Main Archive Specials Wiki | FAQ Links Submit Forum


4th order Linkwitz-Riley filters

Type : LP/HP - LR4
References : Posted by neolit123 at gmail dot com

Notes :
Original from T. Lossius - ttblue project

Optimized version in pseudo-code.

[! The filter is unstable for fast automation changes in the lower frequency range. Parameter interpolation and/or oversampling should fix this. !]

The sum of the Linkwitz-Riley (Butterworth squared) HP and LP outputs, will result an all-pass filter at Fc and flat magnitude response - close to ideal for crossovers.

Lubomir I. Ivanov


Code :
//-----------------------------------------
// [code]
//-----------------------------------------

//fc -> cutoff frequency
//pi -> 3.14285714285714
//srate -> sample rate

//================================================
// shared for both lp, hp; optimizations here
//================================================
wc=2*pi*fc;
wc2=wc*wc;
wc3=wc2*wc;
wc4=wc2*wc2;
k=wc/tan(pi*fc/srate);
k2=k*k;
k3=k2*k;
k4=k2*k2;
sqrt2=sqrt(2);
sq_tmp1=sqrt2*wc3*k;
sq_tmp2=sqrt2*wc*k3;
a_tmp=4*wc2*k2+2*sq_tmp1+k4+2*sq_tmp2+wc4;

b1=(4*(wc4+sq_tmp1-k4-sq_tmp2))/a_tmp;
b2=(6*wc4-8*wc2*k2+6*k4)/a_tmp;
b3=(4*(wc4-sq_tmp1+sq_tmp2-k4))/a_tmp;
b4=(k4-2*sq_tmp1+wc4-2*sq_tmp2+4*wc2*k2)/a_tmp;

//================================================
// low-pass
//================================================
a0=wc4/a_tmp;
a1=4*wc4/a_tmp;
a2=6*wc4/a_tmp;
a3=a1;
a4=a0;

//=====================================================
// high-pass
//=====================================================
a0=k4/a_tmp;
a1=-4*k4/a_tmp;
a2=6*k4/a_tmp;
a3=a1;
a4=a0;

//=====================================================
// sample loop - same for lp, hp
//=====================================================
tempx=input;

tempy=a0*tempx+a1*xm1+a2*xm2+a3*xm3+a4*xm4-b1*ym1-b2*ym2-b3*ym3-b4*ym4;
xm4=xm3;
xm3=xm2;
xm2=xm1;
xm1=tempx;
ym4=ym3;
ym3=ym2;
ym2=ym1;
ym1=tempy;

output=tempy;



Comments


Added on : 29/05/09 by neolit123[ AT ]gmail[ DOT ]com
Comment :
LR2 with DFII:

//------------------------------
// LR2
// fc -> cutoff frequency
// pi -> 3.14285714285714
// srate -> sample rate
//------------------------------
fpi = pi*fc;
wc = 2*fpi;
wc2 = wc*wc;
wc22 = 2*wc2;
k = wc/tan(fpi/srate);
k2 = k*k;
k22 = 2*k2;
wck2 = 2*wc*k;
tmpk = (k2+wc2+wck2);
//b shared
b1 = (-k22+wc22)/tmpk;
b2 = (-wck2+k2+wc2)/tmpk;
//---------------
// low-pass
//---------------
a0_lp = (wc2)/tmpk;
a1_lp = (wc22)/tmpk;
a2_lp = (wc2)/tmpk;
//----------------
// high-pass
//----------------
a0_hp = (k2)/tmpk;
a1_hp = (-k22)/tmpk;
a2_hp = (k2)/tmpk;

//=========================
// sample loop, in -> input
//=========================
//---lp
lp_out = a0_lp*in + lp_xm0;
lp_xm0 = a1_lp*in - b1*lp_out + lp_xm1;
lp_xm1 = a2_lp*in - b2*lp_out;
//---hp
hp_out = a0_hp*in + hp_xm0;
hp_xm0 = a1_hp*in - b1*hp_out + hp_xm1;
hp_xm1 = a2_hp*in - b2*hp_out;

// the two are with 180 degrees phase shift,
// so you need to invert the phase of one.
out = lp_out + hp_out*(-1);

//result is allpass at Fc




Added on : 13/07/11 by dave[ AT ]quikquak[ DOT ]com
Comment :
I've converted this Linkwits Riley 4 into intrinsics. It's set up with the cross over point and the sample rate. The function 'ProcessSplit' returns the low and high parts. It uses _mm_malloc to align the variables to 16 bytes, as putting them into the class as __m128 vars doesn't guarantee alignment.
Enjoy!  :)


