# Robaux — DCSN-3 - [Manual PDF](../../manuals/robauxDCSN3Manual.pdf) --- [Manual PDF](#) ## Robaux Decision Tree — using it for melodic components The attached manual pages describe the **Robaux Decision Tree**, a trigger/gate routing and clock-divider module. It is **not a pitch generator by itself**, but it is very useful for building **melodic structure** when paired with oscillators, quantizers, envelopes, sequencers, shift registers, sample & hold, or precision adders. ## What the module does At its core, the Decision Tree: - takes a **gate/trigger** at **input a** - routes that trigger **randomly** to: - one of the three main outputs **b / c / d** - then to one of the related suboutputs: - **e f g** - **h i j** - **k l m** This creates branching rhythmic paths. It can operate in: - **Mono/Mono** - **Poly/Mono** - **Poly/Poly** - and their **Latch** versions It also has: - a **random ↔ repeating 16-step pattern** control - a **new random sequence** button - an **auto-reset / hidden reset** behavior - **clock divider modes**: - Classic Divider - 2/3/5 Divider - Spread ## Why this matters for melody Melody in Eurorack is often built from several layers: - **pitch source**: sequencer, random voltage, S&H, Turing machine, keyboard, quantizer - **timing**: clocks, triggers, rhythmic gates - **articulation**: envelopes, accent, ratchets, note length - **structure**: variation, phrase resets, call-and-response The Decision Tree is strongest in the last three categories: - it decides **when notes happen** - it creates **branching phrase structures** - it can generate **different melodic lanes** - it can create **repeating or semi-random phrase skeletons** - it can divide one clock into multiple related time layers So, together with other melodic modules, it becomes a very powerful **melodic organizer**. --- ## Best uses for melodic patching ## 1. Branch one pitch stream into multiple melodic voices ### Patch idea Use one pitch source and let the Decision Tree decide which voice gets played. **Patch:** - clock or trigger sequencer → **input a** - same pitch CV source → several oscillators or voices - outputs **b/c/d** → different envelope generators or LPGs for 3 different voices ### Result Each incoming note event is sent to a different voice. If each voice has: - a different timbre - different octave - different quantizer scale - different filter setting then a single pitch line becomes a **distributed melody**. ### Musical effect - call-and-response - hocketing melodies - pseudo-polyphony from a single source - generative ensemble lines This is especially effective in **Mono/Mono** mode if you want one clear note event per step. --- ## 2. Create melodic branches with separate pitch processors Instead of only routing triggers to voices, route them to **different pitch-generation chains**. ### Patch idea - master clock → **a** - outputs **b/c/d** trigger: - sample & hold 1 - sample & hold 2 - sequencer advance - suboutputs **e–m** trigger: - quantizer change - transposition source - octave switch - glide envelope - accent envelope ### Result Different branches of the tree can produce: - different notes - different octave ranges - different scales - different note lengths This turns the module into a **melodic decision engine**. ### Example - **b** triggers a bass-note generator - **c** triggers a midrange melody - **d** triggers a high-answer phrase - lower outputs **e–m** trigger ornaments, grace notes, transpositions, or alternate quantizer inputs --- ## 3. Use repeat/random morphing for phrase evolution The manual says knob **n** smoothly adjusts between: - **completely random** - and a **repeating 16-step pattern** This is excellent for melody because it gives you a continuum between: - generative unpredictability - stable motif repetition ### Musical application Patch a regular clock into **a**, and let the outputs trigger: - a quantized random CV source - a step sequencer reset/advance - envelope triggers for note articulation Then use knob **n** to move between: - evolving melodic fragments - recognizable looping phrases ### Why this is powerful A lot of modular melodies fail by being either: - too random to remember - too repetitive to stay interesting The Decision Tree is designed to sit between those extremes. --- ## 4. Generate melody plus ornament layers Because there are 12 outputs arranged in a hierarchy, you can assign different outputs to different melodic functions. ### Suggested structure - **b/c/d** = main note lanes - **e/h/k** = accents - **f/i/l** = grace notes or ratchets - **g/j/m** = transposition, reset, or harmony events ### Result You can create melodies where: - some events trigger the main pitch - some add ornament notes - some trigger higher-octave doubles - some trigger secondary voices This is a good way to create **musically legible complexity** instead of pure randomness. --- ## 5. Use latch modes as a melodic switch The latch modes hold the signal until a new input appears. The manual specifically notes this makes the module good as a **random switch**. For melody, this is extremely useful. ### Patch idea Use the module to hold one selected route active, so one melodic branch stays selected until the next trigger changes it. **Patch examples:** - send one CV source into several switched destinations - hold one drone note active - keep one transposition lane enabled - select one of several sequencers for a phrase ### Melodic effect In **Latch Mono/Mono** or **Latch Poly/Mono**: - a route stays active longer - phrases feel more intentional - melodic fragments persist before changing This can create: - sustained pedal tones - held harmonic states - phrase-based melodic switching --- ## 6. Use Poly modes for chords or layered melodic bursts In: - **Poly/Mono**, one or more main outputs can fire, but each selected branch still picks one suboutput - **Poly/Poly**, multiple outputs can fire across the whole tree ### Melodic use These modes are ideal when your system has: - multiple oscillators - a chord voice - several quantized pitch sources - multiple envelope/VCA paths ### Patch idea - one clock into **a** - several outputs trigger separate envelopes opening different tuned oscillators - each oscillator is tuned to a scale degree or interval ### Result The tree produces: - dyads - triad fragments - staggered chord tones - clustered melodic bursts In latch versions, these can become sustained harmonic textures. --- ## 7. Build a self-playing melodic ecosystem The Decision Tree excels in generative patches where one timing source branches into many related events. ### Example full melodic ecosystem - clock → **a** - **b** advances sequencer 1 - **c** samples random voltage into quantizer 2 - **d** triggers bass envelope - **e/f/g** trigger: - ornament envelope - octave shift - glide - **h/i/j** trigger: - alternate scale selection - precision adder offset - sequence reset - **k/l/m** trigger: - second quantizer - harmonic voice - phrase reset ### Result A single incoming pulse creates: - main melody - variation - harmony - resets and phrase structure This is one of the strongest uses of the module in melodic music. --- ## 8. Use the clock divider modes as melodic timing sources The divider section is very valuable for melody because it creates multiple related trigger streams at different rates. ## Classic Divider mode According to the manual: - **b** = /2 - **e** = /4 - **f** = /8 - **g** = /16 And the lower outputs: - **c h i j k l m d** act like an **8-step clock sequencer**, cycling one after the other. ### Melodic uses - Use /2, /4, /8, /16 to trigger: - bass notes - melody notes - phrase changes - chord changes - Use the 8-step cycling outputs to: - step through 8 note sources - trigger 8 stored voltages - create an addressable melody from static CVs ### Especially useful for - techno arps - clocked Berlin-style sequences - phrase resets and structured melodic repetition --- ## 9. Use 2/3/5 Divider mode for polyrhythmic melody The manual specifies: - **b e f g** = /2 /4 /8 /16 - **c h i j** = /3 /6 /12 /24 - **d k l m** = /5 /10 /20 /40 This is excellent for creating **multi-timescale melodies**. ### Patch idea - /2 group → bassline triggers - /3 group → lead melody - /5 group → accent or transposition events ### Result Even with simple pitch material, the differing divisions create: - phase relationships - evolving counterpoint - non-repeating melodic interplay ### Why this works musically Pitch content can remain simple, while timing complexity creates the sense of intelligent melodic motion. This mode is especially good with: - quantized random voltages - shift-register melodies - fixed 8-step sequencers - transposition inputs --- ## 10. Use Spread mode for harmonic and overtone-based melody The manual says Spread mode creates successive divisions: - 1/2, 1/3, 1/4, 1/5, 1/6, 1/7, etc. It also explicitly suggests using it at **audio rate for polyphonic chords using one