_Two_Combs.html>
+// * <https://ccrma.stanford.edu/~jos/pasp/Schroeder_Allpass_Sections.html>
+// * <https://ccrma.stanford.edu/~jos/filters/Four_Direct_Forms.html>
+//------------------------------------------------------------
+declare allpass_fcomb author "Julius O. Smith III";
+declare allpass_fcomb copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare allpass_fcomb license "MIT-style STK-4.3 license";
+allpass_fcomb(maxdel,N,aN) = (+ <: de.fdelay(maxdel,N-1),*(aN)) ~ *(-aN) : mem,_ : +;
+
+
+//-----------------------`(fi.)rev2`-----------------------
+// Special case of `allpass_comb` (`rev2(maxlen,len,g)`).
+// The "rev2 section" dates back to the 1960s in computer-music reverberation.
+// See the `jcrev` and `brassrev` in `reverbs.lib` for usage examples.
+//------------------------------------------------------------
+declare rev2 author "Julius O. Smith III";
+declare rev2 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare rev2 license "MIT-style STK-4.3 license";
+rev2(maxlen,len,g) = allpass_comb(maxlen,len,-g);
+
+//-------------------`(fi.)allpass_fcomb5` and `(fi.)allpass_fcomb1a`-----------------
+// Same as `allpass_fcomb` but use `fdelay5` and `fdelay1a` internally
+// (Interpolation helps - look at an fft of faust2octave on
+//
+// ```
+// `1-1' <: allpass_fcomb(1024,10.5,0.95), allpass_fcomb5(1024,10.5,0.95);`).
+// ```
+//------------------------------------------------------------
+declare allpass_fcomb5 author "Julius O. Smith III";
+declare allpass_fcomb5 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare allpass_fcomb5 license "MIT-style STK-4.3 license";
+allpass_fcomb5(maxdel,N,aN) = (+ <: de.fdelay5(maxdel,N-1),*(aN)) ~ *(-aN) : mem,_ : +;
+
+declare allpass_fcomb1a author "Julius O. Smith III";
+declare allpass_fcomb1a copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare allpass_fcomb1a license "MIT-style STK-4.3 license";
+allpass_fcomb1a(maxdel,N,aN) = (+ <: de.fdelay1a(maxdel,N-1),*(aN)) ~ *(-aN) : mem,_ : +;
+
+
+//========================Direct-Form Digital Filter Sections=============================
+//========================================================================================
+
+// Specified by transfer-function polynomials B(z)/A(z) as in matlab
+
+//----------------------------`(fi.)iir`-------------------------------
+// Nth-order Infinite-Impulse-Response (IIR) digital filter,
+// implemented in terms of the Transfer-Function (TF) coefficients.
+// Such filter structures are termed "direct form".
+//
+// `iir` is a standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : iir(bcoeffs,acoeffs) : _
+// ```
+//
+// Where:
+//
+// * `bcoeffs`: (b0,b1,...,b_order) = TF numerator coefficients
+// * `acoeffs`: (a1,...,a_order) = TF denominator coeffs (a0=1)
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/filters/Four_Direct_Forms.html>
+//------------------------------------------------------------
+declare iir author "Julius O. Smith III";
+declare iir copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare iir license "MIT-style STK-4.3 license";
+iir(bv,av) = ma.sub ~ fir(av) : fir(bv);
+
+//-----------------------------`(fi.)fir`---------------------------------
+// FIR filter (convolution of FIR filter coefficients with a signal). `fir` is standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : fir(bv) : _
+// ```
+//
+// Where:
+//
+// * `bv` = b0,b1,...,bn is a parallel bank of coefficient signals.
+//
+// #### Note
+//
+// `bv` is processed using pattern-matching at compile time,
+//       so it must have this normal form (parallel signals).
+//
+// #### Example test program
+//
+// Smoothing white noise with a five-point moving average:
+//
+// ```
+// bv = .2,.2,.2,.2,.2;
+// process = noise : fir(bv);
+// ```
+//
+// Equivalent (note double parens):
+//
+// ```
+// process = noise : fir((.2,.2,.2,.2,.2));
+// ```
+//------------------------------------------------------------
+//fir(bv) = conv(bv);
+declare fir author "Julius O. Smith III";
+declare fir copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare fir license "MIT-style STK-4.3 license";
+fir((b0,bv)) = _ <: *(b0), R(1,bv) :> _ with {
+	R(n,(bn,bv)) = (@(n):*(bn)), R(n+1,bv);
+	R(n, bn)     = (@(n):*(bn)); };
+fir(b0) = *(b0);
+
+//---------------`(fi.)conv` and `(fi.)convN`-------------------------------
+// Convolution of input signal with given coefficients.
+//
+// #### Usage
+//
+// ```
+// _ : conv((k1,k2,k3,...,kN)) : _ // Argument = one signal bank
+// _ : convN(N,(k1,k2,k3,...)) : _ // Useful when N < count((k1,...))
+// ```
+//------------------------------------------------------------
+//convN(N,kv,x) = sum(i,N,take(i+1,kv) * x@i); // take() defined in math.lib
+
+declare convN author "Julius O. Smith III";
+declare convN copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare convN license "MIT-style STK-4.3 license";
+convN(N,kv) = sum(i,N, @(i)*take(i+1,kv)); // take() defined in math.lib
+//conv(kv,x) = sum(i,count(kv),take(i+1,kv) * x@i); // count() from math.lib
+
+declare conv author "Julius O. Smith III";
+declare conv copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare conv license "MIT-style STK-4.3 license";
+conv(kv) = fir(kv);
+
+//----------------`(fi.)tf1`, `(fi.)tf2` and `(fi.)tf3`----------------------
+// tfN = N'th-order direct-form digital filter.
+//
+// #### Usage
+//
+// ```
+// _ : tf1(b0,b1,a1) : _
+// _ : tf2(b0,b1,b2,a1,a2) : _
+// _ : tf3(b0,b1,b2,b3,a1,a2,a3) : _
+// ```
+//
+// Where:
+//
+// * `a`: the poles
+// * `b`: the zeros
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/fp/Direct_Form_I.html>
+//------------------------------------------------------------
+declare tf1 author "Julius O. Smith III";
+declare tf1 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare tf1 license "MIT-style STK-4.3 license";
+tf1(b0,b1,a1) = _ <: *(b0), (mem : *(b1)) :> + ~ *(0-a1);
+
+declare tf2 author "Julius O. Smith III";
+declare tf2 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare tf2 license "MIT-style STK-4.3 license";
+tf2(b0,b1,b2,a1,a2) = iir((b0,b1,b2),(a1,a2));
+// tf2 is a variant of tf22 below with duplicated mems
+
+declare tf3 author "Julius O. Smith III";
+declare tf3 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare tf3 license "MIT-style STK-4.3 license";
+tf3(b0,b1,b2,b3,a1,a2,a3) = iir((b0,b1,b2,b3),(a1,a2,a3));
+
+// "Original" version for music.lib. This is here for comparison but people should
+// use tf2 instead
+TF2(b0,b1,b2,a1,a2) = sub ~ conv2(a1,a2) : conv3(b0,b1,b2)
+with {
+	conv3(k0,k1,k2,x) = k0*x + k1*x' + k2*x'';
+	conv2(k0,k1,x)    = k0*x + k1*x';
+	sub(x,y)          = y-x;
+};
+
+//------------`(fi.)notchw`--------------
+// Simple notch filter based on a biquad (`tf2`).
+// `notchw` is a standard Faust function.
+//
+// #### Usage:
+//
+// ```
+// _ : notchw(width,freq) : _
+// ```
+//
+// Where:
+//
+// * `width`: "notch width" in Hz (approximate)
+// * `freq`: "notch frequency" in Hz
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/pasp/Phasing_2nd_Order_Allpass_Filters.html>
+//------------------------------------------------------------
+declare notchw author "Julius O. Smith III";
+declare notchw copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare notchw license "MIT-style STK-4.3 license";
+notchw(width,freq) = tf2(b0,b1,b2,a1,a2)
+with {
+  fb = 0.5*width; // First design a dcblockerat(width/2)
+  wn = ma.PI*fb/ma.SR;
+  b0db = 1.0 / (1 + wn);
+  p = (1 - wn) * b0db; // This is our pole radius.
+  // Now place unit-circle zeros at desired angles:
+  tn = 2*ma.PI*freq/ma.SR;
+  a2 = p * p;
+  a2p1 = 1+a2;
+  a1 = -a2p1*cos(tn);
+  b1 = a1;
+  b0 = 0.5*a2p1;
+  b2 = b0;
+};
+
+//======================Direct-Form Second-Order Biquad Sections==========================
+// Direct-Form Second-Order Biquad Sections
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/filters/Four_Direct_Forms.html>
+//========================================================================================
+
+//----------------`(fi.)tf21`, `(fi.)tf22`, `(fi.)tf22t` and `(fi.)tf21t`----------------------
+// tfN = N'th-order direct-form digital filter where:
+//
+// * `tf21` is tf2, direct-form 1
+// * `tf22` is tf2, direct-form 2
+// * `tf22t` is tf2, direct-form 2 transposed
+// * `tf21t` is tf2, direct-form 1 transposed
+//
+// #### Usage
+//
+// ```
+// _ : tf21(b0,b1,b2,a1,a2) : _
+// _ : tf22(b0,b1,b2,a1,a2) : _
+// _ : tf22t(b0,b1,b2,a1,a2) : _
+// _ : tf21t(b0,b1,b2,a1,a2) : _
+// ```
+//
+// Where:
+//
+// * `a`: the poles
+// * `b`: the zeros
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/fp/Direct_Form_I.html>
+//------------------------------------------------------------
+declare tf21 author "Julius O. Smith III";
+declare tf21 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare tf21 license "MIT-style STK-4.3 license";
+tf21(b0,b1,b2,a1,a2) = // tf2, direct-form 1:
+    _ <:(mem<:((mem:*(b2)),*(b1))),*(b0) :>_
+    : ((_,_,_:>_) ~(_<:*(-a1),(mem:*(-a2))));
+    
+declare tf22 author "Julius O. Smith III";
+declare tf22 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare tf22 license "MIT-style STK-4.3 license";
+tf22(b0,b1,b2,a1,a2) = // tf2, direct-form 2:
+    _ : (((_,_,_:>_)~*(-a1)<:mem,*(b0))~*(-a2))
+      : (_<:mem,*(b1)),_ : *(b2),_,_ :> _;
+
+declare tf22t author "Julius O. Smith III";
+declare tf22t copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare tf22t license "MIT-style STK-4.3 license";
+tf22t(b0,b1,b2,a1,a2) = // tf2, direct-form 2 transposed:
+    _ : (_,_,(_ <: *(b2)',*(b1)',*(b0))
+      : _,+',_,_ :> _)~*(-a1)~*(-a2) : _;
+
+declare tf21t author "Julius O. Smith III";
+declare tf21t copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare tf21t license "MIT-style STK-4.3 license";
+tf21t(b0,b1,b2,a1,a2) = // tf2, direct-form 1 transposed:
+    tf22t(1,0,0,a1,a2) : tf22t(b0,b1,b2,0,0); // or write it out if you want
+
+//=========================== Ladder/Lattice Digital Filters =============================
+// Ladder and lattice digital filters generally have superior numerical
+// properties relative to direct-form digital filters.  They can be derived
+// from digital waveguide filters, which gives them a physical interpretation.
+
+// #### Reference
+// * F. Itakura and S. Saito: "Digital Filtering Techniques for Speech Analysis and Synthesis",
+//     7th Int. Cong. Acoustics, Budapest, 25 C 1, 1971.
+// * J. D. Markel and A. H. Gray: Linear Prediction of Speech, New York: Springer Verlag, 1976.
+// * <https://ccrma.stanford.edu/~jos/pasp/Conventional_Ladder_Filters.html>
+//========================================================================================
+
+//-------------------------------`(fi.)av2sv`-----------------------------------
+// Compute reflection coefficients sv from transfer-function denominator av.
+//
+// #### Usage
+//
+// ```
+// sv = av2sv(av)
+// ```
+//
+// Where:
+//
+// * `av`: parallel signal bank `a1,...,aN`
+// * `sv`: parallel signal bank `s1,...,sN`
+//
+// where `ro = ith` reflection coefficient, and
+//       `ai` = coefficient of `z^(-i)` in the filter
+//          transfer-function denominator `A(z)`.
+//
+// #### Reference
+//   <https://ccrma.stanford.edu/~jos/filters/Step_Down_Procedure.html>
+//   (where reflection coefficients are denoted by k rather than s).
+//------------------------------------------------------------
+declare av2sv author "Julius O. Smith III";
+declare av2sv copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare av2sv license "MIT-style STK-4.3 license";
+av2sv(av) = par(i,M,s(i+1)) with {
+  M = ba.count(av);
+  s(m) = sr(M-m+1); // m=1..M
+  sr(m) = Ari(m,M-m+1); // s_{M-1-m}
+  Ari(m,i) = ba.take(i+1,Ar(m-1));
+  //step-down recursion for lattice/ladder digital filters:
+  Ar(0) = (1,av); // Ar(m) is order M-m (i.e. "reverse-indexed")
+  Ar(m) = 1,par(i,M-m, (Ari(m,i+1) - sr(m)*Ari(m,M-m-i))/(1-sr(m)*sr(m)));
+};
+
+//----------------------------`(fi.)bvav2nuv`--------------------------------
+// Compute lattice tap coefficients from transfer-function coefficients.