//-----------------------------------------------
//-----------------------------------------------
//-----------------------------------------------
//FIL_Linkwitz_Riley4.h

#pragma once

#include <xmmintrin.h>

class   FIL_Linkwitz_Riley4
{
    __m128 *ab;
    __m128 *al;
    __m128 *ah;
    __m128 *xm;
    __m128 *yml;
    __m128 *ymh;

    float a0l;
    float a0h;

public:

    FIL_Linkwitz_Riley4::FIL_Linkwitz_Riley4(float fc, float srate);
    ~FIL_Linkwitz_Riley4();

    void ResetSplit();

    __inline void FIL_Linkwitz_Riley4::ProcessSplit(const float in, float &low, float &high)
    {
        __m128 m1;

        m1 = _mm_sub_ps(_mm_mul_ps(*al, *xm), _mm_mul_ps(*ab, *yml));
        low  = a0l * in + m1.m128_f32[0] + m1.m128_f32[1] + m1.m128_f32[2] + m1.m128_f32[3];

        m1 = _mm_sub_ps(_mm_mul_ps(*ah, *xm), _mm_mul_ps(*ab, *ymh));
        high = a0h * in + m1.m128_f32[0] + m1.m128_f32[1] + m1.m128_f32[2] + m1.m128_f32[3];

        *xm  = _mm_shuffle_ps(*xm, *xm, _MM_SHUFFLE(2,1,0,0));
        (*xm).m128_f32[0] = in;
        *yml = _mm_shuffle_ps(*yml, *yml, _MM_SHUFFLE(2,1,0,0));
        (*yml).m128_f32[0] = low;
        *ymh = _mm_shuffle_ps(*ymh, *ymh, _MM_SHUFFLE(2,1,0,0));
        (*ymh).m128_f32[0] = high;
    }


};

//-----------------------------------------------
//-----------------------------------------------
//-----------------------------------------------


// FIL_Linkwitz_Riley4.cpp

#include "FIL_Linkwitz_Riley4.h"
#include <math.h>


FIL_Linkwitz_Riley4::FIL_Linkwitz_Riley4(float fc, float srate)
{
    ab  = (__m128*)_mm_malloc(16, 16);
    al  = (__m128*)_mm_malloc(16, 16);
    ah  = (__m128*)_mm_malloc(16, 16);
    xm  = (__m128*)_mm_malloc(16, 16);
    yml = (__m128*)_mm_malloc(16, 16);
    ymh = (__m128*)_mm_malloc(16, 16);

    float wc  = 2.0f * PI * fc;
    float wc2 = wc*wc;
    float wc3 = wc2*wc;
    float wc4 = wc2*wc2;
    float k  = wc / tanf(PI * fc / srate);
    float k2 = k*k;
    float k3 = k2*k;
    float k4 = k2*k2;
    float sqrt2   = sqrtf(2.0f);
    float sq_tmp1 = sqrt2 *wc3 * k;
    float sq_tmp2 = sqrt2 *wc * k3;
    float a_tmp      = 4.0f * wc2 * k2 + 2.0f * sq_tmp1 + k4 + 2.0f * sq_tmp2 + wc4;
                
    (*ab).m128_f32[0] = (4.0f *(wc4+sq_tmp1-k4-sq_tmp2))/a_tmp;
    (*ab).m128_f32[1] = (6.0f *wc4-8*wc2*k2+6*k4)/a_tmp;
    (*ab).m128_f32[2] = (4.0f *(wc4-sq_tmp1+sq_tmp2-k4))/a_tmp;
    (*ab).m128_f32[3] = (k4 -2.0f * sq_tmp1 + wc4 - 2.0f * sq_tmp2 + 4.0f * wc2 * k2) / a_tmp;

    //================================================
    // low-pass
    //================================================
    a0l      = wc4/a_tmp;
    (*al).m128_f32[0] = 4.0f * wc4 / a_tmp;
    (*al).m128_f32[1] = 6.0f * wc4 / a_tmp;
    (*al).m128_f32[2] = (*al).m128_f32[0];
    (*al).m128_f32[3] = a0l;

