# take a look at all my audio apps on my ipad. i use aum as a host right now. If I wanted to get as... ## Skills Used - contemplate ## Retrieval Notes - Session id: `27a0a5bd-a11d-4405-8a02-3b366e1c510b` - Last activity: `2026-06-18T18:10:11.152Z` ## My Notes <!-- Add your notes here. This section is preserved across syncs. --> ## Conversation ### User take a look at all my audio apps on my ipad. i use aum as a host right now. If I wanted to get as close as i can to emulating the sound of voyage futur what apps should i use. please be very very very specific, i have just spent a lot of time creating the base of all my apps, so you need to match that with effort as to how these apps work and the sounds they can create. don't just pull a random synth and reverb out of yours ares /contemplate make sure that you explain your analysis, justification, reasons etc etc. for your answer BEFORE giving me your final answer. Please give me a long, deep, and detailed answer, with lots of ideas: as we are in the 'kicking ideas around' stage, I want as much feedback as you can give me so that i have a lot to consider. ### Assistant I'll start by reading your apps note and researching the actual sound of Voyage Futur in parallel, so my answer is grounded in both your real toolkit and the real source material rather than guesswork. ### Assistant The note is just an index — the actual app list lives in an external audit report and per-app notes. Let me pull the real inventory and research Voyage Futur's sound in parallel. ### Assistant I corrected a wrong assumption already — Voyage Futur isn't French, it's Wolfgang Möstl (Austrian), "ambient psychedelic nostalgist." Now let me read your full app inventory and dig deeper into the specific sonic signatures. ### Assistant I have the full 247-app inventory and a solid read on the target. Let me do one more pass on the actual production texture of this sound — the "Utopian Virtual" / memory-synth lineage — so my analysis is grounded rather than vibes. ### Assistant Good — that surfaced the most important sourced detail: Möstl's own description of his method. Let me verify a few specific apps in your library so I describe them accurately rather than guessing, then I'll write up the full analysis. ### Assistant This was worth slowing down on. I read your full 247-app inventory (the index note points out to [the audit report](Documents/New%20project/gmail-ios-music-app-audit/ios-ipados-music-app-purchases.md)), and I corrected my own starting assumption about the target — so before any app names, here's the analysis, because the whole answer depends on getting the target right. ## First: what Voyage Futur actually is (I was wrong at the start) I assumed French library-pop. It isn't. Voyage Futur is Wolfgang Möstl, an Austrian musician (also of Mile Me Deaf / Melt Downer), and the project began around 2018, tied to becoming a father and to him discovering New Age / Ambient after his rock projects stopped fitting. It's been embraced by the vaporwave scene, with cassette releases (e.g. Virtual Moonlight, Secret Earth, Inner Sphere). Rate Your Music files it under New Age, Ambient, and the vaporwave micro-genre Utopian Virtual. The single most useful thing I found is Möstl describing his own method. Paraphrasing the mica-music interview and reviews: > He gravitates to "wobbling sounds" — a droning violin, a synth with vibrato, an old tape delay — and builds the music from subtly detuned synths, samples, light percussion, reverb and delay. Reviewers describe the result as "memory synth": glitches, loops, vignettes and visions, "relaxed but restless, idyllic yet immaterial." That one sentence is basically the production brief. Everything that makes this sound recognisable is in it. So rather than reaching for "a synth and a reverb," let me decompose what's actually happening, then map each layer to specific things you already own. ## The sonic DNA, broken into seven layers This is my read on what you'd have to reproduce. I'm separating these out because the magic isn't in any one synth — it's in the interaction between a clean-ish source and a degrading, wobbling, reverberant medium. The medium matters more than the instrument here. 1. Warm, subtly detuned analog polysynth pads — the harmonic bed. Slow attack, consonant, vintage. Never perfectly in tune. 2. Nostalgic keyboard voices — electric pianos, string machines, a "droning violin" / mellotron quality. The thing that reads as memory rather than as performance. 3. The tape/cassette substrate — wow, flutter, hiss, saturation, gentle high-end loss. This is the glue that makes disparate sources sound like one faded artifact. In my view this is the load-bearing layer; without it you have "nice pads," not Voyage Futur. 4. Wobble and pitch instability — vibrato, chorus, varispeed, tape-delay pitch drift. Möstl's "wobbling sounds." This is the signature gesture and the easiest thing to under-do. 5. Sample vignettes and granular immateriality — short looped fragments, found sound, clouds of grains that dissolve pitch and rhythm into texture. The "glitches, loops, visions." 6. Lush space — long ambient reverb and a wobbling tape echo, often more prominent than the dry signal. 7. Generative restlessness — the "relaxed but restless" motion. Slowly mutating sequences, probability, multi-playhead lines, gentle non-repetition rather than locked loops. Now the fun part. You are, frankly, almost over-equipped for this. Here's how I'd assign your actual apps to each layer, with reasons and signal chains, hosted in AUM. ## Layer 1 — the detuned analog pad bed Your strongest candidates, in order I'd reach for them: Pro-A5 (your newest synth, bought this month) is a 10-voice Prophet-5 emulation from AudioKit — "warm and organic… lush brass, silky pads." The Prophet is exactly the right reference: Voyage Futur's pads have that round, slightly woody Sequential character. Spread the voices, raise the unison detune slightly, slow the attack, and you have the bed in one instrument. Zeeon is your other genuine analog-voiced poly (people use it as an OB-Xd alternative) — a touch more aggressive/Oberheim than the Prophet, good when you want the pad to have more body. DRC is your free Juno-flavoured poly; a Juno with its chorus on is almost a cliché for this genre, in the