+//
+// #### Usage
+//
+// ```
+// nuv = bvav2nuv(bv,av)
+// ```
+//
+// Where:
+//
+// * `av`: parallel signal bank `a1,...,aN`
+// * `bv`: parallel signal bank `b0,b1,...,aN`
+// * `nuv`: parallel signal bank  `nu1,...,nuN`
+//
+// where `nui` is the i'th tap coefficient,
+//       `bi` is the coefficient of `z^(-i)` in the filter numerator,
+//       `ai` is the coefficient of `z^(-i)` in the filter denominator
+//------------------------------------------------------------
+declare bvav2nuv author "Julius O. Smith III";
+declare bvav2nuv copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare bvav2nuv license "MIT-style STK-4.3 license";
+bvav2nuv(bv,av) = par(m,M+1,nu(m)) with {
+  M = ba.count(av);
+  nu(m) = ba.take(m+1,Pr(M-m)); // m=0..M
+  // lattice/ladder tap parameters:
+  Pr(0) = bv; // Pr(m) is order M-m, 'r' means "reversed"
+  Pr(m) = par(i,M-m+1, (Pri(m,i) - nu(M-m+1)*Ari(m,M-m-i+1)));
+  Pri(m,i) = ba.take(i+1,Pr(m-1));
+  Ari(m,i) = ba.take(i+1,Ar(m-1));
+  //step-down recursion for lattice/ladder digital filters:
+  Ar(0) = (1,av); // Ar(m) is order M-m (recursion index must start at constant)
+  Ar(m) = 1,par(i,M-m, (Ari(m,i+1) - sr(m)*Ari(m,M-m-i))/(1-sr(m)*sr(m)));
+  sr(m) = Ari(m,M-m+1); // s_{M-1-m}
+};
+
+//--------------------`(fi.)iir_lat2`-----------------------
+// Two-multiply latice IIR filter of arbitrary order.
+//
+// #### Usage
+//
+// ```
+// _ : iir_lat2(bv,av) : _
+// ```
+//
+// Where:
+//
+// * bv: zeros as a bank of parallel signals
+// * av: poles as a bank of parallel signals
+//------------------------------------------------------------
+declare iir_lat2 author "Julius O. Smith III";
+declare iir_lat2 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare iir_lat2 license "MIT-style STK-4.3 license";
+iir_lat2(bv,av) = allpassnt(M,sv) : sum(i,M+1,*(ba.take(M-i+1,tg)))
+with {
+  M = ba.count(av);
+  sv = av2sv(av); // sv = vector of sin(theta) reflection coefficients
+  tg = bvav2nuv(bv,av); // tg = vector of tap gains
+};
+
+//-----------------------`(fi.)allpassnt`--------------------------
+// Two-multiply lattice allpass (nested order-1 direct-form-ii allpasses).
+//
+// #### Usage
+//
+// ```
+// _ : allpassnt(n,sv) : _
+// ```
+//
+// Where:
+//
+// * `n`: the order of the filter
+// * `sv`: the reflection coefficients (-1 1)
+//------------------------------------------------------------
+declare allpassnt author "Julius O. Smith III";
+declare allpassnt copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare allpassnt license "MIT-style STK-4.3 license";
+allpassnt(0,sv) = _;
+allpassnt(n,sv) = _ : ((+ <: (allpassnt(n-1,sv),*(s)))~*(-s)) : fsec(n)
+with {
+  fsec(1) = ro.crossnn(1) : _, (_<:mem,_) : +,_;
+  fsec(n) = ro.crossn1(n) : _, (_<:mem,_),par(i,n-1,_) : +, par(i,n,_);
+  innertaps(n) = par(i,n,_);
+  s = ba.take(n,sv); // reflection coefficient s = sin(theta)
+};
+
+//--------------------`(fi.)iir_kl`-----------------------
+// Kelly-Lochbaum ladder IIR filter of arbitrary order.
+//
+// #### Usage
+//
+// ```
+// _ : iir_kl(bv,av) : _
+// ```
+//
+// Where:
+//
+// * bv: zeros as a bank of parallel signals
+// * av: poles as a bank of parallel signals
+//------------------------------------------------------------
+declare iir_kl author "Julius O. Smith III";
+declare iir_kl copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare iir_kl license "MIT-style STK-4.3 license";
+iir_kl(bv,av) = allpassnklt(M,sv) : sum(i,M+1,*(tghr(i)))
+with {
+  M = ba.count(av);
+  sv = av2sv(av); // sv = vector of sin(theta) reflection coefficients
+  tg = bvav2nuv(bv,av); // tg = vector of tap gains for 2mul case
+  tgr(i) = ba.take(M+1-i,tg);
+  tghr(n) = tgr(n)/pi(n);
+  pi(0) = 1;
+  pi(n) = pi(n-1)*(1+ba.take(M-n+1,sv)); // all sign parameters '+'
+};
+
+//-----------------------`(fi.)allpassnklt`--------------------------
+// Kelly-Lochbaum ladder allpass.
+//
+// #### Usage:
+//
+// ```
+// _ : allpassnklt(n,sv) : _
+// ```
+//
+// Where:
+//
+// * `n`: the order of the filter
+// * `sv`: the reflection coefficients (-1 1)
+//------------------------------------------------------------
+declare allpassnklt author "Julius O. Smith III";
+declare allpassnklt copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare allpassnklt license "MIT-style STK-4.3 license";
+allpassnklt(0,sv) = _;
+allpassnklt(n,sv) = _ <: *(s),(*(1+s) : (+
+                   : allpassnklt(n-1,sv))~(*(-s))) : fsec(n)
+with {
+  fsec(1) = _, (_<:mem*(1-s),_) : sumandtaps(n);
+  fsec(n) = _, (_<:mem*(1-s),_), par(i,n-1,_) : sumandtaps(n);
+  s = ba.take(n,sv);
+  sumandtaps(n) = +,par(i,n,_);
+};
+
+//--------------------`(fi.)iir_lat1`-----------------------
+// One-multiply latice IIR filter of arbitrary order.
+//
+// #### Usage
+//
+// ```
+// _ : iir_lat1(bv,av) : _
+// ```
+//
+// Where:
+//
+// * bv: zeros as a bank of parallel signals
+// * av: poles as a bank of parallel signals
+//------------------------------------------------------------
+declare iir_lat1 author "Julius O. Smith III";
+declare iir_lat1 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare iir_lat1 license "MIT-style STK-4.3 license";
+iir_lat1(bv,av) = allpassn1mt(M,sv) : sum(i,M+1,*(tghr(i+1)))
+with {
+  M = ba.count(av);
+  sv = av2sv(av); // sv = vector of sin(theta) reflection coefficients
+  tg = bvav2nuv(bv,av); // tg = vector of tap gains
+  tgr(i) = ba.take(M+2-i,tg); // i=1..M+1 (for "takability")
+  tghr(n) = tgr(n)/pi(n);
+  pi(1) = 1;
+  pi(n) = pi(n-1)*(1+ba.take(M-n+2,sv)); // all sign parameters '+'
+};
+
+//-----------------------`(fi.)allpassn1mt`--------------------------
+// One-multiply lattice allpass with tap lines.
+//
+// #### Usage
+//
+// ```
+// _ : allpassn1mt(N,sv) : _
+// ```
+//
+// Where:
+//
+// * `N`: the order of the filter (fixed at compile time)
+// * `sv`: the reflection coefficients (-1 1)
+//------------------------------------------------------------
+declare allpassn1mt author "Julius O. Smith III";
+declare allpassn1mt copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare allpassn1mt license "MIT-style STK-4.3 license";
+allpassn1mt(0,sv) = _;
+allpassn1mt(n,sv) = _ <: _,_ : ((+:*(s) <: _,_),_ : _,+ : ro.crossnn(1)
+		  : allpassn1mt(n-1,sv),_)~(*(-1)) : fsec(n)
+with {
+  fsec(1) = ro.crossnn(1) : _, (_<:mem,_) : +,_;
+  fsec(n) = ro.crossn1(n) : _, (_<:mem,_),par(i,n-1,_) : +, par(i,n,_);
+  innertaps(n) = par(i,n,_);
+  s = ba.take(n,sv); // reflection coefficient s = sin(theta)
+};
+
+//-------------------------------`(fi.)iir_nl`-------------------------
+// Normalized ladder filter of arbitrary order.
+//
+// #### Usage
+//
+// ```
+// _ : iir_nl(bv,av) : _
+// ```
+//
+// Where:
+//
+// * bv: zeros as a bank of parallel signals
+// * av: poles as a bank of parallel signals
+//
+// #### References
+// * J. D. Markel and A. H. Gray, Linear Prediction of Speech, New York: Springer Verlag, 1976.
+// * <https://ccrma.stanford.edu/~jos/pasp/Normalized_Scattering_Junctions.html>
+//------------------------------------------------------------
+declare iir_nl author "Julius O. Smith III";
+declare iir_nl copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare iir_nl license "MIT-style STK-4.3 license";
+iir_nl(bv,av) = allpassnnlt(M,sv) : sum(i,M+1,*(tghr(i)))
+with {
+  M = ba.count(av);
+  sv = av2sv(av); // sv = vector of sin(theta) reflection coefficients
+  tg = bvav2nuv(bv,av); // tg = vector of tap gains for 2mul case
+  tgr(i) = ba.take(M+1-i,tg);
+  tghr(n) = tgr(n)/pi(n);
+  pi(0) = 1;
+  s(n) = ba.take(M-n+1,sv); // reflection coefficient = sin(theta)
+  c(n) = sqrt(max(0,1-s(n)*s(n))); // compiler crashes on sqrt(-)
+  pi(n) = pi(n-1)*c(n);
+};
+
+//-------------------------------`(fi.)allpassnnlt`-------------------------
+// Normalized ladder allpass filter of arbitrary order.
+//
+// #### Usage:
+//
+// ```
+// _ : allpassnnlt(N,sv) : _
+// ```
+//
+// Where:
+//
+// * `N`: the order of the filter (fixed at compile time)
+// * `sv`: the reflection coefficients (-1,1)
+//
+// #### References
+// * J. D. Markel and A. H. Gray, Linear Prediction of Speech, New York: Springer Verlag, 1976.
+// * <https://ccrma.stanford.edu/~jos/pasp/Normalized_Scattering_Junctions.html>
+//------------------------------------------------------------
+declare allpassnnlt author "Julius O. Smith III";
+declare allpassnnlt copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare allpassnnlt license "MIT-style STK-4.3 license";
+allpassnnlt(0,sv) = _;
+allpassnnlt(n,scl*(sv)) = allpassnnlt(n,par(i,count(sv),scl*(sv(i))));
+allpassnnlt(n,sv) = _ <: *(s),(*(c) : (+
+                   : allpassnnlt(n-1,sv))~(*(-s))) : fsec(n)
+with {
+  fsec(1) = _, (_<:mem*(c),_) : sumandtaps(n);
+  fsec(n) = _, (_<:mem*(c),_), par(i,n-1,_) : sumandtaps(n);
+  s = ba.take(n,sv);
+  c = sqrt(max(0,1-s*s));
+  sumandtaps(n) = +,par(i,n,_);
+};
+
+//=============================Useful Special Cases=======================================
+//========================================================================================
+
+//--------------------------------`(fi.)tf2np`------------------------------------
+// Biquad based on a stable second-order Normalized Ladder Filter
+// (more robust to modulation than `tf2` and protected against instability).
+//
+// #### Usage
+//
+// ```
+// _ : tf2np(b0,b1,b2,a1,a2) : _
+// ```
+//
+// Where:
+//
+// * `a`: the poles
+// * `b`: the zeros
+//------------------------------------------------------------
+declare tf2np author "Julius O. Smith III";
+declare tf2np copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare tf2np license "MIT-style STK-4.3 license";
+tf2np(b0,b1,b2,a1,a2) = allpassnnlt(M,sv) : sum(i,M+1,*(tghr(i)))
+with {
+  smax = 1.0-ma.EPSILON; // maximum reflection-coefficient magnitude allowed
+  s2 = max(-smax, min(smax,a2)); // Project both reflection-coefficients
+  s1 = max(-smax, min(smax,a1/(1+a2))); // into the defined stability-region.
+  sv = (s1,s2); // vector of sin(theta) reflection coefficients
+  M = 2;
+  nu(2) = b2;
+  nu(1) = b1 - b2*a1;
+  nu(0) = (b0-b2*a2) - nu(1)*s1;
+  tg = (nu(0),nu(1),nu(2));
+  tgr(i) = ba.take(M+1-i,tg); // vector of tap gains for 2mul case
+  tghr(n) = tgr(n)/pi(n);  // apply pi parameters for NLF case
+  pi(0) = 1;
+  s(n) = ba.take(M-n+1,sv);
+  c(n) = sqrt(1-s(n)*s(n));
+  pi(n) = pi(n-1)*c(n);
+};
+
+//-----------------------------`(fi.)wgr`---------------------------------
+// Second-order transformer-normalized digital waveguide resonator.
+//
+// #### Usage
+//
+// ```
+// _ : wgr(f,r) : _
+// ```
+//
+// Where:
+//
+// * `f`: resonance frequency (Hz)
+// * `r`: loss factor for exponential decay (set to 1 to make a numerically stable oscillator)
+//
+// #### References
+// * <https://ccrma.stanford.edu/~jos/pasp/Power_Normalized_Waveguide_Filters.html>
+// * <https://ccrma.stanford.edu/~jos/pasp/Digital_Waveguide_Oscillator.html>
+//------------------------------------------------------------
+declare wgr author "Julius O. Smith III";
+declare wgr copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare wgr license "MIT-style STK-4.3 license";
+wgr(f,r,x) = (*(G),_<:_,((+:*(C))<:_,_),_:+,_,_:+(x),-) ~ cross : _,*(0-gi)
+with {
+  C = cos(2*ma.PI*f/ma.SR);
+  gi = sqrt(max(0,(1+C)/(1-C))); // compensate amplitude (only needed when
+  G = r*(1-1' + gi')/gi;         // frequency changes substantially)
+  cross = _,_ <: !,_,_,!;
+};
+
+//-----------------------------`(fi.)nlf2`--------------------------------
+// Second order normalized digital waveguide resonator.
+//
+// #### Usage
+//
+// ```
+// _ : nlf2(f,r) : _
+// ```
+//
+// Where:
+//
+// * `f`: resonance frequency (Hz)
+// * `r`: loss factor for exponential decay (set to 1 to make a sinusoidal oscillator)
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/pasp/Power_Normalized_Waveguide_Filters.html>
+//------------------------------------------------------------
+declare nlf2 author "Julius O. Smith III";
+declare nlf2 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare nlf2 license "MIT-style STK-4.3 license";
+nlf2(f,r,x) = ((_<:_,_),(_<:_,_) : (*(s),*(c),*(c),*(0-s)) :>
+              (*(r),+(x))) ~ cross
+with {
+  th = 2*ma.PI*f/ma.SR;
+  c = cos(th);
+  s = sin(th);
+  cross = _,_ <: !,_,_,!;
+};
+
+
+//------------`(fi.)apnl`---------------
+// Passive Nonlinear Allpass based on Pierce switching springs idea.
