mike_mccue
drewfx1
mike_mccue
Back in the days before digital, a rotary switch on an analog EQ may have been be thought of as ideal for providing both precise settings and recall.
Now, in the era of digital precision and total recall from memory, emulating rotary switches seems fairly dim witted.
There are actually potential advantages from a DSP perspective, depending.
:-)
You know that a statement like that just makes me want to learn more about it. Please?
:-)
These things aren't always a problem, but...
Filters work by multiplying the signal by a series of filter
coefficients and adding the result back into the signal. Calculating these coefficients directly can be difficult to do on the fly so they are often calculated in advance.
If you have a relatively limited number of possible settings, you can just calculate them in advance and store them in a lookup table. But if you have a more continually varying, say frequency, you either:
1. Need a huge lookup table with values for every possible combination of settings down to the limits of adjustment.
2. Need to calculate coefficients on the fly.
3. Fudge it by interpolating pre-calculated coefficients in one way or another.
In most cases filters are implemented as IIR filters, because they are much more efficient in terms of processing than FIR filters. The first "I" in IIR is for "Infinite" (vs. "Finite") Impulse Response and the reason it's infinite (in theory) - and more efficient - is because there is a feedback loop in the filter.
Under certain conditions this feedback loop can cause the filter to be unstable, similar to what happens with an analog delay overloading when you crank up the feedback. One area where they can potentially become unstable is when rapidly changing settings.