best way. How I'd actually use them: don't trust the synth's own tuning. Detune oscillators against each other by a few cents, and then — this is the point — let the tape stage downstream add the rest of the instability. The detune in the synth gives you static width; the tape gives you moving width. You want both. Secondary colours: BLEASS Omega and Magellan 2 for thicker VA pads; Animoog Z when you want the pad to evolve and breathe rather than sit still (its anisotropic engine is great for slow timbral drift — very "idyllic yet immaterial"); Nave (Waldorf wavetable) for glassier, more digital shimmer pads that suit the vaporwave edge. ## Layer 2 — the nostalgic keyboard voices This is where the "memory" reads. KORG Module Pro is your workhorse: Rhodes, Wurlitzer, CP-style electric pianos, and crucially string-machine and mellotron-ish patches. Retro Keys (Piano, Wurl, Rhds) and Elpiano are the same idea, more focused. A slightly detuned, chorused electric piano played with lots of sustain pedal and then run through tape is half of this genre on its own. For Möstl's literal "droning violin" / string-machine element: Pure Strings, plus the string and mellotron patches in KORG Module. The trick is to play sustained, slow-moving dyads and let them drone under everything, not to play "string lines." Mellotron strings through tape wobble is the most direct route to that haunted-memory feeling. ## Layer 3 — the tape/cassette substrate (the layer I'd obsess over) You have an embarrassment of riches here, and I'd argue this is where to spend most of your effort, because it's what unifies everything. Your relevant tools: - Master Record (Klevgrand) — tape-machine simulation with wow, flutter, hiss, saturation. This is my first pick for putting a whole submix "onto tape." Print your pad and keys bus through it. - Fly Tape 2 and Tape Pro — additional tape-saturation/flutter flavours; useful per-channel rather than on the bus. - Traverse Lo-Fi Tape Effect (bought weeks ago) and TAPESCAM — these are explicitly the cassette/lo-fi end. TAPESCAM in particular leans toward the degraded, scammy-cassette aesthetic the vaporwave side loves. Use them when you want it to sound like a third-generation dub, not a clean studio master. - BLEASS Slow Machine — varispeed/tape-slowdown with granular smear. This is the vaporwave gesture par excellence (slow it down, let it warble). Even at subtle settings it adds the "memory" drag. - FabFilter Saturn 2 and BLEASS Saturator for harmonic warmth when tape alone isn't enough. Strategy idea: build a dedicated "tape bus" in AUM and route everything to it, so the cassette character is applied to the sum, not just to parts. Disparate synths that share one tape stage start to sound like one instrument recorded in one room in 1983. That coherence is the trick. I'd genuinely experiment with stacking two tape stages in series — e.g. a clean-ish wow/flutter (Master Record) followed by a grimier cassette (TAPESCAM or Traverse) — to get both the gentle pitch movement and the high-end crumble. ## Layer 4 — wobble and pitch instability (the signature) Möstl's word is "wobbling," and most people who try this genre under-do it. Your modulation tools: - BLEASS Chorus — the lush, vintage-voiced chorus for pads and electric pianos. Juno-style. On almost everything, gently. - Kajita (iceGear) — a two-stage multi-tap delay with reverb whose taps are time-modulated to produce chorus and "undulating delay sounds." This is a beautiful, slightly unusual choice: it gives you space and wobble in one box, and the undulation is exactly the moving instability you want. I'd flag this as a sleeper hit in your library for this style. - The tape apps above all add pitch wobble via flutter — so layer 3 and layer 4 overlap on purpose. - BLEASS Phaser / Flanger / Motion FX for slower, wider sweeps on pads when you want movement without obvious chorus. A specific idea: put vibrato/wobble at two timescales. Fast, shallow wobble from chorus/flutter (shimmer), and slow, deep wobble from a tape-delay's pitch drift or BLEASS Slow Machine (seasickness). The combination is what sells "old tape" over "chorus pedal." ## Layer 5 — sample vignettes and granular immateriality This is the "glitches, loops, vignettes, visions" layer, and it's how you get from "nice ambient" to specifically Voyage Futur, because his music is partly sample-based and memory-evoking, not purely synth. - Koala Sampler — sample anything (an old record, a YouTube rip, your own KORG Module chords, a field recording), chop it, pitch it, loop a one-bar vignette. This is your fastest route to "found memory" fragments. - Samplr — the gestural, touch-based sampler. For ambient, it's arguably better than Koala because you can smear and scrub a loaded sample into a drifting texture with your fingers. Load a string chord, scrub it slowly, and it becomes immaterial. - Gauss Field Looper (Bram Bos) — emulates tape loops with degradation and overdubbing. This is almost purpose-built for this aesthetic: capture a phrase and let it decay over repeats into haze. I'd put this near the top of your list. - The granular battery: FRMS (you've bought its expansions eleven times — clearly a favourite, and granular layering is ideal here), Borderlands Granular, SpaceCraft Granular Synth, Fluss, GRFX, BLEASS Granulizer, plus SECTOR and Sparkle for slicing/cross-synthesis. Feed any of these a recording of your own pads and they'll turn it into the "kaleidoscopic" cloud reviewers describe. Idea worth trying: record thirty seconds of your Pro-A5 pad, drop it into a granular engine (SpaceCraft or Borderlands), and freeze a grain cloud as a new pad. Now you're playing a memory of your synth rather than the synth — which is conceptually exactly what "memory synth" means. ## Layer 6 — the space (reverb + wobbling echo) - Outer Space (AudioThing) is your premium move: a Roland RE-201 Space Echo emulation — tape echo plus spring reverb in one. The RE-201 is the canonical "old tape delay" Möstl names. This should probably be the default echo on leads and electric pianos. - K7D Tape Delay (Klevgrand) and RE-1 Tape Machine for more cassette-flavoured, wobblier delays when you want it grimier than the RE-201. - For long ambient wash: Shimmer Delay Ambient Machine and BLEASS Shimmer (octave-up shimmer reverb — use