Keeping levels low when creating patches which contain several layered oscillators and/or layers is extremely important if you don’t want to end up with a digitally distorted mess of sound. Also compensating for level boosts when using distortion and waveshaping modules or other effects which induced level boosting will help you to keep things under control. It’s much easier to boost levels at the end of a chain then trying to tame them and fail at removing unwanted distortions which occurred much earlier in the chain.
There is no rule for this, but if you start a patch which will contain several oscillators or an oscillator running in unison mode or a multi-granular oscillator with several grain streams, begin with something like -6 dB for oscillator level, -3 dB for keygroup level and -2 dB for layer level. Wavetable oscillators with many unison voices and other unison synths can also get very loud if you play many voices at once, so begin with oscillator levels around -8dB or even lower. Some of Falcon’s filters react very sensitive if you drive them too hot on input level (some of the hybrid filters, the resonance filters too), so if you’re using these filter types, turn down oscillator level to a healthy medium level.
If you’re using a waveshaper on keygroup level with mix set to 1005% and a fair amount of depth, -12 – 10 dB seems like a sensible level to begin with. If you are modulation waveshaper depth/mix something like an LFO, assign either keygroup or oscillator level to the same LFO with an inverted polarity. and a modulation value of something around 0.3 – 0.5, you may want to use a modulation mapper to shape the modulation slope of the level compensation.
Using several filters or effect modules inside an FX rack also requires to think about the gain staging, using 3 filters in parallel and setting their level to 0 dB each will of course totally overload the output if all 3 signals play together. I often assign Macros or other modulators to the individual volume of each component inside an FX chain and will start with maximum levels for each signal of about -6 – -8 dB.
If you’re using a tuned comb-filter (it’s frequency assigned to linear key follow, modulation value set to 0.501) with a high resonance value, you will get extreme level boosts/overloads once you reach the 90%+ mark, so if you have a Macro assigned to comb-filter resonance, assign oscillator level (inverted) to the same Macro with a mapper shaping the slope, as the real level boosting only starts towards the end of the modulation range. You will find that modulation values between – 0.8 – 0.9 will be needed to tame the overloads created by adding comb-filter resonance and your signal will remain crystal clear and free of unwanted distortion.
To be continued…