# Xaoc Devices — Batumi - [Manual PDF](../../manuals/xaoc_batumi2_poti2_manual.pdf) --- [Batumi II & Poti II Manual (Xaoc Devices)](https://xaocdevices.com/manuals/xaoc_batumi2_poti2_manual.pdf) # Using Xaoc Batumi II + Poti II to create melodic material Batumi II is “just” a quad LFO on paper, but in practice it can be a **4-channel modulation brain** that easily produces **pitch movement, stepped note patterns, clock-related sequences, pseudo-counterpoint, and audio-rate melodic voices**. Poti II turns it into a more performance-friendly melodic tool by adding **per-channel attenuation, per-channel assignable ASGN waveform choice, and waveform-shape CV control**. ## What matters musically From the manual, the key features for melody-making are: - **4 channels** of oscillation/modulation - Each channel outputs simultaneously: - **Sine** - **ASGN** waveform - **Rect** - **ASGN choices**: - triangle - downward saw - upward saw - trapezoid - stepped random - smooth random - Modes: - **Free** - **Phase** - **Divide** - **Mult** - **Wide range**: - LFO rates up to **audio rate** - **1V/oct tracking** on CV inputs - **Reset/sync per channel** - In external sync mode, channels can become **clock divisions** - With **Poti II**: - attenuate pitch/rate CV - attenuate sine/asgn outputs - select **different ASGN waveform per channel** - shape CV can modulate: - sine = wavefolding - asgn = waveform switching - rect = pulse width So the module can function as: 1. **4 independent pitch CV sources** 2. **A bank of related melodic modulators** 3. **A clock-derived sequencer substitute** 4. **A harmonic oscillator bank** 5. **A canon/fugue engine** via phase offsets and synced random 6. **A melody source plus rhythmic gates** from the rect outs --- # The core idea: Batumi as a melodic source Batumi II does not quantize by itself, so for conventional tonal melody you’ll usually patch it through a: - **quantizer** - precision adder - sample & hold / track & hold - sequential switch - clock divider / trigger source - envelope + VCA + oscillator voice Typical melodic patch flow: **Batumi output → quantizer → oscillator 1V/oct** And often simultaneously: **Batumi rect out → envelope trigger / LPG / clock input** This is where Batumi becomes very musical: one channel can create the pitch contour, while its rect output or another channel controls when notes happen. --- # Best waveform choices for melody ## 1. Stepped random ASGN This is the most direct route to notes. - Patch **ASGN out** with waveform set to **step-random** - Send to a **quantizer** - Clock note events using: - that same channel’s **rect out** - or an external trigger stream - Result: evolving random melodies Musically: - In **free mode**, each random channel is independent, so you get **multiple unrelated melodic lines** - In **phase/divide/mult**, the random outputs are related, so you get **themed variations** rather than chaos ## 2. Smooth random ASGN Great for “wandering” melodies. - Smooth random into a quantizer gives gliding voltages forced into scale notes - Excellent for ambient or generative lead lines - Add sample & hold after it if you want more discrete note steps ## 3. Triangle / saw / trapezoid These are excellent for **repeating scalar patterns**. When quantized: - **Triangle** = up/down melodic motion - **Upward saw** = rising ramps with sudden reset - **Downward saw** = descending ramps - **Trapezoid** = repeated held notes with quick transitions These produce very usable sequence-like shapes once quantized. ## 4. Sine Sine into a quantizer can be surprisingly musical: - smoother revisiting of nearby notes - cyclical basslines - nice for slow transposition patterns --- # The four modes and how they help with melody ## 1. Free mode: four independent melodic sources In **free mode**, all four channels are independent. This is ideal for: - four separate quantized melodies - pitch CV + gate pattern + transposition + ornament - one module driving an entire small generative patch ### Melodic uses - **Channel A**: stepped random → quantizer → lead pitch - **Channel B**: slow triangle → precision adder transpose input - **Channel C**: rect out → clock a sample & hold or envelope - **Channel D**: smooth random → filter cutoff or second voice pitch Because Batumi tracks **1V/oct**, you can also treat a channel as an oscillator at audio rate. That opens another trick: - Use **A** as audio oscillator - Use **B/C/D** as slow melodic modulators affecting pitch, fold, filter, or FM depth ### Strong patch idea: 4-voice generative ensemble - Set all four channels to **step-random** - Quantize