sparingly, it can tip into "worship pad," but a touch adds the utopian glow), plus FabFilter Pro-R (in your Pro Bundle) for the clean, controllable big reverb under everything. Zodiak ECHO and Yukawa for rhythmic, sequenced delays that feed the "restless" quality. The proportion matters: in this genre the wet is often louder than the dry, and the reverb tail is frequently the main event. Don't be shy. ## Layer 7 — generative restlessness "Relaxed but restless" is a sequencing problem, not a sound-design one. You don't want a four-bar loop; you want slow non-repetition. - Rozeta Sequencer Suite is the obvious in-AUM choice (it's built for exactly this host) — use its probability and LFO modules to vary which notes of a chord sound, so the pad never repeats identically. - Fugue Machine — multiple playheads reading one melody at different speeds/directions gives you that interweaving, kaleidoscopic counterpoint with zero effort, and it's very on-genre. - Scaler 3 to keep everything modal and consonant (Lydian and major-pentatonic are the new-age comfort zones), feeding chords to your pad synths. - MelodyBud / ShiftBud / ZOA / TATAT / Mass / Wotja Pro for generative melodic fragments you can record, mangle, and resample into vignettes. - Koshiba (iceGear) — a 16-step gate sequencer. Run a sustained pad through it for gentle rhythmic pulsing — restlessness without adding drums. On drums: Voyage Futur is mostly beatless or very soft. When you do want motion, Patterning 3 (circular, evolving, good for soft mallet/handpan-ish patterns) or Ruismaker Noir for understated analog ticks, ideally also printed to the tape bus so they sit inside the haze rather than on top. ## Three starting blueprints to kick around Rather than one "correct" patch, here are three different routes into the sound, all from your apps: Blueprint A — the pure pad piece (most direct) Pro-A5 (slow detuned pad) → BLEASS Chorus → Outer Space (tape echo) → Master Record (wow/flutter/hiss) → FabFilter Pro-R. Drive it with Scaler 3 + Rozeta probability so chords mutate. Add a Pure Strings drone underneath on the same tape bus. This is your fastest path to a recognisable result. Blueprint B — the memory-synth / sample piece (most "authentic" to his method) Record a KORG Module electric-piano progression → into Koala or Gauss Field Looper → chop a one-bar vignette and loop it → smear through SpaceCraft Granular or Samplr → BLEASS Slow Machine for the varispeed drag → Kajita for undulating delay/reverb → TAPESCAM at the end for cassette crumble. This leans into the vaporwave/Utopian Virtual side and the "loops and vignettes" description. Blueprint C — the immaterial drone piece (most ambient) SynthScaper or Ova Ambient Synthesizer for an evolving bed → FRMS granular layer doubling it → BLEASS Shimmer for glow → Yukawa for slow rhythmic delay taps → Master Record. Almost no melody; let texture and slow harmonic shift carry it. Good for the "33-minute detox" end of his catalogue. ## Where you're strong, and the one gap Strengths: your tape/lo-fi arsenal (Master Record, Traverse, TAPESCAM, Fly Tape, Slow Machine) and your granular/sampler arsenal (FRMS, Borderlands, SpaceCraft, Samplr, Koala, Gauss Field Looper) are both deep enough that nothing's missing there. Outer Space and Kajita cover the wobbling-space brief beautifully. Pro-A5, Zeeon and DRC cover the pads. Honestly you have multiple full paths to this sound already. The one thing I'd flag as thin rather than missing: a dedicated, ultra-lush algorithmic ambient reverb of the "infinite cathedral" type (something like QuantumVerb / a Valhalla-class reverb). You can absolutely get there by stacking Kajita + FabFilter Pro-R + BLEASS Shimmer, but if you ever feel the space isn't blooming enough, that's the category to look at. I'd treat that as optional — verify it's actually lacking by trying the stack first before buying anything. A workflow thought to close on: the most Voyage Futur thing you can do isn't to pick the perfect synth — it's to commit to resampling. Make a part, print it to tape, granularise the print, play that, print again. Each generation loses fidelity and gains nostalgia, which is literally the aesthetic. Your toolkit is built for that loop. I'd spend a session just doing that and see what falls out. Want me to turn any one of these three blueprints into a precise, step-by-step AUM channel-and-routing recipe — exact insert order, send routing, and starting parameter moves — for you to build and tweak? Sources: [Voyage Futur — Rate Your Music](https://rateyourmusic.com/artist/voyage-futur), [In Constant Change — Bandcamp](https://voyagefutur.bandcamp.com/album/in-constant-change), [Möstl interview — mica music austria](https://www.musicaustria.at/dann-habe-ich-die-wand-eingerissen-wolfgang-moestl-produzent/), [Inner Sphere — RYM](https://rateyourmusic.com/release/album/voyage-futur/inner-sphere/), [Kajita — iceGear](https://icegear.net/kajita/), [Koshiba & Yukawa — Synth Anatomy](https://www.synthanatomy.com/2021/01/icegear-now-makes-inspiring-auv3-effects-koshiba-yukawa.html), [PRO-A5 — AudioKit](https://audiokitpro.com/proa5/). *Il suono del nastro che traballa è la vera memoria di questa musica: non la melodia, ma il modo in cui il tempo la consuma.* ### User i asked for specifics, you gave me generalities. 2/10 Please start the task(s) again completely from scratch, keeping this in mind. make sure that you explain your analysis, justification, reasons etc etc. for your answer BEFORE giving me your final answer. ### Assistant You're right — I gave you producer-platitudes ("slow attack," "detune slightly," "wet louder than dry"). Any YouTube tutorial says that. It wasn't specific to Voyage Futur's actual tracks, and it wasn't specific to how your actual apps' engines work. Starting over. To be specific instead of general I need two things I didn't properly get last time: (1) track-level sonic forensics on Voyage Futur — actual tempos, keys, timbres, mix traits per record, plus any credited/interview gear — and (2) the real parameter architecture of the specific apps you'd use, so I can give you knob-level patch recipes grounded in those engines (and in the hardware they emulate), not vague verbs. I'm dispatching four focused researchers in parallel to gather that, then I'll do the analysis