No decent simple xover filter??

[Ryu]

Hello everybody. I cannot beleive someone still hasn't made a simple decent xover filter. Like something that might look like this:
Ahh.. I can't manage to insert an image. Can anyone tell me how to?

If someone knows of a decent xover filter that has the following options:

Bandpass
Highpass
Lowpass

Filter type selection:
Butterworth
cebychev
Riley
ect.

Frequency range
eg. 600-4000

Slope
Eg: Order 2 (12db/oct)

and maybe a ripple also ?

That simple!
Nothing? Anywhere??

Bye to all of you!

[MacButch]

Hello everybody. I cannot beleive someone still hasn't made a simple decent xover filter.

mmmm.... I started writing one, it seems that no one is really interested in this kind of thing as I didn't bother to polish it properly. It is just something that I use myself now. Though it doesn't have all the filter types etc that you mentioned.

I guess I could clean it up one day... I'll post here if I do get to it...

[Music Engineer]

Hello everybody. I cannot beleive someone still hasn't made a simple decent xover filter.

Filter type selection:
Butterworth
cebychev
Riley
ect.

maybe it's because it isn't that simple - if you want to do it right ("decent"). all of the above mentioned filters are IIR-types which means that they have a nonlinear phase. this in turn leads to the effect, that if you use a -say- butterworth-lowpass and a butterworth-highpass with matching cutoff-frequencies as crossover, the frequency-response of the sum of both filters would NOT be a constant. instead you would get some (quite a lot) of ripple near the x-over frequency. one way to avoid this would be to derive your highpass-signal by subtracting a lowpass-signal from the original - but this "highpass filter"s freq-response would then not look like a nice highpass. the only way i know, to do it right involves linear-phase filtering which mandates the use of FIR-filter technology (which is CPU-intensive) - but maybe someone can point us to more efficient techniques.

[MacButch]

maybe it's because it isn't that simple - if you want to do it right ("decent"). all of the above mentioned filters are IIR-types which means that they have a nonlinear phase. this in turn leads to the effect, that if you use a -say- butterworth-lowpass and a butterworth-highpass with matching cutoff-frequencies as crossover, the frequency-response of the sum of both filters would NOT be a constant. instead you would get some (quite a lot) of ripple near the x-over frequency. one way to avoid this would be to derive your highpass-signal by subtracting a lowpass-signal from the original - but this "highpass filter"s freq-response would then not look like a nice highpass. the only way i know, to do it right involves linear-phase filtering which mandates the use of FIR-filter technology (which is CPU-intensive) - but maybe someone can point us to more efficient techniques.

You're right, of course, certainly the xover that I was working on had some interesting artifacts due to the way I was deriving the high-pass filters. I consider this character but thats probably just me

[Music Engineer]

You're right, of course, certainly the xover that I was working on had some interesting artifacts due to the way I was deriving the high-pass filters. I consider this character but thats probably just me

hihi...it is of course always possible to consider technical imperfections as "character" - and sometimes they are even introduced intentionally. mmmhhh...maybe with bessel-filters it would be possible to come close to the ideal with IIR-technology - as they are optimized for linear phase. but i must admit, that i didn't spend much time on thinking about IIR-crossovers yet

[Music Engineer]

i found another solution which has 2 useful properties:

1. the hpf and lpf outputs are well behaved (their freq-responses look like as one would expect)
2. the hpf and lpf outputs sum up to the original signal

...and 1 major drawback:

the separation slope is only 6 dB/oct

it works as follows:
take a first order allpass-filter, adjust the characteristic frequency (the freq where the phase passes 90 degrees) to the desired crossover frequency.
then calculate:

Code: Select all

lpfOut = 0.5 * (original + apfOut);
hpfOut = 0.5 * (original - apfOut);

it works, because a first order allpass has a phase-response which goes form 0 to -180 degrees.

[Ryu]

AUTO-ADMIN: Non-MP3, WAV, OGG, SoundCloud, YouTube, Vimeo, Twitter and Facebook links in this post have been protected automatically. Once the member reaches 5 posts the links will function as normal.

Hello guys. What do you think about this crossover?

http://www.blocc.biz/ (http://www.blocc.biz/)

Go on: Products - hipass or lopass filter, and then on the tech specs.

It seems to be quite good.
Any comments?
Thanks everyone!