Based on/inspired by the Empress Echosystem algorithm of the same name. Multiswell uses four taps (as opposed to the ES’s five — CPU casualty), and instead of using a flanger to create stereo imaging, the taps are given individual panning, which allows for subtle or dramatic shifting of the stereo field.
Because the delay swells in, the effect is somewhat like “smearing” — the sounds produced resemble (to me) reverbs more than delays, but undulating reverbs, like a series of waves. You can produce some serious soundscapes with this business.
This patch also uses Empress’s (fairly) unique multi-tap system — instead of choosing between a bunch of preset clock divisions, you tap in your own rhythm, which creates more organic and unique multi-tap lines.
The dry signal is stereo throughout. The wet signal is summed to mono then output in stereo via the panning.
Left — tap in your rhythm! As each tap is activated, a white pixel will turn on, indicating that tap is now activate. You don’t have to use all four taps — use it as a two or three tap delay if you prefer. (If you want to use it as a single-tap delay, I recommend my Attack Delay patch, which has more features in the same vein: https://patchstorage.com/attack-delay-an-auto-swell-envelope-and-delay/)
Middle — tap once to reset the tap rhythm
Right — loop the delay buffer. This also closes the send to the delay, allowing you to play over the loop, which can create some tremendous pads/drones.
In the top left corner are controls for the swell envelope.
You can set an attack time of up to 3 seconds. As the taps overlap, you get lush waves of sound. Or you can set the attack to 0 and use it as a traditional multi-tap delay.
You can also change the slope of the attack, from exponential to linear. As you move from exponential to linear, the swell becomes smoother but less dramatic in its rise.
You can disable the sustain stage of the envelope for pseudo-reverse delay sounds.
The pixel shows the state of the envelope, which can be useful for dialing in sounds and setting sensitivity.
In the middle of the patch are controls for the taps.
The first tap is always at unity (relative to the wet send level; more on this in a moment), but each tap has a “level” button that allows you to switch between unity (on) and -6 dB (off) for that output. This can help create some subtle variation in the pattern.
The first tap is also always centered, but the three subsequent taps can be panned across the stereo field, creating shifting stereo images.
There are global controls for the delay across the bottom.
“Tone” controls a simple low-pass filter. Since the delays capture the ringing tails of sounds, they can become quite sharp — the tone control helps them sit beneath your playing better.
There are also controls for feedback, mod rate, and mod depth — these all work as you might expect. Because the delay is “lossless” it will loop when feedback is set to 1 and it will not oscillate.
In the top right corner are global controls for the patch.
The patch has a dry level instead of a mix. 1 is unity gain.
Instead of a wet “level” there is a wet send — this is important to keep in mind because the tone control and level are set before the delay network. When you loop the delay buffer you cannot subsequently affect its level or tone. (This was _much_ cheaper, computationally, than having an output stage to control these, but it does mean you have to know what you want before looping the delay.)
There is also a sensitivity control for the onset detector. You have to play to this control somewhat but it’s worth it.