+// Switch between allpass coefficient `a1` and `a2` at signal zero crossings.
+//
+// #### Usage
+//
+// ```
+// _ : apnl(a1,a2) : _
+// ```
+//
+// Where:
+//
+// * `a1` and `a2`: allpass coefficients
+//
+// #### Reference
+// * "A Passive Nonlinear Digital Filter Design ..." by John R. Pierce and Scott
+// A. Van Duyne, JASA, vol. 101, no. 2, pp. 1120-1126, 1997
+//------------------------------------------------------------
+declare apnl author "Julius O. Smith III";
+declare apnl copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare apnl license "MIT-style STK-4.3 license";
+apnl(a1,a2,x) = nonLinFilter
+with {
+   condition = _>0;
+   nonLinFilter = (x - _ <: _*(condition*a1 + (1-condition)*a2),_')~_ :> +;
+};
+
+
+//============================Ladder/Lattice Allpass Filters==============================
+// An allpass filter has gain 1 at every frequency, but variable phase.
+// Ladder/lattice allpass filters are specified by reflection coefficients.
+// They are defined here as nested allpass filters, hence the names allpassn*.
+//
+// #### References
+// * <https://ccrma.stanford.edu/~jos/pasp/Conventional_Ladder_Filters.html>
+// * <https://ccrma.stanford.edu/~jos/pasp/Nested_Allpass_Filters.html>
+// * Linear Prediction of Speech, Markel and Gray, Springer Verlag, 1976
+//========================================================================================
+
+//---------------`(fi.)allpassn`-----------------
+// Two-multiply lattice - each section is two multiply-adds.
+//
+// #### Usage:
+//
+// ```
+// _ : allpassn(n,sv) : _
+// ```
+// #### Where:
+//
+// * `n`: the order of the filter
+// * `sv`: the reflection coefficients (-1 1)
+//
+// #### References
+// * J. O. Smith and R. Michon, "Nonlinear Allpass Ladder Filters in FAUST", in
+// Proceedings of the 14th International Conference on Digital Audio Effects
+// (DAFx-11), Paris, France, September 19-23, 2011.
+//----------------------------------------------
+declare allpassn author "Julius O. Smith III and Romain Michon";
+declare allpassn copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu> and Romain Michon <rmichon@ccrma.stanford.edu>";
+declare allpassn license "MIT-style STK-4.3 license";
+allpassn(0,sv) = _;
+allpassn(n,sv) = _ <: ((+ <: (allpassn(n-1,sv)),*(s))~(*(-s))) : _',_ :+
+with { s = ba.take(n,sv); };
+
+//---------------`(fi.)allpassnn`-----------------
+// Normalized form - four multiplies and two adds per section,
+// but coefficients can be time varying and nonlinear without
+// "parametric amplification" (modulation of signal energy).
+//
+// #### Usage:
+//
+// ```
+// _ : allpassnn(n,tv) : _
+// ```
+//
+// Where:
+//
+// * `n`: the order of the filter
+// * `tv`: the reflection coefficients (-PI PI)
+//----------------------------------------------
+// power-normalized (reflection coefficients s = sin(t)):
+declare allpassnn author "Julius O. Smith III";
+declare allpassnn copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare allpassnn license "MIT-style STK-4.3 license";
+allpassnn(0,tv) = _;
+allpassnn(n,tv) = _ <: *(s), (*(c) : (+
+        : allpassnn(n-1,tv))~(*(-s))) : _, mem*c : +
+with { c = cos(ba.take(n,tv));  s = sin(ba.take(n,tv)); };
+
+//---------------`(fi.)allpassnkl`-----------------
+// Kelly-Lochbaum form - four multiplies and two adds per
+// section, but all signals have an immediate physical
+// interpretation as traveling pressure waves, etc.
+//
+// #### Usage:
+//
+// ```
+// _ : allpassnkl(n,sv) : _
+// ```
+//
+// Where:
+//
+// * `n`: the order of the filter
+// * `sv`: the reflection coefficients (-1 1)
+//----------------------------------------------
+// Kelly-Lochbaum:
+declare allpassnnkl author "Julius O. Smith III";
+declare allpassnnkl copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare allpassnnkl license "MIT-style STK-4.3 license";
+allpassnkl(0,sv) = _;
+allpassnkl(n,sv) = _ <: *(s),(*(1+s) : (+
+                   : allpassnkl(n-1,sv))~(*(-s))) : _, mem*(1-s) : +
+with { s = ba.take(n,sv); };
+
+//---------------`(fi.)allpass1m`-----------------
+// One-multiply form - one multiply and three adds per section.
+// Normally the most efficient in special-purpose hardware.
+//
+// #### Usage:
+//
+// ```
+// _ : allpassn1m(n,sv) : _
+// ```
+//
+// Where:
+//
+// * `n`: the order of the filter
+// * `sv`: the reflection coefficients (-1 1)
+//----------------------------------------------
+// one-multiply:
+declare allpassn1m author "Julius O. Smith III";
+declare allpassn1m copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare allpassn1m license "MIT-style STK-4.3 license";
+allpassn1m(0,sv) = _;
+allpassn1m(n,sv) = _ <: _,_ : ((+:*(s) <: _,_),_ : _,+ : cross
+		  : allpassn1m(n-1,sv),_)~(*(-1)) : _',_ : +
+with { s = ba.take(n,sv); cross = _,_ <: !,_,_,!; };
+
+//===========Digital Filter Sections Specified as Analog Filter Sections==================
+//========================================================================================
+
+//-------------------------`(fi.)tf2s` and `(fi.)tf2snp`--------------------------------
+// Second-order direct-form digital filter,
+// specified by ANALOG transfer-function polynomials B(s)/A(s),
+// and a frequency-scaling parameter. Digitization via the
+// bilinear transform is built in.
+//
+// #### Usage
+//
+// ```
+// _ : tf2s(b2,b1,b0,a1,a0,w1) : _
+// ```
+// Where:
+//
+// ```
+//         b2 s^2 + b1 s + b0
+// H(s) = --------------------
+//            s^2 + a1 s + a0
+// ```
+//
+// and `w1` is the desired digital frequency (in radians/second)
+// corresponding to analog frequency 1 rad/sec (i.e., `s = j`).
+//
+// #### Example test program
+//
+// A second-order ANALOG Butterworth lowpass filter,
+// normalized to have cutoff frequency at 1 rad/sec,
+// has transfer function:
+//
+// ```
+//              1
+// H(s) = -----------------
+//         s^2 + a1 s + 1
+// ```
+//
+// where `a1 = sqrt(2)`. Therefore, a DIGITAL Butterworth lowpass
+// cutting off at `SR/4` is specified as `tf2s(0,0,1,sqrt(2),1,PI*SR/2);`
+//
+// #### Method
+//
+// Bilinear transform scaled for exact mapping of w1.
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/pasp/Bilinear_Transformation.html>
+//----------------------------------------------
+declare tf2s author "Julius O. Smith III";
+declare tf2s copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare tf2s license "MIT-style STK-4.3 license";
+tf2s(b2,b1,b0,a1,a0,w1) = tf2(b0d,b1d,b2d,a1d,a2d)
+with {
+  c   = 1/tan(w1*0.5/ma.SR); // bilinear-transform scale-factor
+  csq = c*c;
+  d   = a0 + a1 * c + csq;
+  b0d = (b0 + b1 * c + b2 * csq)/d;
+  b1d = 2 * (b0 - b2 * csq)/d;
+  b2d = (b0 - b1 * c + b2 * csq)/d;
+  a1d = 2 * (a0 - csq)/d;
+  a2d = (a0 - a1*c + csq)/d;
+};
+
+// tf2snp = tf2s but using a protected normalized ladder filter for tf2:
+tf2snp(b2,b1,b0,a1,a0,w1) = tf2np(b0d,b1d,b2d,a1d,a2d)
+with {
+  c   = 1/tan(w1*0.5/ma.SR); // bilinear-transform scale-factor
+  csq = c*c;
+  d   = a0 + a1 * c + csq;
+  b0d = (b0 + b1 * c + b2 * csq)/d;
+  b1d = 2 * (b0 - b2 * csq)/d;
+  b2d = (b0 - b1 * c + b2 * csq)/d;
+  a1d = 2 * (a0 - csq)/d;
+  a2d = (a0 - a1*c + csq)/d;
+};
+
+//-----------------------------`(fi.)tf1snp`-------------------------------
+// First-order special case of tf2snp above.
+//
+// #### Usage
+//
+// ```
+// _ : tf1snp(b1,b0,a0) : _
+// ```
+//----------------------------------------------
+declare tf1snp author "Julius O. Smith III";
+declare tf1snp copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare tf1snp license "MIT-style STK-4.3 license";
+tf1snp(b1,b0,a0,w1) = fi.tf2snp(b1,b0,0,a0,0,w1); // FIXME: Faust compiler does not fully optimize - does C++?
+
+//-----------------------------`(fi.)tf3slf`-------------------------------
+// Analogous to `tf2s` above, but third order, and using the typical
+// low-frequency-matching bilinear-transform constant 2/T ("lf" series)
+// instead of the specific-frequency-matching value used in `tf2s` and `tf1s`.
+// Note the lack of a "w1" argument.
+//
+// #### Usage
+//
+// ```
+// _ : tf3slf(b3,b2,b1,b0,a3,a2,a1,a0) : _
+// ```
+//----------------------------------------------
+declare tf3slf author "Julius O. Smith III";
+declare tf3slf copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare tf3slf license "MIT-style STK-4.3 license";
+tf3slf(b3,b2,b1,b0,a3,a2,a1,a0) = tf3(b0d,b1d,b2d,b3d,a1d,a2d,a3d) with {
+  c   = 2.0 * ma.SR; // bilinear-transform scale-factor ("lf" case)
+  csq = c*c;
+  cc  = csq*c;
+  // Thank you maxima:
+  b3d = (b3*c^3-b2*c^2+b1*c-b0)/d;
+  b2d = (-3*b3*c^3+b2*c^2+b1*c-3*b0)/d;
+  b1d = (3*b3*c^3+b2*c^2-b1*c-3*b0)/d;
+  b0d = (-b3*c^3-b2*c^2-b1*c-b0)/d;
+  a3d = (a3*c^3-a2*c^2+a1*c-a0)/d;
+  a2d = (-3*a3*c^3+a2*c^2+a1*c-3*a0)/d;
+  a1d = (3*a3*c^3+a2*c^2-a1*c-3*a0)/d;
+  d = (-a3*c^3-a2*c^2-a1*c-a0);
+};
+
+//-----------------------------`(fi.)tf1s`--------------------------------
+// First-order direct-form digital filter,
+// specified by ANALOG transfer-function polynomials B(s)/A(s),
+// and a frequency-scaling parameter.
+//
+// #### Usage
+//
+// ```
+// _ : tf1s(b1,b0,a0,w1) : _
+// ```
+// Where:
+//
+//        b1 s + b0
+// H(s) = ----------
+//           s + a0
+//
+// and `w1` is the desired digital frequency (in radians/second)
+// corresponding to analog frequency 1 rad/sec (i.e., `s = j`).
+//
+// #### Example test program
+//
+// A first-order ANALOG Butterworth lowpass filter,
+// normalized to have cutoff frequency at 1 rad/sec,
+// has transfer function:
+//
+//           1
+// H(s) = -------
+//         s + 1
+//
+// so `b0 = a0 = 1` and `b1 = 0`.  Therefore, a DIGITAL first-order
+// Butterworth lowpass with gain -3dB at `SR/4` is specified as
+//
+// ```
+// tf1s(0,1,1,PI*SR/2); // digital half-band order 1 Butterworth
+// ```
+//
+// #### Method
+//
+// Bilinear transform scaled for exact mapping of w1.
+//
+// #### Reference
+// <https://ccrma.stanford.edu/~jos/pasp/Bilinear_Transformation.html>
+//----------------------------------------------
+declare tf1s author "Julius O. Smith III";
+declare tf1s copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare tf1s license "MIT-style STK-4.3 license";
+tf1s(b1,b0,a0,w1) = tf1(b0d,b1d,a1d)
+with {
+  c   = 1/tan(w1*0.5/ma.SR); // bilinear-transform scale-factor
+  d   = a0 + c;
+  b1d = (b0 - b1*c) / d;
+  b0d = (b0 + b1*c) / d;
+  a1d = (a0 - c) / d;
+};
+
+//-----------------------------`(fi.)tf2sb`--------------------------------
+// Bandpass mapping of `tf2s`: In addition to a frequency-scaling parameter
+// `w1` (set to HALF the desired passband width in rad/sec),
+// there is a desired center-frequency parameter wc (also in rad/s).
+// Thus, `tf2sb` implements a fourth-order digital bandpass filter section
+// specified by the coefficients of a second-order analog lowpass prototype
+// section.  Such sections can be combined in series for higher orders.
+// The order of mappings is (1) frequency scaling (to set lowpass cutoff w1),
+// (2) bandpass mapping to wc, then (3) the bilinear transform, with the
+// usual scale parameter `2*SR`.  Algebra carried out in maxima and pasted here.