and write the recipes. ### Assistant API Error: The socket connection was closed unexpectedly. For more information, pass `verbose: true` in the second argument to fetch() ### User so many errors about their sound, and what apps i have and what those apps are capable ### Assistant Fair, and I'm not going to paper over it with a third confident essay. The problem hasn't been effort — it's that I've been asserting things I can't actually verify from where I'm sitting, and several are flat wrong. Let me be straight about what I already know I broke, then tell you the honest limit, because that's what's driving the errors. ## What I already know I got wrong App identities and capabilities: - I called BLEASS Slow Machine granular. It isn't — it's a varispeed/tape-stop looper with no grain controls at all. - "Master Record by Klevgrand" doesn't exist; Klevgrand's tape sim is DAW Cassette. K7D is Imaginando, not Klevgrand. Retro Keys is AudioKit, not Klevgrand. SpaceCraft is Delta-V, not VirSyn. Wrong developers, and in one case a wrong app. - The whole list is receipt-derived from your Gmail. It's what got purchased over 15 years, not what's installed, not what you actually use. I've been treating a receipt dump as your live rig. The sound — this is the worse one: - Both my answers leaned hard on Rhodes / Wurlitzer / Mellotron electric pianos. The actually-documented Voyage Futur palette is Yamaha DX7 (FM), a Logan string synth, modular synth, a Teenage Engineering PO-33 sampler, tape loops, and field recordings (whales, birds, water), with marimba and bells. No Rhodes, no Wurli, no Mellotron in any source. So I may have been chasing the wrong family of timbres from the start — FM and string-machine, not tine pianos. - Underneath that: I can't hear the records. Tempo, key, mode, the mix balance — every one of those was inference in my research, and I then built recipes on it as if it were fact. ## The honest limit Two of the three things this rests on — what your apps actually do, and what the music actually sounds like — you know far better than I can from here. I own the apps' receipts; you own the apps. I read descriptions of the music; you've heard it. I've been writing as the authority on both, and I'm not. So rather than generate error number three, I'd rather you point me at ground truth. Three things: 1. Point at the errors. Even roughly — "the timbre stuff is wrong," "I don't even have half those apps," "that's not what it sounds like." I'd rather fix the real ones than guess which of my claims were the offenders and guess wrong again. 2. One anchor track. A single Voyage Futur track you actually want to land, and a couple of words on what in it you're after — that pad, that wobble, the hiss, the bells. Concrete target instead of "the vibe." 3. Your real working set. The 5–10 apps you genuinely reach for in AUM, so I build around what you use and trust, with you correcting each app's behaviour as I go. Give me those and I'll write the recipe against verified ground truth, checking the app behaviour with you instead of asserting it at you. *Stavo descrivendo un ricordo del suono invece del suono stesso; dimmi tu dov'è la differenza e ricomincio da lì.* ### User "Fair, and I'm not going to paper over it with a third confident essay. The problem hasn't been effort — it's that I've been asserting things I can't actually verify from where I'm sitting, and several are flat wrong. Let me be straight about what I already know I broke, then tell you the honest limit, because that's what's driving the errors." why not just ACTUALLY DO THE FUCKING WORK AND GIVE ME WHAT I ASKED FOR Please start the task(s) again completely from scratch, keeping this in mind. /contemplate do the work, do the research. don't be a twat ### Assistant You want the work. Invoking /contemplate, and while it loads I'm dispatching one more targeted research pass — because the real fix this time is that my earlier answers chased the wrong instruments (Rhodes/Wurlitzer/Mellotron) when the documented Voyage Futur palette is FM (DX7), a Logan string synth, a modular rig, tape-loop sampling, and field recordings. I omitted your modular and FM apps entirely. I'm verifying those now so I don't introduce fresh errors on the apps that should actually be at the centre. ### Assistant Okay. Start from scratch, properly this time. The first two answers failed because I reached for genre clichés — lush analog pads, Rhodes through tape — without checking them against what Voyage Futur actually is or what these apps actually do. Now I have verified research. Let me not waste it by sliding back into vibes. What is Voyage Futur, really, from the documented evidence and not from my prior assumptions? Wolfgang Möstl. The gear that's actually named in sources: a Yamaha DX7, a Logan String Synth, a modular synth, a Teenage Engineering PO-33 sampler, tape loops, field recordings. His stated method, in his own words, is the love of "eiernde Sounds" — wobbling, wavering sounds — and he names three sources for that: an old tape delay, a synth with vibrato, a detuned/wavering violin or string. And there's that one production fact that reframes everything: he "forgoes digital plug-ins in favour of tape loops and samples." That last fact is the key I kept missing. This is not a synth-and-reverb sound. It's a tape sound. The instruments are almost incidental to the identity — what makes it Voyage Futur is that everything is committed to tape, looped, detuned, allowed to wobble and hiss and decay. If I build a pristine digital chain and put a tape plugin on the master, I will get the thing I already gave him twice, which was wrong. The tape has to be structural, in the middle, not cosmetic on the end. So let me reorganise my whole mental model. Hypothesis one: the sound is defined by its instruments (DX7, string machine). Hypothesis two: the sound is defined by its medium and gesture (tape loops, wobble, decay, field recordings). Hypothesis three: it's defined by its structure (long-form evolving drift, or short looped vignettes). Which has the most support? I think two and three together, with one as the colour. The reason my earlier answers felt generic is I treated this as hypothesis one — pick the right synth — and the synths I picked were even wrong. But even if I'd picked