square VCO**. ### Melodic/harmonic application 1: subharmonic melody Feed an audio-rate square wave into the input and use the outputs as subharmonic tones. This can create: - chordal intervals - pseudo-organ harmonies - subharmonic melodic structures If you switch or sequence these outputs through VCAs, you can build melodies from one oscillator’s divided frequencies. ### Melodic/harmonic application 2: trigger-derived note hierarchies At clock rate, use the spread outputs to trigger different voices at increasingly slower rates: - frequent notes in one voice - slower harmony notes in another - very slow root movement in a third This gives a natural melodic hierarchy: - ornament - phrase - harmonic anchor --- ## 11. Hidden reset input for phrase control The hidden reset input uses **jack m** in a special mode, and is useful for returning to the first step in loop mode. This matters for melody because reset is how you create: - downbeats - phrase boundaries - recurring motifs - synchronized loops with other sequencers ### Use cases - reset the module every 16 or 32 beats - align it with a master sequencer - force phrase restarts at section changes - keep generative melodies from drifting too far When activated, **m** is no longer available as an output, so it’s a tradeoff: - fewer melodic branches - more phrase control Usually worth it if your patch needs stronger song structure. --- ## 12. Practical melodic patch recipes ## Patch 1: Three-lane melody router **Goal:** one pitch source, three melodic personalities - clock → **a** - quantized CV → 3 oscillators - **b/c/d** → 3 envelopes/VCAs - tune each oscillator differently: - b = root register - c = fifth above - d = octave/high voice Use **Mono/Mono** for clean hocketing. Use **Poly/Poly** for chord fragments. --- ## Patch 2: Generative melody with stable phrase memory **Goal:** random notes that become recurring motifs - trigger clock → **a** - outputs trigger a sample & hold and quantizer setup - set knob **n** midway toward repeating pattern - press **o** to regenerate until the phrase feels musical This gives a melody that sounds composed rather than chaotic. --- ## Patch 3: Ornament engine **Goal:** main notes plus decorative notes - **b/c/d** trigger main melody voice events - **e/f/g** trigger short envelope for upper-octave oscillator - **h/i/j** trigger filter ping or FM accent - **k/l/m** trigger transposition CV changes Now the tree creates melodic detail around the core line. --- ## Patch 4: Polyrhythmic melodic trio **Goal:** three related melodies with different pulse rates In **2/3/5 Divider** mode: - /2 outputs → bassline sequencer advance - /3 outputs → lead sequencer advance - /5 outputs → transposition or accent logic This produces slowly evolving melodic counterpoint. --- ## Patch 5: Subharmonic chord melody **Goal:** melody/harmony from one oscillator In **Spread** mode: - square VCO → input - outputs → separate VCAs/mixer/filter paths - use external gates to open selected divided outputs rhythmically This creates dark, organ-like melodic harmonies from one source. --- ## Strengths for melodic composition The Robaux Decision Tree is especially strong for: - **generative melodies** - **branching phrase logic** - **polyrhythmic note triggering** - **hocketing between voices** - **melody/harmony separation** - **self-playing patches with structure** - **subharmonic harmonic material in Spread mode** ## Limitations By itself, it does **not** generate pitch CV. To make it fully melodic, pair it with: - quantizers - random CV sources - sequencers - sample & hold - oscillators - switch modules - precision adders - envelope/VCA voice chains Think of it as a **melody traffic controller**, not the melody source itself. ## Best partner modules It pairs especially well with: - quantizers - Turing Machine / random CV modules - sequencers with reset/advance inputs - sample & hold - shift registers - envelope generators - logic modules - VCAs/LPGs - precision adders - multi-oscillator voice setups ## Bottom line The Decision Tree can be a powerful melodic module when used as a **rhythmic branching brain**. It shines when you want melodies that feel: - alive - structured - varied - pseudo-composed Its strongest musical role is to distribute timing and phrase decisions across several pitch paths, turning ordinary melodic sources into evolving, layered musical systems. [Generated With Eurorack Processor](https://github.com/nstarke/eurorack-processor)