+//
+// #### Usage
+//
+// ```
+// _ : tf2sb(b2,b1,b0,a1,a0,w1,wc) : _
+// ```
+//----------------------------------------------
+declare tf2sb author "Julius O. Smith III";
+declare tf2sb copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare tf2sb license "MIT-style STK-4.3 license";
+tf2sb(b2,b1,b0,a1,a0,w1,wc) =
+  iir((b0d/a0d,b1d/a0d,b2d/a0d,b3d/a0d,b4d/a0d),(a1d/a0d,a2d/a0d,a3d/a0d,a4d/a0d)) with {
+  T = 1.0/float(ma.SR);
+  b0d = (4*b0*w1^2+8*b2*wc^2)*T^2+8*b1*w1*T+16*b2;
+  b1d = 4*b2*wc^4*T^4+4*b1*wc^2*w1*T^3-16*b1*w1*T-64*b2;
+  b2d = 6*b2*wc^4*T^4+(-8*b0*w1^2-16*b2*wc^2)*T^2+96*b2;
+  b3d = 4*b2*wc^4*T^4-4*b1*wc^2*w1*T^3+16*b1*w1*T-64*b2;
+  b4d = (b2*wc^4*T^4-2*b1*wc^2*w1*T^3+(4*b0*w1^2+8*b2*wc^2)*T^2-8*b1*w1*T+16*b2)
+        + b2*wc^4*T^4+2*b1*wc^2*w1*T^3;
+  a0d = wc^4*T^4+2*a1*wc^2*w1*T^3+(4*a0*w1^2+8*wc^2)*T^2+8*a1*w1*T+16;
+  a1d = 4*wc^4*T^4+4*a1*wc^2*w1*T^3-16*a1*w1*T-64;
+  a2d = 6*wc^4*T^4+(-8*a0*w1^2-16*wc^2)*T^2+96;
+  a3d = 4*wc^4*T^4-4*a1*wc^2*w1*T^3+16*a1*w1*T-64;
+  a4d = wc^4*T^4-2*a1*wc^2*w1*T^3+(4*a0*w1^2+8*wc^2)*T^2-8*a1*w1*T+16;
+};
+
+//-----------------------------`(fi.)tf1sb`--------------------------------
+// First-to-second-order lowpass-to-bandpass section mapping,
+// analogous to tf2sb above.
+//
+// #### Usage
+//
+// ```
+// _ : tf1sb(b1,b0,a0,w1,wc) : _
+// ```
+//----------------------------------------------
+declare tf1sb author "Julius O. Smith III";
+declare tf1sb copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare tf1sb license "MIT-style STK-4.3 license";
+tf1sb(b1,b0,a0,w1,wc) = tf2(b0d/a0d,b1d/a0d,b2d/a0d,a1d/a0d,a2d/a0d) with {
+  T = 1.0/float(ma.SR);
+  a0d = wc^2*T^2+2*a0*w1*T+4;
+  b0d = b1*wc^2*T^2 +2*b0*w1*T+4*b1;
+  b1d = 2*b1*wc^2*T^2-8*b1;
+  b2d = b1*wc^2*T^2-2*b0*w1*T+4*b1;
+  a1d = 2*wc^2*T^2-8;
+  a2d = wc^2*T^2-2*a0*w1*T+4;
+};
+
+//==============================Simple Resonator Filters==================================
+//========================================================================================
+
+//------------------`(fi.)resonlp`-----------------
+// Simple resonant lowpass filter based on `tf2s` (virtual analog).
+// `resonlp` is a standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : resonlp(fc,Q,gain) : _
+// _ : resonhp(fc,Q,gain) : _
+// _ : resonbp(fc,Q,gain) : _
+//
+// ```
+//
+// Where:
+//
+// * `fc`: center frequency (Hz)
+// * `Q`: q
+// * `gain`: gain (0-1)
+//---------------------------------------------------------------------
+// resonlp = 2nd-order lowpass with corner resonance:
+declare resonlp author "Julius O. Smith III";
+declare resonlp copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare resonlp license "MIT-style STK-4.3 license";
+resonlp(fc,Q,gain) = tf2s(b2,b1,b0,a1,a0,wc)
+with {
+     wc = 2*ma.PI*fc;
+     a1 = 1/Q;
+     a0 = 1;
+     b2 = 0;
+     b1 = 0;
+     b0 = gain;
+};
+
+
+//------------------`(fi.)resonhp`-----------------
+// Simple resonant highpass filters based on `tf2s` (virtual analog).
+// `resonhp` is a standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : resonlp(fc,Q,gain) : _
+// _ : resonhp(fc,Q,gain) : _
+// _ : resonbp(fc,Q,gain) : _
+//
+// ```
+//
+// Where:
+//
+// * `fc`: center frequency (Hz)
+// * `Q`: q
+// * `gain`: gain (0-1)
+//---------------------------------------------------------------------
+// resonhp = 2nd-order highpass with corner resonance:
+declare resonhp author "Julius O. Smith III";
+declare resonhp copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare resonhp license "MIT-style STK-4.3 license";
+resonhp(fc,Q,gain,x) = gain*x-resonlp(fc,Q,gain,x);
+
+
+//------------------`(fi.)resonbp`-----------------
+// Simple resonant bandpass filters based on `tf2s` (virtual analog).
+// `resonbp` is a standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : resonlp(fc,Q,gain) : _
+// _ : resonhp(fc,Q,gain) : _
+// _ : resonbp(fc,Q,gain) : _
+//
+// ```
+//
+// Where:
+//
+// * `fc`: center frequency (Hz)
+// * `Q`: q
+// * `gain`: gain (0-1)
+//---------------------------------------------------------------------
+// resonbp = 2nd-order bandpass
+declare resonbp author "Julius O. Smith III";
+declare resonbp copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare resonbp license "MIT-style STK-4.3 license";
+resonbp(fc,Q,gain) = tf2s(b2,b1,b0,a1,a0,wc)
+with {
+     wc = 2*ma.PI*fc;
+     a1 = 1/Q;
+     a0 = 1;
+     b2 = 0;
+     b1 = gain;
+     b0 = 0;
+};
+
+
+//======================Butterworth Lowpass/Highpass Filters==============================
+//========================================================================================
+
+//----------------`(fi.)lowpass`--------------------
+// Nth-order Butterworth lowpass filter.
+// `lowpass` is a standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : lowpass(N,fc) : _
+// ```
+//
+// Where:
+//
+// * `N`: filter order (number of poles), nonnegative constant numerical expression
+// * `fc`: desired cut-off frequency (-3dB frequency) in Hz
+//
+// #### References
+// * <https://ccrma.stanford.edu/~jos/filters/Butterworth_Lowpass_Design.html>
+// * `butter` function in Octave `("[z,p,g] = butter(N,1,'s');")`
+//------------------------------
+declare lowpass author "Julius O. Smith III";
+declare lowpass copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare lowpass license "MIT-style STK-4.3 license";
+lowpass(N,fc) = lowpass0_highpass1(0,N,fc);
+
+
+//----------------`(fi.)highpass`--------------------
+// Nth-order Butterworth highpass filters.
+// `highpass` is a standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : highpass(N,fc) : _
+// ```
+//
+// Where:
+//
+// * `N`: filter order (number of poles), nonnegative constant numerical expression
+// * `fc`: desired cut-off frequency (-3dB frequency) in Hz
+//
+// #### References
+// * <https://ccrma.stanford.edu/~jos/filters/Butterworth_Lowpass_Design.html>
+// * `butter` function in Octave `("[z,p,g] = butter(N,1,'s');")`
+//------------------------------
+declare highpass author "Julius O. Smith III";
+declare highpass copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare higpass license "MIT-style STK-4.3 license";
+highpass(N,fc) = lowpass0_highpass1(1,N,fc);
+
+
+//-------------`(fi.)lowpass0_highpass1`--------------
+declare lowpass0_highpass1 author "Julius O. Smith III";
+declare lowpass0_highpass1 "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare lowpass0_highpass1 "MIT-style STK-4.3 license";
+lowpass0_highpass1(s,N,fc) = lphpr(s,N,N,fc)
+with {
+  lphpr(s,0,N,fc) = _;
+  lphpr(s,1,N,fc) = tf1s(s,1-s,1,2*ma.PI*fc);
+  lphpr(s,O,N,fc) = lphpr(s,(O-2),N,fc) : tf2s(s,0,1-s,a1s,1,w1) with {
+    parity = N % 2;
+    S = (O-parity)/2; // current section number
+    a1s = -2*cos((ma.PI)*-1 + (1-parity)*ma.PI/(2*N) + (S-1+parity)*ma.PI/N);
+    w1 = 2*ma.PI*fc;
+  };
+};
+
+
+//================Special Filter-Bank Delay-Equalizing Allpass Filters====================
+// These special allpass filters are needed by filterbank et al. below.
+// They are equivalent to (`lowpass(N,fc)` +|- `highpass(N,fc))/2`, but with
+// canceling pole-zero pairs removed (which occurs for odd N).
+//========================================================================================
+
+//--------------------`(fi.)lowpass_plus`|`minus_highpass`----------------
+declare highpass_plus_lowpass author "Julius O. Smith III";
+declare highpass_plus_lowpass copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare highpass_plus_lowpass license "MIT-style STK-4.3 license";
+highpass_plus_lowpass(1,fc) = _;
+highpass_plus_lowpass(3,fc) = tf2s(1,-1,1,1,1,w1) with { w1 = 2*ma.PI*fc; };
+highpass_plus_lowpass(5,fc) = tf2s(1,-a11,1,a11,1,w1)
+with {
+  a11 = 1.618033988749895;
+  w1 = 2*ma.PI*fc;
+};
+
+// Catch-all definitions for generality - even order is done:
+highpass_plus_lowpass(N,fc) = _ <: switch_odd_even(N%2,N,fc) with {
+  switch_odd_even(0,N,fc) = highpass_plus_lowpass_even(N,fc);
+  switch_odd_even(1,N,fc) = highpass_plus_lowpass_odd(N,fc);
+};
+
+declare highpass_minus_lowpass author "Julius O. Smith III";
+declare highpass_minus_lowpass copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare highpass_minus_lowpass license "MIT-style STK-4.3 license";
+highpass_minus_lowpass(3,fc) = tf1s(-1,1,1,w1) with { w1 = 2*ma.PI*fc; };
+highpass_minus_lowpass(5,fc) = tf1s(1,-1,1,w1) : tf2s(1,-a12,1,a12,1,w1)
+with {
+  a12 = 0.618033988749895;
+  w1 = 2*ma.PI*fc;
+};
+
+// Catch-all definitions for generality - even order is done:
+highpass_minus_lowpass(N,fc) = _ <: switch_odd_even(N%2,N,fc) with {
+  switch_odd_even(0,N,fc) = highpass_minus_lowpass_even(N,fc);
+  switch_odd_even(1,N,fc) = highpass_minus_lowpass_odd(N,fc);
+};
+
+declare highpass_plus_lowpass_even author "Julius O. Smith III";
+declare highpass_plus_lowpass_even copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare highpass_plus_lowpass_even license "MIT-style STK-4.3 license";
+highpass_plus_lowpass_even(N,fc) = highpass(N,fc) + lowpass(N,fc);
+
+declare highpass_minus_lowpass_even author "Julius O. Smith III";
+declare highpass_minus_lowpass_even copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare highpass_plus_lowpass_even license "MIT-style STK-4.3 license";
+highpass_minus_lowpass_even(N,fc) = highpass(N,fc) - lowpass(N,fc);
+
+declare highpass_plus_lowpass_odd author "Julius O. Smith III";
+declare highpass_plus_lowpass_odd copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare highpass_plus_lowpass_odd license "MIT-style STK-4.3 license";
+// FIXME: Rewrite the following, as for orders 3 and 5 above,
+//        to eliminate pole-zero cancellations:
+highpass_plus_lowpass_odd(N,fc) = highpass(N,fc) + lowpass(N,fc);
+
+declare highpass_minus_lowpass_odd author "Julius O. Smith III";
+declare highpass_minus_lowpass_odd copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare highpass_plus_lowpass_odd license "MIT-style STK-4.3 license";
+// FIXME: Rewrite the following, as for orders 3 and 5 above,
+//        to eliminate pole-zero cancellations:
+highpass_minus_lowpass_odd(N,fc) = highpass(N,fc) - lowpass(N,fc);
+
+
+//==========================Elliptic (Cauer) Lowpass Filters==============================
+// Elliptic (Cauer) Lowpass Filters
+//
+// #### References
+// * <http://en.wikipedia.org/wiki/Elliptic_filter>
+// * functions `ncauer` and `ellip` in Octave.
+//========================================================================================
+
+//-----------------------------`(fi.)lowpass3e`-----------------------------
+// Third-order Elliptic (Cauer) lowpass filter.
+//
+// #### Usage
+//
+// ```
+// _ : lowpass3e(fc) : _
+// ```
+//
+// Where:
+//
+// * `fc`: -3dB frequency in Hz
+//
+// #### Design
+//
+// For spectral band-slice level display (see `octave_analyzer3e`):
+//
+// ```
+// [z,p,g] = ncauer(Rp,Rs,3);  % analog zeros, poles, and gain, where
+// Rp = 60  % dB ripple in stopband
+// Rs = 0.2 % dB ripple in passband
+// ```
+//---------------------------------------------------------------------
+declare lowpass3e author "Julius O. Smith III";
+declare lowpass3e copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare lowpass3e license "MIT-style STK-4.3 license";
+lowpass3e(fc) = tf2s(b21,b11,b01,a11,a01,w1) : tf1s(0,1,a02,w1)
+with {
+  a11 = 0.802636764161030; // format long; poly(p(1:2)) % in octave
+  a01 = 1.412270893774204;
+  a02 = 0.822445908998816; // poly(p(3)) % in octave
+  b21 = 0.019809144837789; // poly(z)
+  b11 = 0;
+  b01 = 1.161516418982696;
+  w1 = 2*ma.PI*fc;
+};
+
+//-----------------------------`(fi.)lowpass6e`-----------------------------
+// Sixth-order Elliptic/Cauer lowpass filter.
+//
+// #### Usage
+//
+// ```
+// _ : lowpass6e(fc) : _
+// ```
+//
+// Where:
+//
+// * `fc`: -3dB frequency in Hz
+//
+// #### Design
+//
+// For spectral band-slice level display (see octave_analyzer6e):
+//
+// ```
+// [z,p,g] = ncauer(Rp,Rs,6);  % analog zeros, poles, and gain, where
+//  Rp = 80  % dB ripple in stopband
+//  Rs = 0.2 % dB ripple in passband
+// ```
+//----------------------------------------------------------------------
+declare lowpass6e author "Julius O. Smith III";
+declare lowpass6e copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare lowpass6e license "MIT-style STK-4.3 license";
+lowpass6e(fc) =
+              tf2s(b21,b11,b01,a11,a01,w1) :
+              tf2s(b22,b12,b02,a12,a02,w1) :
+              tf2s(b23,b13,b03,a13,a03,w1)
+with {
+  b21 = 0.000099999997055;
+  a21 = 1;
+  b11 = 0;
+  a11 = 0.782413046821645;
+  b01 = 0.000433227200555;
+  a01 = 0.245291508706160;
+  b22 = 1;
+  a22 = 1;
+  b12 = 0;
+  a12 = 0.512478641889141;
+  b02 = 7.621731298870603;
+  a02 = 0.689621364484675;
+  b23 = 1;
+  a23 = 1;
+  b13 = 0;
+  a13 = 0.168404871113589;
+  b03 = 53.536152954556727;
+  a03 = 1.069358407707312;
+  w1 = 2*ma.PI*fc;
+};
+
+
+//=========================Elliptic Highpass Filters======================================
+//========================================================================================
+
+//-----------------------------`(fi.)highpass3e`-----------------------------
+// Third-order Elliptic (Cauer) highpass filter. Inversion of `lowpass3e` wrt unit
+// circle in s plane (s <- 1/s).