    //=====================================================
    // high-pass
    //=====================================================
    a0h      = k4 / a_tmp;
    (*ah).m128_f32[0] = -4.0f * k4 / a_tmp;
    (*ah).m128_f32[1] =  6.0f * k4 / a_tmp;
    (*ah).m128_f32[2] = (*ah).m128_f32[0];
    (*ah).m128_f32[3] = a0h;

    ResetSplit();
}

FIL_Linkwitz_Riley4::~FIL_Linkwitz_Riley4()
{
    _mm_free((void*)ab);
    _mm_free((void*)al);
    _mm_free((void*)ah);
    _mm_free((void*)xm);
    _mm_free((void*)yml);
    _mm_free((void*)ymh);
}


void FIL_Linkwitz_Riley4::ResetSplit()
{
    // Reset history...
    *xm  = _mm_set1_ps(0.0f);
    *yml = _mm_set1_ps(0.0f);
    *ymh = _mm_set1_ps(0.0f);
}
//-----------------------------------------------
//-----------------------------------------------
//-----------------------------------------------





Added on : 13/07/11 by dave[ AT ]quikquak[ DOT ]com
Comment :
I've no idea why I'm accessing those pointers like that! But never mind. :)



Added on : 04/07/12 by cdion[ AT ]quantum-music[ DOT ]ca
Comment :
Your pi value is wrong:
pi -> 3.14285714285714

It should be 3.1415692 ect.




Added on : 31/12/12 by dave[ AT ]quikquak[ DOT ]com
Comment :
Or even 3.1415926535!
LOL.




Added on : 29/07/13 by jim[ AT ]cernproductions[ DOT ]com
Comment :
I don't think this is unstable for changes in frequency.  It's unstable for low frequencies.  

Here's my implementation.


#include <iostream>
#include <stdio.h>
#include <math.h>
#include <assert.h>

class LRCrossoverFilter { // LR4 crossover filter
private:
    struct filterCoefficents {
        float a0, a1, a2, a3, a4;
    } lpco, hpco;
    
    float b1co, b2co, b3co, b4co;
    
    struct {
        float xm1 = 0.0f;
        float xm2 = 0.0f;
        float xm3 = 0.0f;
        float xm4 = 0.0f;
        float ym1 = 0.0f, ym2 = 0.0f, ym3 = 0.0f, ym4 = 0.0f;
    } hptemp, lptemp;
    
    float coFreqRunningAv = 100.0f;
public:
    void setup(float crossoverFrequency, float sr);
    void processBlock(float * in, float * outHP, float * outLP, int numSamples);
    void dumpCoefficents(struct filterCoefficents x) {
        std::cout << "a0: " << x.a0 << "\n";
        std::cout << "a1: " << x.a1 << "\n";
        std::cout << "a2: " << x.a2 << "\n";
        std::cout << "a3: " << x.a3 << "\n";
        std::cout << "a4: " << x.a4 << "\n";
    }
    void dumpInformation() {
        std::cout << "-----\nfrequency: "<< coFreqRunningAv << "\n";
        std::cout << "lpco:\n";
        dumpCoefficents(lpco);
        std::cout << "hpco:\n";
        dumpCoefficents(hpco);
        std::cout << "bco:\nb1: ";
        std::cout << b1co << "\nb2: " << b2co << "\nb3: " <<  b3co << "\nb4: " << b4co << "\n";
    }
    

};


void LRCrossoverFilter::setup(float crossoverFrequency, float sr) {
    
    const float pi = 3.141f;
    
    coFreqRunningAv = crossoverFrequency;
    
    float cowc=2*pi*coFreqRunningAv;
    float cowc2=cowc*cowc;
    float cowc3=cowc2*cowc;
    float cowc4=cowc2*cowc2;
    
    float cok=cowc/tan(pi*coFreqRunningAv/sr);
    float cok2=cok*cok;
    float cok3=cok2*cok;
    float cok4=cok2*cok2;
    float sqrt2=sqrt(2);
    float sq_tmp1 = sqrt2 * cowc3 * cok;
    float sq_tmp2 = sqrt2 * cowc * cok3;
    float a_tmp = 4*cowc2*cok2 + 2*sq_tmp1 + cok4 + 2*sq_tmp2+cowc4;
    
    b1co=(4*(cowc4+sq_tmp1-cok4-sq_tmp2))/a_tmp;

    
    b2co=(6*cowc4-8*cowc2*cok2+6*cok4)/a_tmp;

    
    b3co=(4*(cowc4-sq_tmp1+sq_tmp2-cok4))/a_tmp;

    
    b4co=(cok4-2*sq_tmp1+cowc4-2*sq_tmp2+4*cowc2*cok2)/a_tmp;