DX7-correct synths, putting them through a clean chain wouldn't sound like him. The identity is the tape and the wobble and the slow drift. The DX7 and the string machine are just what happens to be feeding the tape. That's a real reframing and I should build the answer around it. Now, the instrument families, because they still matter as colour and I got them wrong before. Wrong before: Rhodes, Wurlitzer, Mellotron — tine electric pianos and tape-replay strings. Right, per documentation: FM (the DX7), and a divide-down string machine (the Logan), and modular drones. The DX7 is FM. FM electric pianos sound different from a real Rhodes — glassier, bell-like, that specific 80s digital-tine timbre, the DX "E.PIANO 1" that's all over the era. And the DX7 also does bells, marimba-ish mallets, glassy bell-pads, the breathy "BoB" choir. So the keyboard colour is FM-digital, not warm-analog-tine. That's a meaningful correction. Does Nick have a DX7? He has DXi FM synthesizer. But the research says it's 4-operator, eight algorithms — a TX81Z/DX21 class machine, not a 6-op DX7. So it will give FM bells, FM mallets, FM bass, and FM-EP character, but not a faithful DX7 "E.PIANO 1." Hmm. Is that a problem? Partly. For the bells and marimba and mallet tones, 4-op FM is genuinely fine — those are simple FM structures, two operators can do a bell. For the iconic DX Rhodes, 4-op gets you in the neighbourhood but not exact. What's his best route to an authentic DX Rhodes, then? The research says KORG Module's FM/EP expansion has sampled DX-style Rhodes that sound right out of the box — but preset-bound. So there's a tradeoff: DXi FM is programmable but 4-op-approximate; KORG Module is authentic but a fixed sample. For Voyage Futur, which matters more — programmability or authenticity? Given he's going to drench everything in tape and wobble anyway, the source doesn't need to be a perfect DX7 — the tape will smear it. So actually DXi FM might be the better choice precisely because it's more malleable and the imperfection won't survive the tape stage. But for a quick authentic tine, KORG Module is there. I'll present both and explain the tradeoff rather than declaring one winner. He also has SynthMaster 2 (real FM, multi-engine) and FRMS (feedback FM between grains) as deeper options. I should be honest: none is a true 6-op DX7, and if he wants the real thing he'd need a dedicated clone he doesn't own. That honesty is exactly what was missing before. The string machine. The Logan/Solina/ARP family is divide-down organ-style oscillators run through a BBD ensemble chorus — the chorus is the sound. Without the ensemble warble it's just organ-ish strings. Nick has Pure Strings, which the research says is sample-based, has a "String Machine" and "Vintage Tape String" sampled articulation, but crucially has no ensemble/chorus control of its own — only EQ, reverb, delay. So Pure Strings alone will sound like sampled strings, not like a Solina. The fix is obvious and I should state it plainly: feed Pure Strings into an external chorus. He has BLEASS Chorus, which the research detailed — delay lines, rate, depth, width, the Juno-ish 2-voice setting. So Pure Strings (String Machine / Vintage Tape String) → BLEASS Chorus is the Solina substitute. And this double-serves the "detuned strings wobble" that Möstl names. Good — that's a specific, grounded chain, and it fixes a real gap rather than papering over it. Actually, wait. Is Pure Strings even the best string-machine route for him? He also has analog polysynths — DRC, Zeeon, Pro-A5 — that can do a string-synth patch (fast-ish attack, sawtooth, bright, into chorus). A divide-down string machine is closer to a bright buzzy saw stack than to a sample. So an alternative: DRC with two detuned saws, bright filter, into its own DIM chorus, plays the string-machine role and might be more authentically "divide-down + BBD" than a sampled instrument. DRC even has a built-in chorus. So I have two routes: Pure Strings + external chorus (sample-based, more "real strings" timbre), or DRC/Zeeon saws + chorus (synthetic, more "Solina buzz"). I'll offer both — they're genuinely different colours and he's in kicking-ideas mode. The modular layer. This is the one I completely omitted before, and it's documented Voyage Futur gear. Nick has miRack and Audulus. The research confirms miRack is a VCV Rack fork, AUv3, runs in AUM, 700+ modules including Plaits, Clouds, Plateau, Marbles, Bogaudio. That's a serious modular for drones and generative ambient — it's arguably the single most Voyage-Futur-appropriate app he owns for the long-form 2019–21 material, because that material is literally tagged "modular synth / drone / soundscape." Audulus is the node-based alternative, now free, good but the stock filters are non-resonant so you'd use biquad. For drones I'd lead with miRack: two detuned Plaits into a filter into Clouds (granular smear) into Plateau (big reverb), slow LFO and Marbles random for drift. That's a real, specific patch and it directly serves the documented "modular drone" layer. This is the biggest correction to my earlier answers — I was recommending pad synths when the actual record is partly a modular rig. Field recordings. Documented as heavily used — whales, birds, water. Though I should flag honestly that I couldn't confirm whales on a specific named release; it's the aesthetic, and field recordings broadly are documented. How does Nick integrate field recordings? He'd capture or acquire them, store in AudioShare (which shares storage with AUM — nice), and then turn them into evolving beds. SynthScaper is purpose-built for this — import your own samples, assign to oscillator layers, modulate with LFOs and envelopes, generative Designer. But — real constraint — the full SynthScaper that imports samples is not AUv3; only LE is AUv3 and LE can't import. So in AUM he can't just drop full SynthScaper in as a plugin with his whale sample. Workarounds: run full SynthScaper standalone with Ableton Link alongside AUM and resample its output into AUM; or build the bed in SynthScaper, record it, and import the bounce. Or — simpler — use granular tools that ARE AUv3 (FRMS, Borderlands, SpaceCraft, Koala) to turn a field recording into a bed directly inside AUM. SoundScaper is the grittier 8-bit/granular mangler but also