+//
+// #### Usage
+//
+// ```
+// _ : highpass3e(fc) : _
+// ```
+//
+// Where:
+//
+// * `fc`: -3dB frequency in Hz
+//-------------------------------------------------------------------------
+declare highpass3e author "Julius O. Smith III";
+declare highpass3e copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare highpass3e license "MIT-style STK-4.3 license";
+highpass3e(fc) = tf2s(b01/a01,b11/a01,b21/a01,a11/a01,1/a01,w1) :
+                 tf1s(1/a02,0,1/a02,w1)
+with {
+  a11 = 0.802636764161030;
+  a01 = 1.412270893774204;
+  a02 = 0.822445908998816;
+  b21 = 0.019809144837789;
+  b11 = 0;
+  b01 = 1.161516418982696;
+  w1 = 2*ma.PI*fc;
+};
+
+//-----------------------------`(fi.)highpass6e`-----------------------------
+// Sixth-order Elliptic/Cauer highpass filter. Inversion of `lowpass3e` wrt unit
+// circle in s plane (s <- 1/s).
+//
+// #### Usage
+//
+// ```
+// _ : highpass6e(fc) : _
+// ```
+//
+// Where:
+//
+// * `fc`: -3dB frequency in Hz
+//-------------------------------------------------------------------------
+declare highpass6e author "Julius O. Smith III";
+declare highpass6e copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare highpass6e license "MIT-style STK-4.3 license";
+highpass6e(fc) =
+              tf2s(b01/a01,b11/a01,b21/a01,a11/a01,1/a01,w1) :
+              tf2s(b02/a02,b12/a02,b22/a02,a12/a02,1/a02,w1) :
+              tf2s(b03/a03,b13/a03,b23/a03,a13/a03,1/a03,w1)
+with {
+  b21 = 0.000099999997055;
+  a21 = 1;
+  b11 = 0;
+  a11 = 0.782413046821645;
+  b01 = 0.000433227200555;
+  a01 = 0.245291508706160;
+  b22 = 1;
+  a22 = 1;
+  b12 = 0;
+  a12 = 0.512478641889141;
+  b02 = 7.621731298870603;
+  a02 = 0.689621364484675;
+  b23 = 1;
+  a23 = 1;
+  b13 = 0;
+  a13 = 0.168404871113589;
+  b03 = 53.536152954556727;
+  a03 = 1.069358407707312;
+  w1 = 2*ma.PI*fc;
+};
+
+
+//========================Butterworth Bandpass/Bandstop Filters===========================
+//========================================================================================
+
+//--------------------`(fi.)bandpass`----------------
+// Order 2*Nh Butterworth bandpass filter made using the transformation
+// `s <- s + wc^2/s` on `lowpass(Nh)`, where `wc` is the desired bandpass center
+// frequency.  The `lowpass(Nh)` cutoff `w1` is half the desired bandpass width.
+// `bandpass` is a standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : bandpass(Nh,fl,fu) : _
+// ```
+//
+// Where:
+//
+// * `Nh`: HALF the desired bandpass order (which is therefore even)
+// * `fl`: lower -3dB frequency in Hz
+// * `fu`: upper -3dB frequency in Hz
+// Thus, the passband width is `fu-fl`,
+//       and its center frequency is `(fl+fu)/2`.
+//
+// #### Reference
+// <http://cnx.org/content/m16913/latest/>
+//-------------------------------------------------------------------------
+declare bandpass author "Julius O. Smith III";
+declare bandpass copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare bandpass license "MIT-style STK-4.3 license";
+bandpass(Nh,fl,fu) = bandpass0_bandstop1(0,Nh,fl,fu);
+
+
+//--------------------`(fi.)bandstop`----------------
+// Order 2*Nh Butterworth bandstop filter made using the transformation
+// `s <- s + wc^2/s` on `highpass(Nh)`, where `wc` is the desired bandpass center
+// frequency.  The `highpass(Nh)` cutoff `w1` is half the desired bandpass width.
+// `bandstop` is a standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : bandstop(Nh,fl,fu) : _
+// ```
+// Where:
+//
+// * `Nh`: HALF the desired bandstop order (which is therefore even)
+// * `fl`: lower -3dB frequency in Hz
+// * `fu`: upper -3dB frequency in Hz
+// Thus, the passband (stopband) width is `fu-fl`,
+//       and its center frequency is `(fl+fu)/2`.
+//
+// #### Reference
+// <http://cnx.org/content/m16913/latest/>
+//-------------------------------------------------------------------------
+declare bandstop author "Julius O. Smith III";
+declare bandstop copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare bandstop license "MIT-style STK-4.3 license";
+bandstop(Nh,fl,fu) = bandpass0_bandstop1(1,Nh,fl,fu);
+
+declare bandpass0_bandstop1 author "Julius O. Smith III";
+declare bandpass0_bandstop1 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare bandpass0_bandstop1 license "MIT-style STK-4.3 license";
+bandpass0_bandstop1(s,Nh,fl,fu) = bpbsr(s,Nh,Nh,fl,fu)
+with {
+  wl = 2*ma.PI*fl; // digital (z-plane) lower passband edge
+  wu = 2*ma.PI*fu; // digital (z-plane) upper passband edge
+
+  c = 2.0*ma.SR; // bilinear transform scaling used in tf2sb, tf1sb
+  wla = c*tan(wl/c); // analog (s-plane) lower cutoff
+  wua = c*tan(wu/c); // analog (s-plane) upper cutoff
+
+  wc = sqrt(wla*wua); // s-plane center frequency
+  w1 = wua - wc^2/wua; // s-plane lowpass prototype cutoff
+
+  bpbsr(s,0,Nh,fl,fu) = _;
+  bpbsr(s,1,Nh,fl,fu) = tf1sb(s,1-s,1,w1,wc);
+  bpbsr(s,O,Nh,fl,fu) = bpbsr(s,O-2,Nh,fl,fu) : tf2sb(s,0,(1-s),a1s,1,w1,wc)
+  with {
+    parity = Nh % 2;
+    S = (O-parity)/2; // current section number
+    a1s = -2*cos(-1*ma.PI + (1-parity)*ma.PI/(2*Nh) + (S-1+parity)*ma.PI/Nh);
+  };
+};
+
+
+//===========================Elliptic Bandpass Filters====================================
+//========================================================================================
+
+//---------------------`(fi.)bandpass6e`-----------------------------
+// Order 12 elliptic bandpass filter analogous to `bandpass(6)`.
+//--------------------------------------------------------------
+declare bandpass6e author "Julius O. Smith III";
+declare bandpass6e copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare bandpass6e license "MIT-style STK-4.3 license";
+bandpass6e(fl,fu) = tf2sb(b21,b11,b01,a11,a01,w1,wc) : tf1sb(0,1,a02,w1,wc)
+with {
+  a11 = 0.802636764161030; // In octave: format long; poly(p(1:2))
+  a01 = 1.412270893774204;
+  a02 = 0.822445908998816; // poly(p(3))
+  b21 = 0.019809144837789; // poly(z)
+  b11 = 0;
+  b01 = 1.161516418982696;
+
+  wl = 2*ma.PI*fl; // digital (z-plane) lower passband edge
+  wu = 2*ma.PI*fu; // digital (z-plane) upper passband edge
+
+  c = 2.0*ma.SR; // bilinear transform scaling used in tf2sb, tf1sb
+  wla = c*tan(wl/c); // analog (s-plane) lower cutoff
+  wua = c*tan(wu/c); // analog (s-plane) upper cutoff
+
+  wc = sqrt(wla*wua); // s-plane center frequency
+  w1 = wua - wc^2/wua; // s-plane lowpass cutoff
+};
+
+//----------------------`(fi.)bandpass12e`---------------------------
+// Order 24 elliptic bandpass filter analogous to `bandpass(6)`.
+//--------------------------------------------------------------
+declare bandpass12e author "Julius O. Smith III";
+declare bandpass12e copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare bandpass12e license "MIT-style STK-4.3 license";
+bandpass12e(fl,fu) =
+              tf2sb(b21,b11,b01,a11,a01,w1,wc) :
+              tf2sb(b22,b12,b02,a12,a02,w1,wc) :
+              tf2sb(b23,b13,b03,a13,a03,w1,wc)
+with { // octave script output:
+  b21 = 0.000099999997055;
+  a21 = 1;
+  b11 = 0;
+  a11 = 0.782413046821645;
+  b01 = 0.000433227200555;
+  a01 = 0.245291508706160;
+  b22 = 1;
+  a22 = 1;
+  b12 = 0;
+  a12 = 0.512478641889141;
+  b02 = 7.621731298870603;
+  a02 = 0.689621364484675;
+  b23 = 1;
+  a23 = 1;
+  b13 = 0;
+  a13 = 0.168404871113589;
+  b03 = 53.536152954556727;
+  a03 = 1.069358407707312;
+
+  wl = 2*ma.PI*fl; // digital (z-plane) lower passband edge
+  wu = 2*ma.PI*fu; // digital (z-plane) upper passband edge
+
+  c = 2.0*ma.SR; // bilinear transform scaling used in tf2sb, tf1sb
+  wla = c*tan(wl/c); // analog (s-plane) lower cutoff
+  wua = c*tan(wu/c); // analog (s-plane) upper cutoff
+
+  wc = sqrt(wla*wua); // s-plane center frequency
+  w1 = wua - wc^2/wua; // s-plane lowpass cutoff
+};
+
+//------------------------`(fi.)pospass`---------------------------
+// Positive-Pass Filter (single-side-band filter).
+//
+// #### Usage
+//
+// ```
+// _ : pospass(N,fc) : _,_
+// ```
+//
+// where
+//
+// * `N`: filter order (Butterworth bandpass for positive frequencies).
+// * `fc`: lower bandpass cutoff frequency in Hz.
+//   - Highpass cutoff frequency at ma.SR/2 - fc Hz.
+//
+// #### Example test program
+//
+// * See `dm.pospass_demo`
+// * Look at frequency response
+//
+// #### Method
+//
+// A filter passing only positive frequencies can be made from a
+// half-band lowpass by modulating it up to the positive-frequency range.
+// Equivalently, down-modulate the input signal using a complex sinusoid at -SR/4 Hz,
+// lowpass it with a half-band filter, and modulate back up by SR/4 Hz.
+// In Faust/math notation:
+// $$pospass(N) = \ast(e^{-j\frac{\pi}{2}n}) : \mbox{lowpass(N,SR/4)} : \ast(e^{j\frac{\pi}{2}n})$$
+//
+// An approximation to the Hilbert transform is given by the
+// imaginary output signal:
+//
+// ```
+// hilbert(N) = pospass(N) : !,*(2);
+// ```
+//
+// #### References
+// * <https://ccrma.stanford.edu/~jos/mdft/Analytic_Signals_Hilbert_Transform.html>
+// * <https://ccrma.stanford.edu/~jos/sasp/Comparison_Optimal_Chebyshev_FIR_I.html>
+// * <https://ccrma.stanford.edu/~jos/sasp/Hilbert_Transform.html>
+//------------------------------------------------------------
+declare pospass author "Julius O. Smith III";
+declare pospass copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare pospass license "MIT-style STK-4.3 license";
+pospass(N,fc) = fi.pospass0(lpf) with {
+  fcs = ma.SR/4 - fc; // Upper lowpass cutoff = (SR/2 - fc) - SR/4
+  lpf = fi.lowpass(N,fcs); // Butterworth lowpass
+};
+
+declare pospass6e author "Julius O. Smith III";
+declare pospass6e copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare pospass6e license "MIT-style STK-4.3 license";
+pospass6e(fc) = fi.pospass0(lpf) with {
+  lpf = fi.lowpass6e(ma.SR/4 - fc); // Elliptic lowpass, order 6
+};
+
+declare pospass0 author "Julius O. Smith III";
+declare pospass0 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare pospass0 license "MIT-style STK-4.3 license";
+pospass0(lpf) = unmodulate : lpf, lpf : modulate with {
+  c = 1-1' : +~(*(-1):mem); // complex sinusoid rotating at SR/4
+  s = c'; // ||: 0, 1, 0, -1 :||
+  unmodulate = _ <: *(c),*(-s); // subtract SR/4 from all input frequencies
+  modulate(x,y) = c*x-s*y, c*y + s*x; // add SR/4 to all frequencies
+};
+
+
+//=================Parametric Equalizers (Shelf, Peaking)=================================
+// Parametric Equalizers (Shelf, Peaking).
+//
+// #### References
+// * <http://en.wikipedia.org/wiki/Equalization>
+// * <https://webaudio.github.io/Audio-EQ-Cookbook/Audio-EQ-Cookbook.txt>
+// * Digital Audio Signal Processing, Udo Zolzer, Wiley, 1999, p. 124
+// * <https://ccrma.stanford.edu/~jos/filters/Low_High_Shelving_Filters.html>
+// * <https://ccrma.stanford.edu/~jos/filters/Peaking_Equalizers.html>
+// * maxmsp.lib in the Faust distribution
+// * bandfilter.dsp in the faust2pd distribution
+//========================================================================================
+
+//----------------------`(fi.)low_shelf`----------------------
+// First-order "low shelf" filter (gain boost|cut between dc and some frequency)
+// `low_shelf` is a standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : lowshelf(N,L0,fx) : _
+// _ : low_shelf(L0,fx) : _ // default case (order 3)
+// _ : lowshelf_other_freq(N,L0,fx) : _
+// ```
+//
+// Where:
+// * `N`: filter order 1, 3, 5, ... (odd only, default should be 3, a constant numerical expression)
+// * `L0`: desired level (dB) between dc and fx (boost `L0>0` or cut `L0<0`)
+// * `fx`: -3dB frequency of lowpass band (`L0>0`) or upper band (`L0<0`)
+//       (see "SHELF SHAPE" below).
