Intro to "Spiral" Draft (Vangelis Cover) - Roland System-500, Audiothingies Doctor A, OB-6, Lyra-8

video by thesrabbit

"This started off as an exercise on how to do auto-panning within a Eurorack system, without having any mixer modules with pan controls or CV panning inputs. I thought about what sort of musical piece would be a good way to test out my explorations, and the answer is "Spiral" by Vangelis, of course! The intro is probably my favorite synth piece of all time. The breakdown of the patch used here for the main sequence is as follows, from left to right:

Roland System-500 512 Dual VCO:
One oscillator sawtooth and pulse wave out, range set to 16' (should have done 32'). PWM is applied via an LFO.

Roland System-500 521 Dual VCF:
Oscillator sawtooth and pulse coming into the filter's mixer, level of the pulse wave is 100%, and the level of the sawtooth is about 50%
Cutoff set low to start with, resonance about half-way up. The cutoff inputs are fed from the outputs of an envelope generator and an LFO. Playing with the levels of these inputs in real time along with the resonance knob is critical to capturing the vibe of the original sequence and delivers that pulsating sound.

Roland System-500 505 Dual VCF:
Output of the 521 is going to the input of the first filter in the 505. I'm only using this module for its built-in VCA. The cutoff is open all the way. This does add some extra brightness to the sound, but my other VCA's are all used up in this patch. Once the signal runs through the 505's VCA output, it is effectively a complete single synth voice, ready to be recorded or routed to a mixer, etc.

Buffered Multiple:
The first mult input is the output of an envelope generator from the Dreadbox Ataxia. One of the mult's outputs is routed to one of the 521's cutoff inputs and another one of the mult's outputs is routed to the 505's VCA level input. The second mult input is the output of the 505's VCA. Two outputs of the second mult are routed to the inputs of each VCA in a Doepfer Dual VCA module on the far right. Using a buffered mult means that the outputs have exact copies of the input signal. Having 2 identical copies of the single synth voice is critical to the auto-panning process.

Dreadbox Utopia:
The triangle LFO of the Utopia is routed to both the pulse width input of the 512 oscillator and to one of the cutoff inputs on the 521 VCF. The rate of this LFO is pretty slow. This slow modulation of the pulse width and filter is another critical component to recreating the vibe of the original sound.

Dreadbox Ataxia:
Mod 1 is setup as a standard ADSR envelope. Its output is split and sent to one of the 521's cutoff inputs and the 505's VCA level input. This is already mentioned above in the "Buffered Multiple" section. Attack is 0, the decay is set to around 40% and sustain and release are probably no more than 10%. This envelope is triggered by the gate of incoming notes from the sequencer.
Mod 2 is setup as a fairly slow LFO, with a wave shape that's somewhere between a triangle and a sawtooth. Its output is split into 2 with a passive mult and routed to the "rise" and "fall" inputs of the Make Noise Function module.

Make Noise Function:
This is setup to basically function as a slow triangle-ish LFO with a positive and negative output. The positive output is routed to the CV input of VCA 1 of the Doepfer Dual VCA module, and the negative output is routed to the CV input of VCA 2. Cycle mode is enabled and the "rise" and "fall" parameters are continuously manipulated in real-time via the output of an LFO from the Dreadbox Ataxia. The net effect of this is that audio doesn't simply sweep from left to right at a steady even pace. The panning speed changes in accordance with the signal coming from the Dreadbox Ataxia's LFO. This variable panning speed in conjunction with external stereo delay creates a nice swirling spacey effect.

Doepfer Dual VCA:
Each VCA's input is an exact copy of the output of the 505's VCA (the complete synth voice). The CV input levels on each VCA are maxed out and the gain control is maxed on the second VCA. The positive output of the Make Noise Function is routed to VCA 1's CV input and the negative output is routed to VCA 2's CV input. As the positive output of the Make Noise Function increases (moves up and away from zero), what this does is effectively pull the level of VCA 2's output down (by applying negative voltage) while simultaneously pushing the level of VCA 1's output up. As the positive and negative output from the Make Noise Function approaches 0, the level of VCA 1 drops and the level of VCA 2 increases (as the pullback applied to the maxed out gain level of VCA 2 diminishes). This is the final piece of the auto-panning puzzle. The output of VCA 1 is then connected to an input on a desktop mixer that's panned hard left and the output of VCA 2 is connected to an input on the same mixer that's panned hard right."