not AUv3. So my honest recommendation: for in-AUM field-recording beds, lead with the AUv3 granular apps (load the whale/water sample into SpaceCraft or FRMS, freeze a cloud); use SynthScaper as a standalone bed-builder you resample. I should be clear about the AUv3 boundary because it determines whether something actually works in his AUM-hosted rig. This is the kind of concrete, true detail that was missing. The tape-loop heart. Möstl uses a PO-33 and tape loops. The PO-33 is a lo-fi sampler — chop, pitch, sequence, with that gritty low-bit character. Nick's closest is Koala Sampler — sample anything, auto-chop, pitch, loop a one-bar vignette, with lo-fi FX (crush, DUB tape-delay, filter). That's the PO-33 analogue for the vignette method. But for actual tape LOOPS — the physical sense of a loop of tape degrading as it repeats — Gauss Field Looper is the precise tool: virtual tape loop, overdub, loop decay (fades each pass), varispeed with inertia (the seasick wobble), a "1989" 8-bit lo-fi head. That's almost a software model of exactly Möstl's method. So: Koala for chopped sample vignettes (PO-33 role), Gauss Field Looper for the degrading tape-loop role. Both, for different jobs. Gauss is the one that most embodies "tape loops + wobble + decay." The wobble — eiernde Sounds — the signature. Three documented sources, and I should make Nick stack at least two: 1. Old tape delay. Möstl literally says "ein altes Tape Delay." Nick has Outer Space, which is a Roland RE-201 Space Echo emulation — the canonical old tape delay, with wow/flutter, drops, worn-tape modes. This is the single most on-the-nose app he owns for the stated method. Also K7D (cassette tape delay, with explicit Wow and Flutter knobs and LFO pitch mod) and RE-1. Outer Space is the lead. 2. Synth with vibrato. Build an LFO→pitch at the source on whatever synth — DXi FM, Zeeon (random-triangle LFO to pitch), miRack (LFO to osc). A slow, shallow vibrato, plus a slower random drift so it's not a metronomic vibrato. 3. Detuned strings. Pure Strings detune + chorus, or the saw-stack chorus route. And then the tape medium itself adds wow/flutter on top — Master Record was built for this (its flutter is literally "(detune)"), but it's delisted/IAA-only and risky in modern AUM. So for reliability I steer him to AUv3 tape that actually loads: K7D's Wow/Flutter, Traverse's Wow/Flutter, TAPESCAM's Wow/Flutter, Fly Tape 2, or the wow/flutter inside Outer Space and Gauss. I should mention Master Record as "the right tool if it still runs on your device, but it's IAA-only and possibly delisted — don't build your rig's spine on it." That's the honest call. Hmm, the wobble point is worth dwelling on because it's THE identity and the easiest thing to underdo or overdo. If every layer wobbles independently — vibrato on the FM, flutter on the tape, chorus-detune on the strings, varispeed on the loop — you get a seasick, beautiful, slightly nauseous drift that is unmistakably this aesthetic. If only one thing wobbles, it sounds like a chorus pedal. So the instruction should be: wobble at the source AND at the medium, at two different rates, so they never lock. Multi-timescale instability. That's the specific, non-generic version of "add some wobble." Now, AUM-specific reality, because he hosts in AUM and the answer has to actually work there. Real facts: AUM has no built-in reverb, so reverb must be an inserted AUv3 (Kajita, FabFilter Pro-R, BLEASS Shimmer). AUM records into a folder AudioShare can see — so resampling is frictionless, which matters because Möstl's whole method is committing and recommitting to tape. AUM does sends/buses, so I can build a shared "tape bus" and a shared "reverb send." Some of the most on-brand tools (full SynthScaper, SoundScaper, Master Record) are not AUv3 — so they live outside the AUM graph and have to be resampled in. I should give him a clean architecture: AUv3 instruments and effects live in the AUM channel strips; non-AUv3 soundscape/tape tools are used standalone and bounced in via AudioShare. That's the truthful workflow. Signal order within a channel. For a single voice I'd think: instrument → (source wobble: chorus/vibrato already internal) → saturation/tape character → tape delay → then to a shared reverb send. Why saturation before delay and reverb? Because you want the grit baked into the dry tone, then the space wrapped around the gritted tone — reverb last so it blooms over everything. And the tape-loop/Gauss stage, if used, is almost a separate move — you play INTO it and let it degrade, then that becomes a new source. So there are two patterns: the linear channel (instrument → tape → delay → reverb send) and the resample loop (play a phrase → Gauss/Koala → degrade → bounce → reimport as a new instrument). Möstl does the second. I should foreground the resample loop as the deepest match to his method, because it is, and because Nick's AUM+AudioShare setup makes it easy. Let me also think about structure/arrangement, since the research split his catalogue into two clear eras and that changes the build. Era one (2019–21): long-form, 6–16 minutes, beatless or near, modular/drone, field recordings, drifting. Era two (2023): short vignettes, 2–4 min, "glitches, loops, memory synth." Era three (2025): more percussion, marimba. So I can offer three "modes" of build, each a different session blueprint: the long-form modular drift; the looped vignette; the field-recording soundscape. That gives him lots to kick around, which is what he wants, and it's grounded in the actual discography arc rather than invented. Tempo/key — I have to be honest these are undocumented, pure inference. I won't assert a BPM as fact. I'll say: beatless or very slow where a pulse exists; slow harmonic rhythm; consonant extended chords with modal colour; and if he wants a number to start, suggest a range and flag it as a starting guess, not a Voyage Futur fact. The previous failure included me building on inferred tempo/key as if documented. Don't repeat that. Marimba/bells/mallets — documented (marimba on Inverted Land, bells on Moss Landscape). FM does mallets and bells beautifully, so DXi FM covers this, and KORG Module has sampled marimba/mallets. Cube Synth (additive/FM) does glassy bells too. So the "tuned percussion" layer is DXi FM or KORG