+//
+// The gain at SR/2 is constrained to be 1.
+// The generalization to arbitrary odd orders is based on the well known
+// fact that odd-order Butterworth band-splits are allpass-complementary
+// (see filterbank documentation below for references).
+//
+// #### Shelf Shape
+// The magnitude frequency response is approximately piecewise-linear
+// on a log-log plot ("BODE PLOT").  The Bode "stick diagram" approximation
+// L(lf) is easy to state in dB versus dB-frequency lf = dB(f):
+//
+// * L0 > 0:
+// 	* L(lf) = L0, f between 0 and fx = 1st corner frequency;
+// 	* L(lf) = L0 - N * (lf - lfx), f between fx and f2 = 2nd corner frequency;
+// 	* L(lf) = 0, lf > lf2.
+// 	* lf2 = lfx + L0/N = dB-frequency at which level gets back to 0 dB.
+// * L0 < 0:
+// 	* L(lf) = L0, f between 0 and f1 = 1st corner frequency;
+// 	* L(lf) = - N * (lfx - lf), f between f1 and lfx = 2nd corner frequency;
+// 	* L(lf) = 0, lf > lfx.
+// 	* lf1 = lfx + L0/N = dB-frequency at which level goes up from L0.
+//
+//  See `lowshelf_other_freq`.
+//
+// #### References
+// See "Parametric Equalizers" above for references regarding
+// `low_shelf`, `high_shelf`, and `peak_eq`.
+//
+//--------------------------------------------------------------
+declare lowshelf author "Julius O. Smith III";
+declare lowshelf copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare lowshelf license "MIT-style STK-4.3 license";
+lowshelf(N,L0,fx) = filterbank(N,(fx)) : _, *(ba.db2linear(L0)) :> _;
+
+// Special cases and optimization:
+declare low_shelf author "Julius O. Smith III";
+declare low_shelf copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare low_shelf license "MIT-style STK-4.3 license";
+low_shelf  = lowshelf(3); // default = 3rd order Butterworth
+
+declare low_shelf1 author "Julius O. Smith III";
+declare low_shelf1 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare low_shelf1 license "MIT-style STK-4.3 license";
+low_shelf1(L0,fx,x) = x + (ba.db2linear(L0)-1)*lowpass(1,fx,x); // optimized
+
+declare low_shelf1_l author "Julius O. Smith III";
+declare low_shelf1_l copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare low_shelf1_l license "MIT-style STK-4.3 license";
+low_shelf1_l(G0,fx,x) = x + (G0-1)*lowpass(1,fx,x); // optimized
+
+declare lowshelf_other_freq author "Julius O. Smith III";
+declare lowshelf_other_freq copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare lowshelf_other_freq license "MIT-style STK-4.3 license";
+lowshelf_other_freq(N, L0, fx) = ba.db2linear(ba.linear2db(fx) + L0/N); // convenience
+
+//-------------`(fi.)high_shelf`--------------
+// First-order "high shelf" filter (gain boost|cut above some frequency).
+// `high_shelf` is a standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : highshelf(N,Lpi,fx) : _
+// _ : high_shelf(L0,fx) : _ // default case (order 3)
+// _ : highshelf_other_freq(N,Lpi,fx) : _
+// ```
+//
+// Where:
+//
+// * `N`: filter order 1, 3, 5, ... (odd only, a constant numerical expression).
+// * `Lpi`: desired level (dB) between fx and SR/2 (boost Lpi>0 or cut Lpi<0)
+// * `fx`: -3dB frequency of highpass band (L0>0) or lower band (L0<0)
+//        (Use highshelf_other_freq() below to find the other one.)
+//
+// The gain at dc is constrained to be 1.
+// See `lowshelf` documentation above for more details on shelf shape.
+//
+// #### References
+// See "Parametric Equalizers" above for references regarding
+// `low_shelf`, `high_shelf`, and `peak_eq`.
+//
+//--------------------------------------------------------------
+declare highshelf author "Julius O. Smith III";
+declare highshelf copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare highshelf license "MIT-style STK-4.3 license";
+highshelf(N,Lpi,fx) = filterbank(N,(fx)) : *(ba.db2linear(Lpi)), _ :> _;
+// Special cases and optimization:
+high_shelf = highshelf(3); // default = 3rd order Butterworth
+
+declare high_shelf1 author "Julius O. Smith III";
+declare high_shelf1 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare high_shelf1 license "MIT-style STK-4.3 license";
+high_shelf1(Lpi,fx,x) = x + (ba.db2linear(Lpi)-1)*highpass(1,fx,x); // optimized
+
+declare high_shelf1_l author "Julius O. Smith III";
+declare high_shelf1_l copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare high_shelf1_l license "MIT-style STK-4.3 license";
+high_shelf1_l(Gpi,fx,x) = x + (Gpi-1)*highpass(1,fx,x); //optimized
+
+// shelf transitions between frequency fx and this one:
+declare highshelf_other_freq author "Julius O. Smith III";
+declare highshelf_other_freq copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare highshelf_other_freq license "MIT-style STK-4.3 license";
+highshelf_other_freq(N, Lpi, fx) = ba.db2linear(ba.linear2db(fx) - Lpi/N);
+
+
+//-------------------`(fi.)peak_eq`------------------------------
+// Second order "peaking equalizer" section (gain boost or cut near some frequency)
+// Also called a "parametric equalizer" section.
+// `peak_eq` is a standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : peak_eq(Lfx,fx,B) : _
+// ```
+//
+// Where:
+//
+// * `Lfx`: level (dB) at fx (boost Lfx>0 or cut Lfx<0)
+// * `fx`: peak frequency (Hz)
+// * `B`: bandwidth (B) of peak in Hz
+//
+// #### References
+// See "Parametric Equalizers" above for references regarding
+// `low_shelf`, `high_shelf`, and `peak_eq`.
+//
+//--------------------------------------------------------------
+declare peak_eq author "Julius O. Smith III";
+declare peak_eq copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare peak_eq license "MIT-style STK-4.3 license";
+peak_eq(Lfx,fx,B) = tf2s(1,b1s,1,a1s,1,wx) with {
+  T = float(1.0/ma.SR);
+  Bw = B*T/sin(wx*T); // prewarp s-bandwidth for more accuracy in z-plane
+  a1 = ma.PI*Bw;
+  b1 = g*a1;
+  g = ba.db2linear(abs(Lfx));
+  b1s = select2(Lfx>0,a1,b1); // When Lfx>0, pole dominates bandwidth
+  a1s = select2(Lfx>0,b1,a1); // When Lfx<0, zero dominates
+  wx = 2*ma.PI*fx;
+};
+
+//--------------------`(fi.)peak_eq_cq`----------------------------
+// Constant-Q second order peaking equalizer section.
+//
+// #### Usage
+//
+// ```
+// _ : peak_eq_cq(Lfx,fx,Q) : _
+// ```
+//
+// Where:
+//
+// * `Lfx`: level (dB) at fx
+// * `fx`: boost or cut frequency (Hz)
+// * `Q`: "Quality factor" = fx/B where B = bandwidth of peak in Hz
+//
+// #### References
+// See "Parametric Equalizers" above for references regarding
+// `low_shelf`, `high_shelf`, and `peak_eq`.
+//
+//------------------------------------------------------------
+declare peak_eq_cq author "Julius O. Smith III";
+declare peak_eq_cq copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare peak_eq_cq license "MIT-style STK-4.3 license";
+peak_eq_cq(Lfx,fx,Q) = peak_eq(Lfx,fx,fx/Q);
+
+//-------------------`(fi.)peak_eq_rm`--------------------------
+// Regalia-Mitra second order peaking equalizer section.
+//
+// #### Usage
+//
+// ```
+// _ : peak_eq_rm(Lfx,fx,tanPiBT) : _
+// ```
+//
+// Where:
+//
+// * `Lfx`: level (dB) at fx
+// * `fx`: boost or cut frequency (Hz)
+// * `tanPiBT`: `tan(PI*B/SR)`, where B = -3dB bandwidth (Hz) when 10^(Lfx/20) = 0
+//         ~ PI*B/SR for narrow bandwidths B
+//
+// #### Reference
+// P.A. Regalia, S.K. Mitra, and P.P. Vaidyanathan,
+// "The Digital All-Pass Filter: A Versatile Signal Processing Building Block"
+// Proceedings of the IEEE, 76(1):19-37, Jan. 1988.  (See pp. 29-30.)
+// See also "Parametric Equalizers" above for references on shelf
+// and peaking equalizers in general.
+//
+//------------------------------------------------------------
+declare peak_eq_rm author "Julius O. Smith III";
+declare peak_eq_rm copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare peak_eq_rm license "MIT-style STK-4.3 license";
+peak_eq_rm(Lfx,fx,tanPiBT) = _ <: _,A,_ : +,- : *(0.5),*(K/2.0) : + with {
+  A = tf2(k2, k1*(1+k2), 1, k1*(1+k2), k2) <: _,_; // allpass
+  k1 = 0.0 - cos(2.0*ma.PI*fx/ma.SR);
+  k2 = (1.0 - tanPiBT)/(1.0 + tanPiBT);
+  K = ba.db2linear(Lfx);
+};
+
+
+//---------------------`(fi.)spectral_tilt`-------------------------
+// Spectral tilt filter, providing an arbitrary spectral rolloff factor
+// alpha in (-1,1), where
+//  -1 corresponds to one pole (-6 dB per octave), and
+//  +1 corresponds to one zero (+6 dB per octave).
+// In other words, alpha is the slope of the ln magnitude versus ln frequency.
+// For a "pinking filter" (e.g., to generate 1/f noise from white noise),
+// set alpha to -1/2.
+//
+// #### Usage
+//
+// ```
+// _ : spectral_tilt(N,f0,bw,alpha) : _
+// ```
+// Where:
+//
+// * `N`: desired integer filter order (fixed at compile time)
+// * `f0`: lower frequency limit for desired roll-off band > 0
+// * `bw`: bandwidth of desired roll-off band
+// * `alpha`: slope of roll-off desired in nepers per neper,
+//         between -1 and 1 (ln mag / ln radian freq)
+//
+// #### Example test program
+//
+// See `dm.spectral_tilt_demo` and the documentation for `no.pink_noise`.
+//
+// #### Reference
+// J.O. Smith and H.F. Smith,
+// "Closed Form Fractional Integration and Differentiation via Real Exponentially Spaced Pole-Zero Pairs",
+// arXiv.org publication arXiv:1606.06154 [cs.CE], June 7, 2016,
+// <http://arxiv.org/abs/1606.06154>
+//
+//------------------------------------------------------------
+declare spectral_tilt author "Julius O. Smith III";
+declare spectral_tilt copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare spectral_tilt license "MIT-style STK-4.3 license";
+spectral_tilt(N,f0,bw,alpha) = seq(i,N,sec(i)) with {
+  sec(i) = g * tf1s(b1,b0,a0,1) with {
+    g = a0/b0; // unity dc-gain scaling
+    b1 = 1.0;
+    b0 = mzh(i);
+    a0 = mph(i);
+    mzh(i) = prewarp(mz(i),ma.SR,w0); // prewarping for bilinear transform
+    mph(i) = prewarp(mp(i),ma.SR,w0);
+    prewarp(w,SR,wp) = wp * tan(w*T/2) / tan(wp*T/2) with { T = 1/ma.SR; };
+    mz(i) = w0 * r ^ (-alpha+i); // minus zero i in s plane
+    mp(i) = w0 * r ^ i; // minus pole i in s plane
+    f0p = max(f0,ma.EPSILON); // cannot go to zero
+    w0 = 2 * ma.PI * f0p; // radian frequency of first pole
+    f1 = f0p + bw; // upper band limit
+    r = (f1/f0p)^(1.0/float(N-1)); // pole ratio (2 => octave spacing)
+  };
+};
+
+
+//----------------------`(fi.)levelfilter`----------------------
+// Dynamic level lowpass filter.
+// `levelfilter` is a standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : levelfilter(L,freq) : _
+// ```
+//
+// Where:
+//
+// * `L`: desired level (in dB) at Nyquist limit (SR/2), e.g., -60
+// * `freq`: corner frequency (-3dB point) usually set to fundamental freq
+// * `N`: Number of filters in series where L = L/N
+//
+// #### Reference
+// <https://ccrma.stanford.edu/realsimple/faust_strings/Dynamic_Level_Lowpass_Filter.html>
+//------------------------------------------------------------
+declare levelfilter author "Julius O. Smith III";
+declare levelfilter copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare levelfilter license "MIT-style STK-4.3 license";
+levelfilter(L,freq,x) = (L * L0 * x) + ((1.0-L) * lp2out(x))
+with {
+  L0 = pow(L,1/3);
+  Lw = ma.PI*freq/ma.SR; // = w1 T / 2
+  Lgain = Lw / (1.0 + Lw);
+  Lpole2 = (1.0 - Lw) / (1.0 + Lw);
+  lp2out = *(Lgain) : + ~ *(Lpole2);
+};
+
+
+//----------------------`(fi.)levelfilterN`----------------------
+// Dynamic level lowpass filter.
+//
+// #### Usage
+//
+// ```
+// _ : levelfilterN(N,freq,L) : _
+// ```
+//
+// Where:
+//
+// * `N`: Number of filters in series where L = L/N, a constant numerical expression
+// * `freq`: corner frequency (-3dB point) usually set to fundamental freq
+// * `L`: desired level (in dB) at Nyquist limit (SR/2), e.g., -60
+//
+// #### Reference
+// <https://ccrma.stanford.edu/realsimple/faust_strings/Dynamic_Level_Lowpass_Filter.html>
+//------------------------------------------------------------
+declare levelfilterN author "Julius O. Smith III";
+declare levelfilterN copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare levelfilterN license "MIT-style STK-4.3 license";
+levelfilterN(N,freq,L) = seq(i,N,levelfilter((L/N),freq));
+
+
+//=================================Mth-Octave Filter-Banks================================
+// Mth-octave filter-banks split the input signal into a bank of parallel signals, one
+// for each spectral band. They are related to the Mth-Octave Spectrum-Analyzers in
+// `analysis.lib`.
