Filter Slope... HPF, LPF, Crossovers

Oct 26, 2010
4,830
11,378
5,041
Surprise, AZ
Hi All,
With all the effort put into using filters for parallel signal paths, has there been much discussion about what filter slope should be used in each application and at each point in the chain? I haven't noticed any.

For example, when using an HPF on the front side of a dirty chain, should it be -6 dB/oct, -12 dB/oct, etc.?

Thoughts? Feelings? Real experience/testing? Measurements of existing commercial products?

- John
 
Last edited:
Rarely filter frequencies are given and even more rare are the slopes of the filters given. Using filters in parallel signal paths (that are then mixed electronically) can cause huge problems with phase issues. Mixing a crossed-over mix with a dry signal can create a huge notch at the cross-over freq. I measured one notch at approx -40dB.

But if you are just interested in general filters without parallel mixing, then....
A high-pass filter, that's used as a "sub sonic" filter, should have a fairly high slope like -24dB/oct to provide the protection that it is meant to do. Or like a low-pass filter to protect tweeters should also be a high slope. Other than "protection" filters, the type of filter and its slope will be a personal choice for the application.:thumbsup:

-Frank
 
But this is exactly what is being done by many folks for mixing clean and dirty paths. Reference the Tech 21 dUg Pinnick pedal, for example. Wonder what order filter they have used in that.

- John

Yes, parallel mixing of phase modified signals is wildly popular. There just aren't any rules or guidelines for doing so. ANYTIME you mix the same source signal that has been phase modified by filters with another version of the same source signal, you will have boosts and cuts due to phase additions/subtractions. Most everyone that I know will use their ears to see what they want in the signal mix. None of them have ever said "I am going to strap a xyz para EQ with a center freq of 628Hz, at 12dB/oct, with a Q of .8 and a boost of 7.2dB to the dirt path because......"

The designers of effects will usually fine-tune filters within an effects design after many hours, days or months of trial and error to get the EQ profile that they desire. Again, there are no rules or guidelines to what type of filters to use. Sometimes the designer will say something like "after years of refinement", or something along those lines, where the only differences thru the years were EQ modifications.:thumbsup:

-Frank
 
Yes, parallel mixing of phase modified signals is wildly popular. There just aren't any rules or guidelines for doing so. ANYTIME you mix the same source signal that has been phase modified by filters with another version of the same source signal, you will have boosts and cuts due to phase additions/subtractions. Most everyone that I know will use their ears to see what they want in the signal mix. None of them have ever said "I am going to strap a xyz para EQ with a center freq of 628Hz, at 12dB/oct, with a Q of .8 and a boost of 7.2dB to the dirt path because......"

The designers of effects will usually fine-tune filters within an effects design after many hours, days or months of trial and error to get the EQ profile that they desire. Again, there are no rules or guidelines to what type of filters to use. Sometimes the designer will say something like "after years of refinement", or something along those lines, where the only differences thru the years were EQ modifications.:thumbsup:

Yep. I mess around a lot with ambient satellite cabs and clean/dirty rigs. Even if I carefully measure out spacing distances, spinning the knob on my variable HPF to the satellite(s) tends to feel like throwing darts. Now throw in delay compensation, and spontaneous head explosion (mine, not the amp) is a distinct possibility. Or as I've heard all too often in what passes for my playing career: "who's the leader of this anarchy band?"