Repository navigation
Flanger
10 HP. A short modulated delay, 0.1 to 10 ms.
This is the module the whole set was built around, because it is the one the library genuinely does not have.
| Control | Range | Default |
|---|---|---|
| Rate | 0.02 to 10 Hz | 0.4 Hz |
| Depth | how much of the 0.1 to 10 ms span the sweep covers | 60% |
| Feedback | -100% to +100% | +35% |
| Mix | dry to wet | 50% |
| Port | Direction |
|---|---|
| Rate, Depth, Feedback, Mix | CV in, one per knob, each with an attenuverter |
| IN L, IN R | audio in. R is normalled to L |
| OUT L, OUT R | audio out |
Bypass routes IN L to OUT L and IN R to OUT R.
Most flangers give you a feedback knob that runs from zero upwards. That covers the resonant, singing version of the effect and stops there.
Inverting the feedback path inverts the comb filter, which moves the peaks to where the notches were. The result is hollow and thinned rather than resonant, and it sounds different enough that treating it as the same control at a different setting would be misleading. Sweeping the knob through centre gives you both, and the transition through zero feedback is itself useful.
This matters most on sustained material. On a drum loop the difference is subtle. On a pad it is not.
At delay lengths this short, linear interpolation between samples audibly dulls the wet signal. You lose high frequency content exactly where a flanger is supposed to be doing its most obvious work, and the effect ends up sounding softer than it should without any single control being at fault.
Cubic interpolation costs more arithmetic per sample and fixes it. On a longer delay I would not have bothered.
The left and right sweeps sit a quarter cycle apart. There is no control for it.
I tried making it adjustable and removed it. There is one value that gives good width without the two sides drifting into sounding like separate effects, and a knob for it would mostly have been a way to get it wrong. If you want a different relationship, run two instances.
The same thing happens here as in the Phaser, for the same reason.
A feedback loop around a delay line has nothing damping it, so at high feedback the output builds until it is a level problem rather than an effect. During testing this module peaked at 32 V before the trim went in, which is well past anything Rack expects on a cable.
The signal entering the loop is now attenuated as feedback rises, holding the peak near twice unity. Below roughly half feedback the trim does nothing measurable. The measured peak at maximum feedback is now 7.44 V.
This is a departure from the idea that these modules do exactly what the panel says, since the panel does not show you a trim. I decided that was better than a feedback knob whose top third is unusable, and writing it down here is the compromise.
At maximum depth and maximum rate the delay line is moving fast enough to produce obvious pitch modulation on the wet signal. Some people want that. If you do not, pull the rate back rather than the depth, because depth is doing the work you actually came for.
Modules
Reference