+// The documentation of this library contains more details about the implementation.
+// The parameters are:
+//
+// * `M`: number of band-slices per octave (>1), a constant numerical expression
+// * `N`: total number of bands (>2), a constant numerical expression
+// * `ftop`: upper bandlimit of the Mth-octave bands (<SR/2)
+//
+// In addition to the Mth-octave output signals, there is a highpass signal
+// containing frequencies from ftop to SR/2, and a "dc band" lowpass signal
+// containing frequencies from 0 (dc) up to the start of the Mth-octave bands.
+// Thus, the N output signals are
+//
+// ```
+// highpass(ftop), MthOctaveBands(M,N-2,ftop), dcBand(ftop*2^(-M*(N-1)))
+// ```
+//
+// A Filter-Bank is defined here as a signal bandsplitter having the
+// property that summing its output signals gives an allpass-filtered
+// version of the filter-bank input signal.  A more conventional term for
+// this is an "allpass-complementary filter bank".  If the allpass filter
+// is a pure delay (and possible scaling), the filter bank is said to be
+// a "perfect-reconstruction filter bank" (see Vaidyanathan-1993 cited
+// below for details).  A "graphic equalizer", in which band signals
+// are scaled by gains and summed, should be based on a filter bank.
+//
+// The filter-banks below are implemented as Butterworth or Elliptic
+// spectrum-analyzers followed by delay equalizers that make them
+// allpass-complementary.
+//
+// #### Increasing Channel Isolation
+//
+// Go to higher filter orders - see Regalia et al. or Vaidyanathan (cited
+// below) regarding the construction of more aggressive recursive
+// filter-banks using elliptic or Chebyshev prototype filters.
+//
+// #### References
+// * "Tree-structured complementary filter banks using all-pass sections",
+//   Regalia et al., IEEE Trans. Circuits & Systems, CAS-34:1470-1484, Dec. 1987
+// * "Multirate Systems and Filter Banks", P. Vaidyanathan, Prentice-Hall, 1993
+// * Elementary filter theory: <https://ccrma.stanford.edu/~jos/filters/>
+//========================================================================================
+
+//------------------------`(fi.)mth_octave_filterbank[n]`-------------------------
+// Allpass-complementary filter banks based on Butterworth band-splitting.
+// For Butterworth band-splits, the needed delay equalizer is easily found.
+//
+// #### Usage
+//
+// ```
+// _ : mth_octave_filterbank(O,M,ftop,N) : par(i,N,_)     // Oth-order
+// _ : mth_octave_filterbank_alt(O,M,ftop,N) : par(i,N,_) // dc-inverted version
+// ```
+//
+// Also for convenience:
+//
+// ```
+// _ : mth_octave_filterbank3(M,ftop,N) : par(i,N,_) // 3rd-order Butterworth
+// _ : mth_octave_filterbank5(M,ftop,N) : par(i,N,_) // 5th-order Butterworth
+// mth_octave_filterbank_default = mth_octave_filterbank5;
+// ```
+//
+// Where:
+//
+// * `O`: order of filter used to split each frequency band into two, a constant numerical expression
+// * `M`: number of band-slices per octave, a constant numerical expression
+// * `ftop`: highest band-split crossover frequency (e.g., 20 kHz)
+// * `N`: total number of bands (including dc and Nyquist), a constant numerical expression
+//------------------------------------------------------------
+declare mth_octave_filterbank author "Julius O. Smith III";
+declare mth_octave_filterbank copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare mth_octave_filterbank license "MIT-style STK-4.3 license";
+mth_octave_filterbank(O,M,ftop,N) = an.mth_octave_analyzer(O,M,ftop,N) : delayeq(N) 
+with {
+   fc(n) = ftop * 2^(float(n-N+1)/float(M)); // -3dB crossover frequencies
+   ap(n) = highpass_plus_lowpass(O,fc(n));   // delay-equalizing allpass
+   delayeq(N) = par(i,N-2,apchain(i+1)), _, _;
+   apchain(i) = seq(j,N-1-i,ap(j+1));
+};
+
+// dc-inverted version. This reduces the delay-equalizer order for odd O.
+// Negating the input signal makes the dc band noninverting
+// and all higher bands sign-inverted (if preferred).
+declare mth_octave_filterbank_alt author "Julius O. Smith III";
+declare mth_octave_filterbank_alt copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare mth_octave_filterbank_alt license "MIT-style STK-4.3 license";
+mth_octave_filterbank_alt(O,M,ftop,N) = an.mth_octave_analyzer(O,M,ftop,N) : delayeqi(O,N) 
+with {
+    fc(n) = ftop * 2^(float(n-N+1)/float(M)); // -3dB crossover frequencies
+    ap(n) = highpass_minus_lowpass(O,fc(n)); // half the order of 'plus' case
+    delayeqi(N) = par(i,N-2,apchain(i+1)), _, *(-1.0);
+    apchain(i) = seq(j,N-1-i,ap(j+1));
+};
+
+// Note that even-order cases require complex coefficients.
+// See Vaidyanathan 1993 and papers cited there for more info.
+declare mth_octave_filterbank3 author "Julius O. Smith III";
+declare mth_octave_filterbank3 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare mth_octave_filterbank3 license "MIT-style STK-4.3 license";
+mth_octave_filterbank3(M,ftop,N) = mth_octave_filterbank_alt(3,M,ftop,N);
+
+declare mth_octave_filterbank5 author "Julius O. Smith III";
+declare mth_octave_filterbank5 copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare mth_octave_filterbank5 license "MIT-style STK-4.3 license";
+mth_octave_filterbank5(M,ftop,N) = mth_octave_filterbank(5,M,ftop,N);
+
+declare mth_octave_filterbank_default author "Julius O. Smith III";
+declare mth_octave_filterbank_default copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare mth_octave_filterbank_default license "MIT-style STK-4.3 license";
+mth_octave_filterbank_default = mth_octave_filterbank5;
+
+
+//===============Arbitrary-Crossover Filter-Banks and Spectrum Analyzers==================
+// These are similar to the Mth-octave analyzers above, except that the
+// band-split frequencies are passed explicitly as arguments.
+//========================================================================================
+
+// ACKNOWLEDGMENT
+// Technique for processing a variable number of signal arguments due
+// to Yann Orlarey (as is the entire Faust framework!)
+
+//---------------`(fi.)filterbank`--------------------------
+// Filter bank.
+// `filterbank` is a standard Faust function.
+//
+// #### Usage
+//
+// ```
+// _ : filterbank (O,freqs) : par(i,N,_) // Butterworth band-splits
+// ```
+// Where:
+//
+// * `O`: band-split filter order (odd integer required for filterbank[i], a constant numerical expression)
+// * `freqs`: (fc1,fc2,...,fcNs) [in numerically ascending order], where
+//           Ns=N-1 is the number of octave band-splits
+//           (total number of bands N=Ns+1).
+//
+// If frequencies are listed explicitly as arguments, enclose them in parens:
+//
+// ```
+// _ : filterbank(3,(fc1,fc2)) : _,_,_
+// ```
+//---------------------------------------------------
+declare filterbank author "Julius O. Smith III";
+declare filterbank copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare filterbank license "MIT-style STK-4.3 license";
+filterbank(O,lfreqs) = an.analyzer(O,lfreqs) : delayeq(nb) 
+with {
+   nb = ba.count(lfreqs);
+   fc(n) = ba.take(n, lfreqs);
+   ap(n) = highpass_plus_lowpass(O,fc(n));
+   delayeq(1) = _,_; // par(i,0,...) does not fly
+   delayeq(nb) = par(i,nb-1,apchain(nb-1-i)),_,_;
+   apchain(0) = _;
+   apchain(i) = ap(i) : apchain(i-1);
+};
+
+//-----------------`(fi.)filterbanki`----------------------
+// Inverted-dc filter bank.
+//
+// #### Usage
+//
+// ```
+// _ : filterbanki(O,freqs) : par(i,N,_) // Inverted-dc version
+// ```
+//
+// Where:
+//
+// * `O`: band-split filter order (odd integer required for `filterbank[i]`, a constant numerical expression)
+// * `freqs`: (fc1,fc2,...,fcNs) [in numerically ascending order], where
+//           Ns=N-1 is the number of octave band-splits
+//           (total number of bands N=Ns+1).
+//
+// If frequencies are listed explicitly as arguments, enclose them in parens:
+//
+// ```
+// _ : filterbanki(3,(fc1,fc2)) : _,_,_
+// ```
+//---------------------------------------------------
+declare filterbanki author "Julius O. Smith III";
+declare filterbanki copyright "Copyright (C) 2003-2019 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare filterbanki license "MIT-style STK-4.3 license";
+filterbanki(O,lfreqs) = _ <: bsplit(nb) 
+with {
+   nb = ba.count(lfreqs);
+   fc(n) = ba.take(n, lfreqs);
+   lp(n) = lowpass(O,fc(n));
+   hp(n) = highpass(O,fc(n));
+   ap(n) = highpass_minus_lowpass(O,fc(n));
+   bsplit(0) = *(-1.0);
+   bsplit(i) = (hp(i) : delayeq(i-1)), (lp(i) <: bsplit(i-1));
+   delayeq(0) = _; // moving the *(-1) here inverts all outputs BUT dc
+   delayeq(i) = ap(i) : delayeq(i-1);
+};
+
+
+//===============State Variable Filters=========================================================
+// #### References 
+// Solving the continuous SVF equations using trapezoidal integration
+//
+// <https://cytomic.com/files/dsp/SvfLinearTrapOptimised2.pdf>
+//========================================================================================
+
+//-----------------`(fi.)svf`----------------------
+// An environment with `lp`, `bp`, `hp`, `notch`, `peak`, `ap`, `bell`, `ls`, `hs` SVF based filters. 
+// All filters have `freq` and `Q` parameters, the `bell`, `ls`, `hs` ones also have a `gain` third parameter.
+//
+// #### Usage
+//
+// ```
+// _ : svf.xx(freq, Q, [gain]) : _
+// ```
+//
+// Where:
+//
+// * `freq`: cut frequency
+// * `Q`: quality factor
+// * `[gain]`: gain in dB
+//
+/// ```
+//---------------------------------------------------
+declare svf author "Oleg Nesterov";
+declare svf copyright "Copyright (C) 2020 Oleg Nesterov <oleg@redhat.com>";
+declare svf license "MIT-style STK-4.3 license";
+
+svf = environment {
+
+	// Internal implementation
+	svf(T,F,Q,G) = tick ~ (_,_) : !,!,si.dot(3, mix)
+	with {
+		tick(ic1eq, ic2eq, v0) =
+			2*v1 - ic1eq,
+			2*v2 - ic2eq,
+			v0, v1, v2
+		with {
+			v1 = ic1eq + g *(v0-ic2eq) : /(1 + g*(g+k));
+			v2 = ic2eq + g * v1;
+		};
+
+		A = pow(10.0, G/40.0);
+
+		g = tan(F * ma.PI/ma.SR) : case {
+			(7) => /(sqrt(A));
+			(8) => *(sqrt(A));
+			(t) => _;
+		} (T);
+
+		k = case {
+			(6) => 1/(Q*A);
+			(t) => 1/Q;
+		} (T);
+
+		mix = case {
+			(0) => 0, 0, 1;
+			(1) => 0, 1, 0;
+			(2) => 1, -k, -1;
+			(3) => 1, -k, 0;
+			(4) => 1, -k, -2;
+			(5) => 1, -2*k, 0;
+			(6) => 1, k*(A*A-1), 0;
+			(7) => 1, k*(A-1), A*A-1;
+			(8) => A*A, k*(1-A)*A, 1-A*A;
+		} (T);
+	};
+
+	// External API 
+	lp(f,q)     = svf(0, f, q, 0);
+	bp(f,q)     = svf(1, f, q, 0);
+	hp(f,q)     = svf(2, f, q, 0);
+	notch(f,q)  = svf(3, f, q, 0);
+	peak(f,q)   = svf(4, f, q, 0);
+	ap(f,q)     = svf(5, f, q, 0);
+	bell(f,q,g) = svf(6, f, q, g);
+	ls(f,q,g)   = svf(7, f, q, g);
+	hs(f,q,g)   = svf(8, f, q, g);
+};
+
+
+//===========Linkwitz-Riley 4th-order 2-way, 3-way, and 4-way crossovers=====
+//
+// The Linkwitz-Riley (LR) crossovers are designed to produce a fully-flat
+// magnitude response when their outputs are combined. The 4th-order
+// LR filters (LR4) have a 24dB/octave slope and they are rather popular audio
+// crossovers used in multi-band processing.
+//
+// The LR4 can be constructed by cascading two second-order Butterworth
+// filters. For the second-order Butterworth filters, we will use the SVF
+// filter implemented above by setting the Q-factor to 1.0 / sqrt(2.0).
+// These will be cascaded in pairs to build the LR4 highpass and lowpass.
+// For the phase correction, we will use the 2nd-order Butterworth allpass.
+//
+// #### Reference
+// Zavalishin, Vadim. "The art of VA filter design." Native Instruments, Berlin, Germany (2012). 
+//=============================================================================
+
+
+//----------`(fi.)lowpassLR4`---------------------------------------------------
+// 4th-order Linkwitz-Riley lowpass.
+//
+// #### Usage
+//
+// ```
+// _ : lowpassLR4(cf) : _
+// ```
+//
+// Where:
+//
+// * `cf` is the lowpass cutoff in Hz
+//------------------------------------------------------------------------------
+declare lowpassLR4 author "Dario Sanfilippo";
+declare lowpassLR4 copyright
+    "Copyright (C) 2022 Dario Sanfilippo <sanfilippo.dario@gmail.com>";
+declare lowpassLR4 license "MIT-style STK-4.3 license";
+lowpassLR4(cf, x) = x : seq(i, 2, svf.lp(cf, 1.0 / sqrt(2.0)));
+
+
+//----------`(fi.)highpassLR4`--------------------------------------------------
+// 4th-order Linkwitz-Riley highpass.
