TOPIC: Re: 4th order bessel crossover

Hi, a fourth-order Bessel filter crossover would be handy. See:

www.rane.com/note147.html

Hi John,

How about simply plug-in it into the biquad spreadsheet (i.e. update the custom spreadsheet) and you can build it in very little time?

There are many filters out there, and put it simply, we're never going to be able to have all of them implemented. There will always be a new one required..

DevTeam

The spreadsheet only does second-order filters.

John,

Do you know how to do this, or are you aware and just asking miniDSP to incorporate the functionality?

If the former, then it's a little cutting and pasting from the spreadsheet and you can get there in a few minutes.

For example: For a 4th-order, 400Hz Bessel LPF, you'd input 400Hz and Q=0.759 into the spreadsheet which will generate a set of 2nd-order coefficients. Paste those into the biquad1 and biquad2 locations and fill in the rest as usual to paste into the advanced plug-in.

This will yield a 400hz Bessel LPF with Q=0.577.

Your result should look like this:

biquad1,
b0=0.000662395277357,
b1=0.001324790554713,
b2=0.000662395277357,
a1=1.930694674258500,
a2=-0.933344255367930,
biquad2,
b0=0.000662395277357,
b1=0.001324790554713,
b2=0.000662395277357,
a1=1.930694674258500,
a2=-0.933344255367930,
biquad3,
b0=1,
b1=0,
b2=0,
a1=0,
a2=0,
biquad4,
b0=1,
b1=0,
b2=0,
a1=0,
a2=0,
biquad5,
b0=1,
b1=0,
b2=0,
a1=0,
a2=0,
biquad6,
b0=1,
b1=0,
b2=0,
a1=0,
a2=0,
biquad7,
b0=1,
b1=0,
b2=0,
a1=0,
a2=0,
biquad8,
b0=1,
b1=0,
b2=0,
a1=0,
a2=0,

Hope that helps.

Dave.

The biquad spreadsheet calculates 2nd order filter coefficients. Higher order filters for dsp's are constructed by cascading 2nd order filters. So a 4th order filter is constructed by using two 2nd order filters, a 6th order by using three 2nd order filter and so on.

The trick is to use the correct Q values for calculating higher order filters. At section 16.9 in this document you can find these Q values (for Linkwitz-Riley filters look here).

In table 16-4 you'll find that a 4th order Bessel filter is constructed by cascading two 2nd order filters, the first one having a Q of 0.52, the second one with a Q of 0.81.
In the spreadsheet you calculate these two filters (400Hz LPF) giving the following result:

biquad1,
b0=0.000652401784953,
b1=0.001304803569907,
b2=0.000652401784953,
a1=1.901566473590970,
a2=-0.904176080730782,
biquad2,
b0=0.000663788184815,
b1=0.001327576369630,
b2=0.000663788184815,
a1=1.934754605093700,
a2=-0.937409757832964,
biquad3,
b0=1,
b1=0,
b2=0,
a1=0,
a2=0,
biquad4,
b0=1,
b1=0,
b2=0,
a1=0,
a2=0,
biquad5,
b0=1,
b1=0,
b2=0,
a1=0,
a2=0,
biquad6,
b0=1,
b1=0,
b2=0,
a1=0,
a2=0,
biquad7,
b0=1,
b1=0,
b2=0,
a1=0,
a2=0,
biquad8,
b0=1,
b1=0,
b2=0,
a1=0,
a2=0

Yes indeed. Thank you for correcting my mistake.

Cheers,

Dave.

Thanks guys! I'll give that a try.

With regard to the specific suggestion I was suggesting that miniDSP might be interested in implementing the "phase match" normalization described in the Rane paper: "Rane's Bessel crossover is set for phase match between low-pass and high-pass. This minimizes lobing due to driver separation, and also results in a pretty flat combined response."