each to the same scale - Use each rect output as its own gate/trigger source - Send to 4 voices or to a voice allocator / switch - Adjust each channel speed independently This makes Batumi feel like a compact algorithmic sequencer. --- ## 2. Phase mode: canon, fugue, and melodic imitation In **phase mode**, channels B/C/D follow channel A’s frequency but are **phase shifted** relative to it. This is one of the most musically special features. ### Why it matters for melody In this mode, the manual says the random waves are **copies of channel A’s random sequence, delayed by phase amount**. That means you can create: - melodic imitation - canons - staggered repetitions - “same tune, different entrance point” - tightly related counterpoint This is gold for melodic composition. ### Patch: Fugue machine - Set ASGN to **step-random** - Put Batumi in **phase mode** - Quantize outputs of A/B/C/D separately - Set phase offsets on B/C/D to different values - Use similar or divided gate streams for note articulation Result: - the same underlying note stream appears at different time offsets - sounds like canon/counterpoint rather than random unrelated notes ### Patch: rotating arpeggio - Use **triangle** or **saw** as the source waveform - Quantize all channels - Set phase offsets to 90°, 180°, 270° - Each channel becomes a shifted version of the same contour Use with: - one voice and a sequential switch - or multiple oscillators for interlocking harmony ### Advanced note The manual mentions **deep phase modulation** can radically deform the waveform and even act like FM at audio rates. In melodic use, that means: - a simple source contour can morph into more complex note shapes before quantization - subtle CV on phase can create **variation without losing structure** --- ## 3. Divide mode: rhythmic melodic hierarchy In **divide mode**, B/C/D become integer subdivisions of A: - 1, 2, 3, 4, 5, 8, 16, 32 This is amazing for **tempo-locked melody layers**. ### Melodic consequence All channels are related to A, but slower. So you can patch: - **A** = fast melody - **B** = slower bassline - **C** = even slower transposition source - **D** = ultra-slow phrase-level movement ### Patch: melody + bass + phrase transpose - Channel A ASGN step-random → quantizer → lead - Channel B ASGN step-random → quantizer → bass - Channel C triangle → quantizer or precision adder transpose - Channel D rect → reset a sequencer every long cycle Because divide mode keeps things related, the whole patch feels coherent. ### Why the random mode is useful here The manual says divided random channels are **downsampled copies** of channel A’s random sequence. So if A is a fast random note stream, B/C/D become slower extractions of that same source. Musically this creates: - motif at different time scales - phrase-level unity - less “everything independent” chaos This is ideal for generative music where you want a family resemblance across voices. --- ## 4. Mult mode: embellishment, ornament, upper voices In **mult mode**, B/C/D become integer multiples of A. That means: - one base motion in A - faster related motions in the other channels ### Great melodic uses - ornament around a main line - ratcheting note motion - high-register companion lines - harmonic overtone-like pitch movement if run at audio rate ### Patch: lead with ornament voices - A = slow quantized contour for main lead - B = x2 related line - C = x3 related line - D = x4 or x5 related line - Quantize them all to the same scale Result: - upper voices move more rapidly but remain tied to the same underlying cycle ### Random in mult mode The manual says random sequences are **upsampled** relative to A, sharing values periodically. That means: - the faster channels reuse material from A - good for decorations and variations around a motif This makes mult mode very useful for: - grace-note-like melodic behavior - dense top-line texture over a slower foundation --- # Poti II: why it matters for melodic control Poti II makes Batumi much more useful as a melody tool. ## 1. Output attenuation This is extremely important because quantizers respond strongly to voltage range. With Poti II, you can reduce output amplitude for each channel: - narrow melodic range - keep notes within one octave - create voice-dependent contours Example: - Channel A wide range = lead - Channel B attenuated = bassline with small movement - Channel C tiny range = subtle transposition CV Without attenuation, some Batumi outputs may jump too widely for practical tonal melodies. ## 2. Per-channel ASGN waveform selection This is huge. You can choose different