+//
+// #### Usage
+//
+// ```
+// _ : highpassLR4(cf) : _
+// ```
+//
+// Where:
+//
+// * `cf` is the highpass cutoff in Hz
+//------------------------------------------------------------------------------
+declare highpassLR4 author "Dario Sanfilippo";
+declare highpassLR4 copyright
+    "Copyright (C) 2022 Dario Sanfilippo <sanfilippo.dario@gmail.com>";
+declare highpassLR4 license "MIT-style STK-4.3 license";
+highpassLR4(cf, x) = x : seq(i, 2, svf.hp(cf, 1.0 / sqrt(2.0)));
+
+
+//----------`(fi.)crossover2LR4`------------------------------------------------
+// Two-way 4th-order Linkwitz-Riley crossover.
+//
+// #### Usage
+//
+// ```
+// _ : crossover2LR4(cf) : si.bus(2)
+// ```
+//
+// Where:
+//
+// * `cf` is the crossover split cutoff in Hz
+//------------------------------------------------------------------------------
+declare crossover2LR4 author "Dario Sanfilippo";
+declare crossover2LR4 copyright
+    "Copyright (C) 2022 Dario Sanfilippo <sanfilippo.dario@gmail.com>";
+declare crossover2LR4 license "MIT-style STK-4.3 license";
+crossover2LR4(cf, x) = lowpassLR4(cf, x) , highpassLR4(cf, x);
+
+
+//----------`(fi.)crossover3LR4`------------------------------------------------
+// Three-way 4th-order Linkwitz-Riley crossover.
+//
+// #### Usage
+//
+// ```
+// _ : crossover3LR4(cf1, cf2) : si.bus(3)
+// ```
+//
+// Where:
+//
+// * `cf1` is the crossover lower split cutoff in Hz
+// * `cf2` is the crossover upper split cutoff in Hz
+//------------------------------------------------------------------------------
+declare crossover3LR4 author "Dario Sanfilippo";
+declare crossover3LR4 copyright
+    "Copyright (C) 2022 Dario Sanfilippo <sanfilippo.dario@gmail.com>";
+declare crossover3LR4 license "MIT-style STK-4.3 license";
+crossover3LR4(cf1, cf2, x) =
+    crossover2LR4(cf1, x) : svf.ap(cf2, 1.0 / sqrt(2.0)) , crossover2LR4(cf2);
+
+
+//----------`(fi.)crossover4LR4`------------------------------------------------
+// Four-way 4th-order Linkwitz-Riley crossover.
+//
+// #### Usage
+//
+// ```
+// _ : crossover4LR4(cf1, cf2, cf3) : si.bus(4)
+// ```
+//
+// Where:
+//
+// * `cf1` is the crossover lower split cutoff in Hz
+// * `cf2` is the crossover mid split cutoff in Hz
+// * `cf3` is the crossover upper split cutoff in Hz
+//------------------------------------------------------------------------------
+declare crossover4LR4 author "Dario Sanfilippo";
+declare crossover4LR4 copyright
+    "Copyright (C) 2022 Dario Sanfilippo <sanfilippo.dario@gmail.com>";
+declare crossover4LR4 license "MIT-style STK-4.3 license";
+crossover4LR4(cf1, cf2, cf3, x) =
+    crossover2LR4(cf2, x) :
+        svf.ap(cf3, 1.0 / sqrt(2.0)) ,
+        svf.ap(cf1, 1.0 / sqrt(2.0)) :
+            crossover2LR4(cf1) ,
+            crossover2LR4(cf3);
+
+
+//----------`(fi.)crossover8LR4`------------------------------------------------
+// Eight-way 4th-order Linkwitz-Riley crossover.
+//
+// #### Usage
+//
+// ```
+// _ : crossover8LR4(cf1, cf2, cf3, cf4, cf5, cf6, cf7) : si.bus(8)
+// ```
+//
+// Where:
+//
+// * `cf1-cf7` are the crossover cutoff frequencies in Hz
+//------------------------------------------------------------------------------
+declare crossover8LR4 author "Dario Sanfilippo";
+declare crossover8LR4 copyright
+    "Copyright (C) 2022 Dario Sanfilippo <sanfilippo.dario@gmail.com>";
+declare crossover8LR4 license "MIT-style STK-4.3 license";
+crossover8LR4(cf1, cf2, cf3, cf4, cf5, cf6, cf7, x) =
+    crossover2LR4(cf4, x) :
+        (ap(cf6) : ap(cf5) : ap(cf7)) ,
+        (ap(cf2) : ap(cf1) : ap(cf3)) :
+            crossover2LR4(cf2) ,
+            crossover2LR4(cf6) :
+                ap(cf3) ,
+                ap(cf1) ,
+                ap(cf7) ,
+                ap(cf5) :
+                    crossover2LR4(cf1) ,
+                    crossover2LR4(cf3) ,
+                    crossover2LR4(cf5) ,
+                    crossover2LR4(cf7)
+    with {
+        ap(cf) = svf.ap(cf, 1.0 / sqrt(2.0));
+    };
+
+
+//=========================== Standardized Filters ============================
+//=============================================================================
+//
+// This section provides filters that are defined by national or
+// international standards, e.g. for measurement applications.
+
+//----------------------`(fi.)itu_r_bs_1770_4_kfilter`-------------------------
+// The prefilter from Recommendation ITU-R BS.1770-4 for loudness
+// measurement. Also known as "K-filter". The recommendation defines
+// biquad filter coefficients for a fixed sample rate of 48kHz (page
+// 4-5). Here, we construct biquads for arbitrary samplerates.  The
+// resulting filter is normalized, such that the magnitude at 997Hz is
+// unity gain 1.0.
+//
+// Please note, the ITU-recommendation handles the normalization in
+// equation (2) by subtracting 0.691dB, which is not needed with
+// `itu_r_bs_1770_4_kfilter`.
+//
+// One option for future improvement might be, to round those filter
+// coefficients, that are almost equal to one. Second, the maximum
+// magnitude difference at 48kHz between the ITU-defined filter and
+// `itu_r_bs_1770_4_kfilter` is 0.001dB, which obviously could be
+// less.
+//
+// #### Usage
+//
+// ```
+// _ : itu_r_bs_1770_4_kfilter : _
+// ```
+//
+// #### Reference
+// <https://www.itu.int/rec/R-REC-BS.1770>
+// <https://gist.github.com/jkbd/07521a98f7873a2dc3dbe16417930791>
+//-----------------------------------------------------------------------------
+declare itu_r_bs_1770_4_kfilter author "Jakob Dübel";
+declare itu_r_bs_1770_4_kfilter copyright "Copyright (C) 2022 Jakob Dübel";
+declare itu_r_bs_1770_4_kfilter license "ISC license";
+
+itu_r_bs_1770_4_kfilter = stage1 : stage2 : normalize997Hz
+with {
+  freq2k(f_c) = tan((ma.PI * f_c)/ma.SR);
+
+  stage1 = tf22t(b0,b1,b2,a1,a2)
+  with {
+    f_c = 1681.7632251028442; // Hertz
+    gain = 3.9997778685513232; // Decibel
+    K = freq2k(f_c);
+    V_0 = pow(10, (gain/20.0));
+
+    denominator = 1.0 + sqrt(2.0)*K + K^2;
+    b0 = (V_0 + sqrt((2.0*V_0))*K + K^2) / denominator;
+    b1 = 2.0*(K^2 - V_0) / denominator;
+    b2 = (V_0 - sqrt(2.0*V_0)*K + K^2) / denominator;
+
+    a1 = 2*(K^2 - 1) / denominator;
+    a2 = (1 - sqrt(2.0)*K + K^2) / denominator;
+  };
+
+  stage2 = tf22t(b0,b1,b2,a1,a2)
+  with {
+    f_c = 38.135470876002174; // Hertz
+    Q = 0.5003270373223665;
+    K = freq2k(f_c);
+
+    denominator = (K^2) * Q + K + Q;
+    b0 = Q / denominator;
+    b1 = -2*Q / denominator;
+    b2 = b0;
+
+    a1 = (2*Q * (K^2 - 1)) / denominator;
+    a2 = ((K^2) * Q - K + Q) / denominator;
+  };
+
+  normalize997Hz = *(0.9273671710547968);
+};
+
+
+//============================Averaging Functions==============================
+//=============================================================================
+//
+// These are a set of samplerate independent averaging functions based on
+// moving-average and one-pole filters with specific response characteristics.
+
+//----------------------------`(fi.)avg_rect`----------------------------------
+// Moving average.
+//
+// #### Usage
+//
+// ```
+// _ : avg_rect(period) : _
+// ```
+//
+// Where:
+//
+// * `period` is the averaging frame in seconds
+//-----------------------------------------------------------------------------
+declare avg_rect author "Dario Sanfilippo and Julius O. Smith III";
+declare avg_rect copyright "Copyright (C) 2020 Dario Sanfilippo 
+      <sanfilippo.dario@gmail.com> and 
+       2003-2020 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare avg_rect license "MIT-style STK-4.3 license";
+avg_rect(period, x) = x : ba.slidingMean(rint(period * ma.SR));
+
+
+//----------------------------`(fi.)avg_tau`-------------------------------------
+// Averaging function based on a one-pole filter and the tau response time.
+// Tau represents the effective length of the one-pole impulse response,
+// that is, tau is the integral of the filter's impulse response. This
+// response is slower to reach the final value but has less ripples in
+// non-steady signals.
+// 
+// #### Usage
+//
+// ```
+// _ : avg_tau(period) : _
+// ```
+//
+// Where:
+//
+// * `period` is the time, in seconds, for the system to decay by 1/e, 
+// or to reach 1-1/e of its final value.
+//
+// #### Reference
+//
+// <https://ccrma.stanford.edu/~jos/mdft/Exponentials.html>
+//-----------------------------------------------------------------------------
+declare avg_tau author "Dario Sanfilippo and Julius O. Smith III";
+declare avg_tau copyright "Copyright (C) 2020 Dario Sanfilippo 
+      <sanfilippo.dario@gmail.com> and 
+       2003-2020 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare avg_tau license "MIT-style STK-4.3 license";
+avg_tau(period, x) = fi.lptau(period, x);
+
+
+//----------------------------`(fi.)avg_t60`-------------------------------------
+// Averaging function based on a one-pole filter and the t60 response time.
+// This response is particularly useful when the system is required to 
+// reach the final value after about `period` seconds.
+// 
+// #### Usage
+//
+// ```
+// _ : avg_t60(period) : _
+// ```
+//
+// Where:
+//
+// * `period` is the time, in seconds, for the system to decay by 1/1000, 
+// or to reach 1-1/1000 of its final value.
+//
+// #### Reference
+//
+// <https://ccrma.stanford.edu/~jos/mdft/Audio_Decay_Time_T60.html>
+//-----------------------------------------------------------------------------
+declare avg_t60 author "Dario Sanfilippo and Julius O. Smith III";
+declare avg_t60 copyright "Copyright (C) 2020 Dario Sanfilippo 
+      <sanfilippo.dario@gmail.com> and 
+       2003-2020 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare avg_t60 license "MIT-style STK-4.3 license";
+avg_t60(period, x) = fi.lpt60(period, x);
+
+
+//----------------------------`(fi.)avg_t19`-------------------------------------
+// Averaging function based on a one-pole filter and the t19 response time.
+// This response is close to the moving-average algorithm as it roughly reaches
+// the final value after `period` seconds and shows about the same
+// oscillations for non-steady signals.
+//
+// #### Usage
+//
+// ```
+// _ : avg_t19(period) : _
+// ```
+//
+// Where:
+//
+// * `period` is the time, in seconds, for the system to decay by 1/e^2.2, 
+// or to reach 1-1/e^2.2 of its final value.
+//
+// #### Reference
+// Zölzer, U. (2008). Digital audio signal processing (Vol. 9). New York: Wiley.
+//-----------------------------------------------------------------------------
+declare avg_t19 author "Dario Sanfilippo and Julius O. Smith III";
+declare avg_t19 copyright "Copyright (C) 2020 Dario Sanfilippo
+      <sanfilippo.dario@gmail.com> and
+       2003-2020 by Julius O. Smith III <jos@ccrma.stanford.edu>";
+declare avg_t19 license "MIT-style STK-4.3 license";
+avg_t19(period, x) = fi.lpt19(period, x); 
+
+
+/*******************************************************************************
+# Licenses
+
+## STK 4.3 License
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of 
+this software and associated documentation files (the "Software"), to deal in 
+the Software without restriction, including without limitation the rights to 
+use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies 
+of the Software, and to permit persons to whom the Software is furnished to do 
+so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all 
+copies or substantial portions of the Software.
+
+Any person wishing to distribute modifications to the Software is asked to send 
+the modifications to the original developer so that they can be incorporated 
+into the canonical version.  For software copyrighted by Julius O. Smith III, 
+email your modifications to <jos@ccrma.stanford.edu>.  This is, however, not a 
+binding provision of this license.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL THE 
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, 
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE 
+SOFTWARE.
+
+--------------------------------------------------------------------------------
+
+## LGPL License
+
+This program is free software; you can redistribute it and/or modify it under 
+the terms of the GNU Lesser General Public License as published by the Free 
+Software Foundation; either version 2.1 of the License, or (at your option) any 
+later version.
+
+This program is distributed in the hope that it will be useful, but WITHOUT ANY 
+WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A 
+PARTICULAR PURPOSE.  See the GNU Lesser General Public License for more details.
+
+You should have received a copy of the GNU Lesser General Public License along 
+with the GNU C Library; if not, write to the Free Software Foundation, Inc., 
+59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
+
+--------------------------------------------------------------------------------
+
+## ISC License
+
+Permission to use, copy, modify, and/or distribute this software for any
+purpose with or without fee is hereby granted, provided that the above
+copyright notice and this permission notice appear in all copies.
+
+THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH
+REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY
+AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT,
+INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM
+LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR
+OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
+PERFORMANCE OF THIS SOFTWARE.
+
+*******************************************************************************/