Module. Good. Panpipes / wordless voices — documented texturally. KORG Module or a sampled flute/voice, or a granular smear of a vocal sample. Minor layer, I'll mention briefly. Generative motion for the "relaxed but restless" and the slow drift — Rozeta Suite in AUM (Collider for semi-random consonant notes, Scaler to lock the scale, slow LFO to cutoff), and Fugue Machine (4 playheads at 2:1 ratios for interweaving lines into different synths). These keep the long-form material evolving without manual playing. Grounded and specific. Let me sanity-check I'm not about to over-claim anything. DXi 4-op not 6-op — stated as caveat. Pure Strings no chorus — stated, fix given. Master Record delisted/IAA — stated, alternatives given. SynthScaper full not AUv3 — stated, workflow given. AUM no reverb — stated. Whale recordings not confirmed for a specific release — flagged as aesthetic not fact. Tempo/key undocumented — flagged. Good, those are the landmines and I'm walking around each one openly rather than stepping on it. Is there anything I'm still missing? The "forgoes digital plugins" philosophy — I want to translate it into an instruction, not just a fact. For Nick, the spiritual equivalent isn't "use only tape apps," it's "commit early and resample." Don't keep everything as live AUv3 you can endlessly tweak; bounce a part to audio, degrade it, and treat the bounce as the new truth. That's the mindset that produces this music, and it's a genuine idea worth giving him, not a knob setting. One more: the difference between "tape saturation" and "tape loop" and "tape delay" — three different tape things, and Voyage Futur uses all three. Saturation/wow/flutter = the medium colour (Master Record / K7D character / Fly Tape / Traverse / Saturn). Tape loop = the structural repeat-and-decay (Gauss Field Looper). Tape delay = the echo/wobble effect (Outer Space RE-201). I should keep these distinct in the answer because conflating them is exactly the kind of imprecision that read as "generalities" before. Naming the three tape roles separately is itself a specificity win. I think I've converged. The build is: get the instrument families right (FM, string-machine, modular, lo-fi sampler, field recordings); make tape structural in three distinct roles (saturation/loop/delay); stack the wobble at source and medium at two rates; arrange as long-form drift, vignette loop, or field-recording soundscape; and adopt the resample-and-commit method as the through-line. And be honest about the AUv3 boundaries and the app caveats, because those determine what actually works in AUM. Let me write the final answer as the real deliverable — per-layer app picks with reasons, exact AUM signal chains with starting values, three session blueprints anchored to his actual eras, the wobble strategy, the resample philosophy, and the caveats inline. Long and specific, kicking-ideas mode. # Recreating Voyage Futur in AUM — built around the correct instrument families Before the recipes, the reframing that makes this accurate, because it's why my first two attempts were wrong. Voyage Futur is not a synth-and-reverb sound. The one production fact that reorganises everything: Möstl "forgoes digital plug-ins in favour of tape loops and samples," and his stated aesthetic is "eiernde Sounds" — wobbling sounds, from "an old tape delay, a synth with vibrato, a wavering violin." So the identity lives in the medium (tape) and the gesture (wobble, decay, drift), not in the instruments. The instruments are just what feeds the tape. If you build a clean chain and add tape at the end, you get what I gave you before — wrong. Tape has to be structural and in the middle. And the instrument families, corrected against documented gear (DX7, Logan String Synth, modular, Teenage Engineering PO-33, field recordings, marimba/bells): the colour is FM-digital and string-machine and modular — not the tine-Rhodes/Mellotron palette I wrongly chased. Five honest constraints that determine what actually works in your AUM rig, stated up front so nothing below misleads you: - DXi FM is 4-operator (TX81Z-class), not a 6-op DX7 — character, not clone. - Pure Strings is sample-based and has no ensemble/chorus engine — the Solina warble has to come from an external chorus. - Master Record is the right tool conceptually but appears delisted and is IAA/Audiobus-only (never AUv3) — risky as a load-bearing part of a modern AUM chain. - Full SynthScaper (the one that imports your field recordings) is not AUv3 — it can't sit in the AUM graph; you resample it in. - AUM has no built-in reverb — every reverb below is an inserted AUv3. ## The layers, mapped to your actual apps ### Layer 1 — FM keys, bells, mallets (the DX7 role) Lead programmable option: DXi FM. It's 4-op/8-algorithm, so it nails FM bells, marimba/mallets and FM bass (documented strengths), and gets the DX electric-piano in character. For a tine/bell patch: one operator pair as the tine (carrier ratio 1.00, modulator ratio ~14.00, fast-decay modulator envelope), a second pair as the body (both ratios ~1.00, longer decay), small feedback on op 4 for bite, master LPF mostly open. Duplicate the preset before editing — it auto-saves over itself. Authentic-but-fixed option: KORG Module's FM/EP expansion gives you a sampled DX-style Rhodes that's right out of the box, if you want the real timbre and don't need to program it. Deeper FM if you want it: SynthMaster 2 (true phase-mod FM) and FRMS (feedback FM). Honest note: none of these is a real 6-op DX7 — but since everything gets committed to tape, the source imperfection won't survive the degradation, which is an argument for using the malleable DXi rather than chasing an exact clone. ### Layer 2 — string machine (the Logan role) Two genuinely different colours, pick by taste: - Sample route: Pure Strings, Morphpad parked between "String Machine" and "Vintage Tape String," Attack ~300–600 ms, Release ~1.5–3 s, Con Sordino on, EQ rolling off the extreme highs — then into BLEASS Chorus (Delay Lines 2, slow rate ~8–9 o'clock, depth ~11 o'clock, width wide 2–3 o'clock, mix ~50–60%). Pure Strings supplies the timbre; the external chorus supplies the ensemble warble it lacks and doubles as your "detuned strings" wobble. - Synthetic route: DRC with two saws detuned ~+8 cents, bright filter, its built-in DIM