ASGN waveforms per channel: - A = stepped random - B = triangle - C = upward saw - D = smooth random Now Batumi becomes a **mixed melodic ecosystem** instead of four copies of the same waveform type. ## 3. Shape CV inputs The shape inputs are extremely playable for melody generation. Per channel, shape CV can affect one selected destination: ### Sine: wavefolding If you quantize folded sine motion, you get: - more complex repeated note patterns - symmetry breaking - pseudo-arpeggiated shapes from a simple cycle ### ASGN: waveform switching This is especially strong for melodic work. If shape CV is assigned to ASGN: - external CV can switch among waveforms - the melodic contour can jump between triangle, saw, trapezoid, random, etc. That means one melodic line can morph between: - ordered - ramping - held - random Great for arrangement and variation. ### Rect: pulse-width modulation This does not directly create pitch, but it changes gate timing/duty behavior, which affects note length and rhythmic phrasing if rect is used as a gate source. --- # Practical melodic patch recipes ## Patch 1: Simple generative melody **Goal:** one evolving tonal line - Batumi in **free mode** - Channel A ASGN set to **step-random** - A ASGN → quantizer → oscillator pitch - A RECT → envelope trigger - Envelope → VCA → oscillator Optional: - Poti II attenuate A ASGN to keep melody within a narrower range - Slow modulation into A FRQ CV for phrase speed changes Why it works: - stepped random provides note selection - rect provides timing - quantizer keeps it musical --- ## Patch 2: Canon melody **Goal:** related voices with delayed entrances - Batumi in **phase mode** - ASGN = **step-random** - A/B/C/D ASGN → separate quantizer channels or multiple quantizer inputs - Set B/C/D to different phase offsets - Use rects or external triggers to articulate voices Result: - the same sequence appears staggered in time - ideal for minimalist or contrapuntal music This is one of the best “melodic identity” uses of Batumi II. --- ## Patch 3: Bassline + lead from one source family **Goal:** coherent two-part composition - Batumi in **divide mode** - A = medium-fast stepped random to lead quantizer - B = divide by 2 or 4 to bass quantizer - C = triangle, attenuated, to transpose the bass or lead by a few scale steps - D RECT = phrase reset trigger somewhere else in the patch Result: - lead and bass feel related because they derive from shared timing/material --- ## Patch 4: Arpeggiator from continuous waves **Goal:** sequencer-like melodic loops without a sequencer - Batumi in **phase mode** or **free mode** - Use **triangle** or **saw** outputs - Quantize outputs - Clock a sample & hold from rect or external trigger if you want discrete note changes Options: - Triangle gives pendulum-style arps - Saw gives rising or falling arps - Trapezoid gives repeated held notes --- ## Patch 5: Ornament engine **Goal:** primary melody plus embellishment - Batumi in **mult mode** - A ASGN → quantizer → main pitch - B/C ASGN → quantizer → secondary voices at x2 / x3 - D sine or rect modulates filter, LPG, or wavefolder This can sound like: - one melody with ornamentation - clustered lines - baroque-style decorations - shimmer on top of a slower line --- ## Patch 6: Batumi as four VCOs for harmonic melody Because Batumi II goes to **audio rate** and supports **1V/oct**, it can be used as a bank of oscillators. ### Patch - Put channels in **phase**, **divide**, or **mult** - Feed pitch CV to the frq input(s) - Take sine or asgn outputs as audio - Mix channels ### Melodic applications - **Phase mode**: detuned/phase-shifted unison or chord-like movement - **Divide mode**: subharmonic relationships for bass and drones - **Mult mode**: harmonic interval / overtone-like melodic stacks If you send a keyboard or sequencer CV to channel A: - in divide/mult, other channels become related pitch structures - not equal-tempered harmony automatically, but very rich for experimental melody --- # Sync-based melodic strategies Batumi’s **reset/sync** behavior is very useful when you want melody tied to song tempo. ## Reset mode - incoming trigger resets phase - useful for restarting melodic contours at bar lines - good for keeping looping note curves aligned with the groove Example: - A saw → quantizer → pitch - send a bar-reset trigger to A reset - melody restarts every measure ## Sync mode - Batumi adapts to external clock period - sliders then select **division factors** - in non-free modes, only channel A can sync to external tempo This effectively lets Batumi become a **clock-relative melodic