Main Archive Specials Wiki | FAQ Links Submit Forum


Karlsen Fast Ladder

Type : 4 pole ladder emulation
References : Posted by arifovekarlsen[AT]hotmail[DOT]com

Notes :
ATTN Admin: You should remove the old version named "Karlsen" on your website, and rather include this one instead.

Code :
// An updated version of "Karlsen 24dB Filter"
// This time, the fastest incarnation possible.
// The very best greetings, Arif Ove Karlsen.
// arifovekarlsen->hotmail.com

b_rscl = b_buf4; if (b_rscl > 1) {b_rscl = 1;}
b_in = (-b_rscl * b_rez) + b_in;
b_buf1 = ((-b_buf1 + b_in1) * b_cut) + b_buf1;
b_buf2 = ((-b_buf2 + b_buf1) * b_cut) + b_buf2;
b_buf3 = ((-b_buf3 + b_buf2) * b_cut) + b_buf3;
b_buf4 = ((-b_buf4 + b_buf3) * b_cut) + b_buf4;
b_lpout = b_buf4;



Comments


Added on : 08/01/07 by nobody[ AT ]nowhere[ DOT ]com
Comment :
Where are the coefficients? How do I set the cutoff frequency?



Added on : 09/01/07 by scoofy[ AT ]inf[ DOT ]elte[ DOT ]hu
Comment :
The parameters are:
b_cut - cutoff freq
b_rez - resonance
b_in1 - input

Cutoff is normalized frequency in rads (2*pi*cutoff/samplerate). Stability limit for b_cut is around 0.7-0.8.

There's a typo, the input is sometimes b_in, sometimes b_in1. Anyways why do you use a b_ prefix for all your variables? Wouldn't it be more easy to read like this:

resoclip = buf4; if (resoclip > 1) resoclip = 1;
in = in - (resoclip * res);
buf1 = ((in - buf1) * cut) + buf1;
buf2 = ((buf1 - buf2) * cut) + buf2;
buf3 = ((buf2 - buf3) * cut) + buf3;
buf4 = ((buf3 - buf4) * cut) + buf4;
lpout = buf4;

Also note that asymmetrical clipping gives you DC offset (at least that's what I get), so symmetrical clipping is better (and gives a much smoother sound).

-- peter schoffhauzer




Added on : 26/06/07 by arif[ AT ]str8dsp[ DOT ]com
Comment :
Tee b_prefix is simply a procedure I began using when I started programming C. Influenced by the BEOS operating system. However it seemed to also make my code more readable, atleast to me. So I started using various prefixes for various things, making the variables easily reckognizable. Peter, everyone, I am now reachable on www.str8dsp.com - Do also check out the plugin offers there!



Added on : 17/07/07 by aok[ AT ]str8dsp[ DOT ]com
Comment :
Here's even another filter, I will probably never get around to making any product with this one so here it is, pseudo-vintage diode ladder.

            Diode Ladder, (unbuffered)

            // limit resonance, rzl, tweak smearing with fltw, 0.3230 seems to be a good vintage sound.
            in = in - rzl;
            in = in + ((-in +kbuf1) * cutoff);  
            kbuf1 = in + ((-in + kbuf1) * fltw);
            in = in + ((-in +kbuf2) * cutoff);  
            kbuf2 = in + ((-in + kbuf2) * fltw);
            etc..




Added on : 09/09/07 by dev[ AT ]fxpointaudio[ DOT ]com
Comment :
"Cutoff is normalized frequency in rads (2*pi*cutoff/samplerate):

This seems to be valid for very low ( < 200 Hz ) frequencies - higher sample rates seem to be "Closer"

thanks




Added on : 17/07/10 by checkpageforemail[ AT ]checkpageforemail[ DOT ]com
Comment :
I also did a 9th order gaussian filter (minimal phase), using only 5 orders, for my limiter, which is released under the GPL LICENCE. http://www.paradoxuncreated.com



Added on : 14/11/12 by Generalized perfect digital "ladder" filter, with the desired aspects of analog[ DOT ]
Comment :
Hi, I have now generalized the ladder filter, into fast code, and with the desired aspects of analog, but retaining digital perfectness.

Please see my blog: http://paradoxuncreated.com/Blog/wordpress/?p=1360

Peace Be With You.




Added on : 21/06/13 by pleasesee[ AT ]blog[ DOT ]com
Comment :
I have also moved domains now, and consolidated the information on this ultimate digital filter, with "analog sound", here:

http://ovekarlsen.com/Blog/abdullah-filter/

Peace Be With You!




Added on : 13/02/16 by ove hy karlsen [ AT ] facebook[ DOT ]com
Comment :
Karlsen Fast Ladder III - inspired by "transistors set to work as diode" type Roland filters. The best fast and non-nonsensical approximation of popular analog filter sound, as in for instance Roland SH-5, and the smaller TB-303.

//Coupled with oversampling and simple oscs you will probably get the best analog approximation.

//        // for nice low sat, or sharper type low deemphasis saturation, one can use a onepole shelf before the filter.
//        b_lf = b_lf + ((-b_lf + b_v) * b_lfcut); // b_lfcut 0..1
//        double b_lfhp = b_v - b_lf;
//        b_v = b_lf + (b_lf1hp * ((b_lfgain*0.5)+1));  

        double b_rez = b_aflt4 - b_v; // no attenuation with rez, makes a stabler filter.
        b_v = b_v - (b_rez*b_fres); // b_fres = resonance amount. 0..4 typical "to selfoscillation", 0.6 covers a more saturated range.

        double b_vnc = b_v; // clip, and adding back some nonclipped, to get a dynamic like analog.
        if (b_v > 1) {b_v = 1;} else if (b_v < -1) {b_v = -1;}
        b_v = b_vnc + ((-b_vnc + b_v) * 0.9840);

        b_aflt1 = b_aflt1 + ((-b_aflt1 + b_v) * b_fenv); // straightforward 4 pole filter, (4 normalized feedback paths in series)
        b_aflt2 = b_aflt2 + ((-b_aflt2 + b_aflt1) * b_fenv);
        b_aflt3 = b_aflt3 + ((-b_aflt3 + b_aflt2) * b_fenv);
        b_aflt4 = b_aflt4 + ((-b_aflt4 + b_aflt3) * b_fenv);
        b_v = b_aflt4;

// Behave.
// Ove Hy Karlsen.




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