    
    
    //================================================
    // low-pass
    //================================================
    lpco.a0=cowc4/a_tmp;
    lpco.a1=4*cowc4/a_tmp;
    lpco.a2=6*cowc4/a_tmp;
    lpco.a3=lpco.a1;
    lpco.a4=lpco.a0;
    
    //=====================================================
    // high-pass
    //=====================================================
    hpco.a0=cok4/a_tmp;
    hpco.a1=-4*cok4/a_tmp;
    hpco.a2=6*cok4/a_tmp;
    hpco.a3=hpco.a1;
    hpco.a4=hpco.a0;
    
    
}

void LRCrossoverFilter::processBlock(float * in, float * outHP, float * outLP, int numSamples) {
    
    float tempx, tempy;
    for (int i = 0; i<numSamples; i++) {
        tempx=in[i];
        
        // High pass
        
        tempy = hpco.a0*tempx +
        hpco.a1*hptemp.xm1 +
        hpco.a2*hptemp.xm2 +
        hpco.a3*hptemp.xm3 +
        hpco.a4*hptemp.xm4 -
        b1co*hptemp.ym1 -
        b2co*hptemp.ym2 -
        b3co*hptemp.ym3 -
        b4co*hptemp.ym4;
        
        hptemp.xm4=hptemp.xm3;
        hptemp.xm3=hptemp.xm2;
        hptemp.xm2=hptemp.xm1;
        hptemp.xm1=tempx;
        hptemp.ym4=hptemp.ym3;
        hptemp.ym3=hptemp.ym2;
        hptemp.ym2=hptemp.ym1;
        hptemp.ym1=tempy;
        outHP[i]=tempy;

        assert(tempy<10000000);
        
        // Low pass
        
        tempy = lpco.a0*tempx +
        lpco.a1*lptemp.xm1 +
        lpco.a2*lptemp.xm2 +
        lpco.a3*lptemp.xm3 +
        lpco.a4*lptemp.xm4 -
        b1co*lptemp.ym1 -
        b2co*lptemp.ym2 -
        b3co*lptemp.ym3 -
        b4co*lptemp.ym4;
        
        lptemp.xm4=lptemp.xm3; // these are the same as hptemp and could be optimised away
        lptemp.xm3=lptemp.xm2;
        lptemp.xm2=lptemp.xm1;
        lptemp.xm1=tempx;
        lptemp.ym4=lptemp.ym3;
        lptemp.ym3=lptemp.ym2;
        lptemp.ym2=lptemp.ym1;
        lptemp.ym1=tempy;
        outLP[i] = tempy;
        
        assert(!isnan(outLP[i]));
    }
}


int main () {
    LRCrossoverFilter filter;
    float data[2000];
    float lp[2000], hp[2000];
    
    filter.setup(50.0, 44100.0f);
    filter.dumpInformation();
    
    for (int i = 0; i<2000; i++) {
        data[i] = sinf(i/100.f);
    }
    filter.processBlock(data, hp, lp, 2000);
    
}

I'll try and fix it, but this kind of work is new to me, so all suggestions appreciated (Including "You Fool, you've copied the code wrong"). cheers!              




Added on : 03/09/13 by ck[ AT ]chriskerry[ DOT ]f9[ DOT ]co[ DOT ]uk
Comment :
I tried this code for a crossover - firstly the SSE intrinsics version then the full original version. Both have problems with the HPF output.
With a crossover frequency of 200Hz and a pure sine tone input (any pitch) I get loud (-16dBFS)low frequency noise in the HPF output. This noise level reduces as the crossover frequency increases but it is unusable in its current state.
Can anyone post a solution for this problem?
Thanks.....Chris




Added on : 04/09/13 by ck[ AT ]chriskerry[ DOT ]f9[ DOT ]co[ DOT ]uk
Comment :
I also tried the LR2 code, this works better but there is still low frequency noise (-56dBFS & Xover 200Hz) in the HPF output.
Seems there is a fundamental problem with the HPF coefficients in this code :(
The LF Noise for both LR2 and LR4 appears to be a modulating DC offset - maybe that can guide the Filter Gurus to identify and solve the problem.
Cheers.....Chris    




Add your own comment
Comments are displayed in fixed width, no HTML code allowed!
Email:

Comment:

Are you human?



Site created and maintained by Bram
Graphic design by line.out | Server sponsered by fxpansion