generator**. ### Patch: clocked note divisions - External clock into A sync - A/B/C/D outputs to quantizers or modulation destinations - Use the resulting divided cycle lengths as phrase lengths This is excellent for: - tempo-locked ambient melodies - self-playing Berlin-school style structures - polymetric note streams --- # How to get more “musical” melodies from Batumi ## Use a quantizer This is the big one. Batumi makes voltages; the quantizer makes notes. Good scale choices: - pentatonic for instant consonance - dorian/aeolian for melodic ambient lines - chromatic for experimental lines ## Limit voltage range with Poti II Attenuation is your friend. Smaller voltage ranges mean: - fewer leaps - more motif coherence - basslines that stay in register ## Combine with sample & hold Continuous shapes become stepped note sequences when sampled. Examples: - triangle + S&H = repeating scalar melody - smooth random + S&H = wandering but articulated line - phase-shifted channels sampled by same clock = related parallel melodies ## Use rect outputs as note triggers Every channel provides its own rect output. This can drive: - envelopes - logic - clocking - sample & hold So one Batumi channel can generate both: - note voltage - note timing ## Use channel relationships intentionally - **Free** = independence - **Phase** = imitation/canon - **Divide** = structural hierarchy - **Mult** = embellishment/density That gives you compositional roles, not just modulation modes. --- # Best pairings with other Eurorack modules These modules really shine with: - **Quantizer** - Intellijel Scales, Ornament & Crime, ADDAC quantizers, etc. - **Sample & hold / T&H** - **Precision adder** - **Sequential switch** - **Logic / clock divider** - **Envelope + VCA or LPG** - **Oscillator / voice module** - **Mixer** - **Comparator** for deriving extra gates from smooth shapes Especially powerful combinations: ## Batumi + quantizer + switch Four channels become four phrase sources; the switch chooses which phrase is active. ## Batumi + quantizer + precision adder One channel makes note motion, another makes transposition, another makes phrase-length movement. ## Batumi + chord oscillator / multi-voice synth Use multiple Batumi outputs to animate root, inversion, spread, or parallel pitch lines. --- # Performance-oriented melodic ideas ## Morph from ordered to chaotic Using Poti II per-channel ASGN selection and shape CV: - start with triangle/saw for regular melody - morph into stepped random - then into smooth random - return to structured shapes This feels like moving from composition to improvisation. ## Bring voices in by channel role - A = main melody - B = delayed imitation - C = bass - D = ornament Because all are on one module, this is easy to perform as a “mini ensemble.” ## Use resets for phrase boundaries Manual resets into channels create controlled re-alignments: - all voices land together - then drift through their own relationships - then rejoin Very musical for live generative sets. --- # A few especially strong melodic concepts from the manual ## 1. Phase-mode random as delayed copies This is probably the standout melodic feature. It enables: - canon - fugue-like structures - offset repetition - related motif streams ## 2. Divide/mult random as derived versions of one sequence This gives: - coherent family resemblance - phrase hierarchy - ornament vs foundation behavior ## 3. Audio-rate operation with 1V/oct Batumi can cross from modulation into sound generation, so a patch can use the same module for: - pitch CV generation - audio oscillation - internal self-modulation - rhythmic gating That makes it unusually rich for melodic composition. --- # Bottom line **Batumi II + Poti II can absolutely be used as a melodic composition tool, not only as an LFO.** Its strongest melodic use cases are: - **random quantized melodies** - **phase-shifted canons** - **clock-related bass/lead hierarchies** - **ornamented multi-speed related lines** - **audio-rate harmonic or quasi-harmonic voices** - **self-playing patches where one module supplies pitch, timing, and phrase structure** If I were using it specifically for melodic work, I’d think of the modes like this: - **Free** = 4 independent composers - **Phase** = 1 melody, 3 imitators - **Divide** = 1 fast idea, 3 slower structural layers - **Mult** = 1 core idea, 3 embellishers And Poti II makes the difference between “interesting modulation” and **usable melodic control**, mainly because of attenuation and per-channel waveform assignment. --- [Generated With Eurorack Processor](https://github.com/nstarke/eurorack-processor)