Filters

Rowena Larkins rowenal at nimrod.itg.telstra.com.au
Thu Nov 27 02:40:26 CET 1997


Hi Kurt, 

The location of the poles also affects the characteristics above the cutoff
frequency - how ripply it is before settling down. Thus you can get a 
Chebychev filter etc.
The signal analysis is quite complex and requires advanced maths using complex
analysis.

As an example of poles, a simple "tone" control on a cheap radio or stereo is 
usually a single pole, or 6db filter. This is typically a pot and a capacitor

Hi fi amps/equalisers with bass and treble controls are usually 12 db/oct or
 two pole filters. 

The number of poles can generally be related to the number of frequency 
dependant components used in a filter.

A single RC or RL filter gives 6db/oct. 
A more complex RLC filter gives 12db/oct.

Synths commonly have filters with 24 db/oct.

Rowena

> >While doing some research into the archives for a project I'm working
> >on, I noticed some discussion on the subject of Ladder-style VCF's
> >versus 4 Pole. (Is there a 2 pole? 3 pole? 1 pole? Other types
> >entirely?) Can they both be used for LP, HP, BP, notch?
> >
> >I know it's a bit of a technical distinction, but what exactly
> >differentiates these? What's the difference in sound and is it only
> >noticeable when self oscillating?
> 
> You're talking about two different aspects of the filter here. The fact
> that it is a "ladder" filter simply refers to the configuration of the
> transistors used in the filter, not its roll-off rate. This can be referred
> to either in dB/octave (the rate at which filtered frequencies drop off),
> or the number of poles (technically speaking, the number of points in the
> complex plane  where the filter's characteristic equation goes to
> infinity). 1 pole = 6 dB/octave, so a 4 pole filter has a 24 dB/octave
> roll-off, etc.
> 
> You should be able to build just about any filter configuration with a
> reasonable number of poles and with or without ladder circuitry. Notch and
> bandpass filters always have at least two poles (so you get roll-off on
> each side). I don't believe a one-pole filter will oscillate. 
> 
> The number of poles and the filter architecture certainly affect the sound
> even at low resonance values.
>  
>



More information about the Synth-diy mailing list