Conceptual question about filter envelopes in synth building
so I’m currently trying to build a synth in max 5, and it’s working quit well until now, but I have this conceptual dilemma regarding the filter envelope, and the way they are usually implemented in traditional synths:
- does the envelope usually affects the whole output of the filter section, meaning there’s no signal at all if the envelope is at 0 and if the signal is routed all the way through the filter ?,
- or does it only affects some aspects of the filter, like the gain, resonance or cutoff frequency, but leaving the amplitude level at the output of the filter alone ?
- or does it create a balance between the filtered signal and a parallel "dry" signal going around the filter (and multiplied by the same but inverted envelope for instance)
I suppose there’s no absolute right way to do this, but I’m just wondering what was usually done in this matter.
Any suggestion welcome.
You’re right, theres definitely no single ‘right’ way to do this. However a common set up I’ve seen is osc -> filter -> amplifier. In this situation there would be an envelope modulating the filters cutoff frequency, and possibly another (or the same) envelope modulating the filter’s amplitude.
Next, the signal is fed into an amplifier with its own envelope. This envelope is the one that would control the output gain and act as a kind of gate to silence the signal when a note is not playing.
But really, how you decide to connect your envelopes should ultimately be based around the kind of sound you want. Having a different envelope for the cutoff and another for the amplitude is more flexible, but sometimes its interesting to have different parameters shaped with the same contour.
Another thing to consider is efficiency, if your using something like filtercoeff~, there is a gain control built in. As far as I know the amplitude characteristics for biquad~ are linear so it doesn’t matter if you are modulating the gain pre or post filter. So its probably easier to simply connect a gain control to the filtercoeff~ gain input than add a *~ to the output of biquad~.
the envelope modulating the filtercoef~ gain input seems to give good results, indeed.