While external control is limited to modulation of the global delay time, we enjoyed patching in some subtle random modulation from Marbles to impart some pseudo wow-and-flutter-like sounds. There are few delay modules out there that are as malleable as the Multi-Delay, which made it an instant pick in hopes of sparking creative ideas from Katrina.

Speaking of highly flexible modules, we also chose the deeply powerful Rainmaker from Intellijel to complement our use of the Multi-Delay. On the surface, it would seem that both of these modules have some functional overlap, in that they both function as multi-tap delays. While this is true, Rainmaker drastically expands on the idea by including preset memory, granular pitch shifting, and comb filters, thus extending the reach of the module well beyond simple delays and comfortably into the territories of reverb, resonators, and more.

Less of a specific effect and more of a clever patching trick, we wanted to experiment with layering Katrina's voice with more synthesized timbres, and perhaps one of the best ways to do this was by taking a copy of her vocal signal and running it through a comparator. We chose a classic and reliable design for this purpose, leaning upon the Random*Source Serge NCOM, which packs both a comparator and a pulse divider into a single module. An in-depth discussion on comparators can be found in our Learning Synthesis article on logic functions, but a basic explanation should be sufficient for this article. Our audio signal is being compared to a static reference voltage (just above 0V in our case), and every time the waveform crosses above this threshold, a gate signal is generated, and will likewise turn off when it falls below that value. This is super handy for deriving gates and triggers from fluctuating control signals, but at audio rates our gate signal is audible as a square wave that cleanly follows the pitch of an incoming monophonic source.

As long as Katrina vocalizes sound loud enough to surpass the threshold, the resulting comparator audio will smoothly follow her melodies and vibrato. While it doesn't impart the inflections of words onto the synthetic audio as something like a vocoder would, it can be super handy to layer in as a harmonically relevant and interactive texture underneath a vocal line. Placing it before all of the other effects also means that it can be smeared and warped alongside the dry vocal signal. The divider section comes into play after the comparator signal is generated, and can be used to generate sub-octaves or further divisions of the audio frequency, specifically up to 31 divisions on the NCOM. At audio rates, you'd have to be generating an excruciatingly high frequency signal to make use of high values for N, but low divisions the audio signal down in the subharmonic series, for example producing frequencies an octave, an octave and a fifth, and two octaves below the fundamental where N is two, three, and four respectively.

Given the complexity of managing several parallel signal paths, we found ourselves in need of an adaptable mixer with a dash of patch-programmability philosophy. With our need for multiple stereo inputs, a crossfader to blend between the dry and comparator signals, and voltage control over the channel volume levels, few candidates fit the bill as well as the SSF Vortices and its handy expander. But yet, even at this point, after all the effects and complex signal routing, there's still one more thing we wanted to incorporate from our previous video with Katrina: looping.

In the world of guitar pedals, there's an incredible number of straightforward, no-fuss loopers to aid instrumentalists in capturing and layering musical phrases. Although Eurorack looper modules offer increased control over their pedal counterparts, the learning curve and approachability are drastically different, so we wanted to offer Katrina something that was as familiar to her as possible. In the end, we chose to make use of dual 2hp Loop modules, one each for the left and right channels of our mixer output. This way, Katrina could explore various effect combinations and as soon as she stumbles upon an idea she likes, it can be recorded and looped as is, and subsequent layers could use an entirely different mix of effects if desired. We're also making use of the WMD AXYS as a stereo crossfader to blend between the mixer and looper outputs.

So while we do lean upon the potent pairing of Maths and Marbles for continuous control voltages within the main system itself, we also loaded up some manual controller modules into a pair of 4ms Pods as satellite cases for Katrina to use while performing. Given the density and number of patch cable connections involved with the modules we're using, we want to keep connections between the main and control systems as tidy as possible. Our best efforts to consolidate things down is achieved with the Doepfer A-180-9 Multicore modules, condensing what would have been fourteen 3.5mm cables down to a mere two network cables.

Occupying the entirety of a 20hp Pod, Intellijel's Tetrapad is capable of many things, but with a layout of four touch faders, it was an obvious choice for a mixer control surface. Between the dry vocal signal and separate parallel outputs from the Multi-Delay, Rainmaker, and Filter/Fourcaster, as well as the complex patching facilitated by the normalizations inside Vortices, it quickly becomes difficult to keep track of what knob affects what, even to an experienced patcher at the helm of the effects system. But by patching the Tetrapad CV into the corresponding input on the Vortices expander, we can achieve a more straightforward and approachable way to manage signal levels in our effects system, and certainly one that makes sense for Katrina to adjust as she's singing.

In the other Pod, there are two main things we'd like to achieve: offer Katrina some kind of simultaneous control over several parts of the patch, and facilitate a straightforward approach to looping with the 2hp Loops. Our first goal is easily handled by the Intellijel Planar 2—it provides plenty of voltage outputs based upon the joystick position that can be patched to different parameters in a macro-controller fashion. Among other things, we like using this to blend between the dry vocal and comparator signals, as well as modifying the delay time on the Multi-Delay.

To adequately handle recording and blending the loops, we decided upon two modules. While we can't understate how much we love the goose sample that comes loaded in Winterbloom's Big Honking Button, the arcade-style button itself can be used as a manual gate generator, and patching it into the Record input on the 2hp Loop behaves just as it would on a looper pedal. Once Katrina has an idea of a phrase she'd like to capture, she can tap the button once to start recording a loop, and tap it again to stop recording and begin looping. Once the loop length is set, pressing the button again will toggle recording on and off for overdubbing additional lines. We previously mentioned that the WMD AXYS was serving as our crossfader between the loopers and mixer output—Katrina can use the SM800a Wobla to blend the loop in and out as she pleases.