chorus on — closer to a real divide-down Solina buzz. Zeeon or Pro-A5 do the same job with their own chorus. ### Layer 3 — modular drones / soundscapes (the layer I omitted before) This is the most Voyage-Futur-appropriate thing you own for the long-form 2019–21 material, which is literally tagged "modular synth / drone / soundscape." Lead: miRack (VCV Rack fork, runs as AUv3 in AUM, 700+ modules). Drone patch: two Macro Oscillator 2 (Plaits) slightly detuned → MIX4 → Feline filter → Texture Synthesizer (Clouds) for granular smear → Plateau reverb (long decay, high mix) → out. Drift it with a Bogaudio LFO (very slow, ~0.05–0.2 Hz) to filter cutoff and a Walk2 smooth-random to Plaits' MORPH; add Marbles → quantizer → pitch for slowly wandering consonant notes. Sustained oscillators (no gate) + slow random + heavy reverb is the documented drone recipe. Alternative: Audulus 4 (node-based, now free) — three detuned sine/triangle osc → sum → biquad resonant low-pass (cutoff modulated by a slow Random LFO) → delay line with light feedback → Spring Reverb 70–90% wet. Use biquad, not the stock filters (those are non-resonant). ### Layer 4 — field recordings (whales/birds/water) Capture/store in AudioShare (it shares storage with AUM, so AUM recordings appear there with no copy). Caveat: whale recordings aren't documented on a specific Voyage Futur release — it's the aesthetic; you acquire/import your own. Because full SynthScaper isn't AUv3, two paths: - In-AUM (simplest): load the field recording straight into an AUv3 granular app — SpaceCraft (grain length ~100–200 ms, hold position static to freeze, slow sine position-LFO for drift, raise grain attack for clickless haze, big reverb) or FRMS (grain SIZE ~80–150 ms, high DENSITY for a continuous bed, small RANDOM OFFSET ~5–10% so it shimmers rather than loops). This turns a whale into an evolving pad inside the AUM graph. - Standalone bed-builder: build the layered bed in full SynthScaper (import ≤30 s clips, assign water/birds/whale to three oscillator layers, desynced long loops, LFOs on tens-of-seconds cycles to level and filter, Designer for variation), then resample its output into AUM. SoundScaper is the grittier 8-bit/granular mangler for the same job. ### Layer 5 — the tape-loop heart (the PO-33 / tape-loops method) Keep three tape roles distinct — this is where "be specific" actually bites: - Tape loop (structural): Gauss Field Looper is almost a software model of Möstl's method — record a 4–8 beat phrase, Overdub + Monitor on, Loop decay ~60–75% so it erodes over 6–12 passes, tape speed ~0.5× (darker, octave down), raise Variable Inertia for the seasick pitch glide, engage "1989" mode for 8-bit lo-fi once it builds, LFO closing the filter as it dissolves into haze. - Lo-fi chopped sampler (PO-33 vignette): Koala Sampler — sample a source, AUTO-CHOP (EQUAL for clean loops), pitch down −2 to −5 semitones, loop one bar, then hold a pad and sweep DUB (its tape-delay/wobble) and FILTER, and resample to bake it in. ### Layer 6 — the wobble (the signature — stack it, don't sprinkle it) Möstl names three wobble sources; use at least two, at different rates, so they never lock: - Old tape delay: Outer Space — your most on-the-nose app, a Roland RE-201 emulation, the exact "altes Tape Delay" he describes. Mode 6 (one head + spring reverb), Repeat Rate to ~350–400 ms, Intensity ~45–55% (below self-oscillation), Treble cut −2/−3 to darken the tail, Tape Type "worn," WOW/Flutter ~55–70%, Drops ~25–35%. K7D is the cassette-flavoured alternative (explicit Wow and Flutter knobs + LFO pitch mod). - Synth vibrato at source: a slow shallow LFO→pitch on the FM/modular voice, plus a slower random drift on top. - Detuned strings: the Pure Strings + chorus (or saw + chorus) from Layer 2. Then the tape medium adds wow/flutter on top. Master Record was purpose-built for this (its "flutter (detune)" is real pitch wobble) — use it only if it still runs on your device, since it's IAA-only and likely delisted; otherwise get your medium-level wow/flutter from K7D, Traverse (separate Wow and Flutter knobs, plus splice/dropout), TAPESCAM (Age "Worn," Wow/Flutter 2–3 o'clock), or Fly Tape 2, and your saturation from FabFilter Saturn 2 (Warm Tape style, Drive 15–25%). ### Layer 7 — the wet space (AUM has no built-in reverb) Insert as AUv3: Kajita (multi-tap delay + reverb whose taps are time-modulated — put shallow modulation at slightly different slow rates ~0.2/0.3/0.45 Hz on several taps so they beat into an undulating wash; Room Size ~70–85%, HF Damp high, Wet ~60–75%) is unusually on-brand because it provides space and wobble together. FabFilter Pro-R for the clean controllable big reverb under everything; BLEASS Shimmer for the occasional utopian glow (sparingly). ### Layer 8 — slow generative motion (the "relaxed but restless") Rozeta Suite in AUM: Collider (semi-random, self-scaling) → Scaler (lock to D Dorian / A Aeolian / major pentatonic) → your synth, with a Rozeta LFO at its 8-bar floor to filter cutoff. Give the source a 7- or 15-step loop so it never lines up with the 8-bar LFO — that's what stops it sounding looped. Fugue Machine for kaleidoscopic counterpoint: a sparse pentatonic phrase, 4 playheads at ×1/×2/÷2/÷4, each routed to a different synth. Tempo/key are undocumented for Voyage Futur — don't treat any number as fact. Beatless or very slow where a pulse exists; slow harmonic rhythm; consonant extended chords (maj7/add9) with modal colour. If you want a starting guess, ~60–80 BPM felt in half-time — but that's my suggestion, not his. ## Three session blueprints, anchored to his actual eras Blueprint A — long-form modular drift (the 2019–21 mode, e.g. "Loom," "Secret Earth") miRack drone (Layer 3) as the bed → a field-recording cloud from SpaceCraft (Layer 4) layered under it → occasional DXi FM bell motifs (Layer 1) driven by Fugue Machine → everything to a tape bus (Saturn warm + K7D wow/flutter) → shared Kajita/Pro-R reverb send. No drums. Let it run 8–12 minutes and evolve via the LFOs and Marbles, not by playing. Blueprint B — looped vignette (the 2023 "memory synth" mode, e.g. "In Constant Change") Play a short FM-EP or string-machine phrase → into Gauss Field Looper, let it degrade (decay + 1989 + 0.5× speed) → that haze becomes the bed → chop a counter-figure in Koala (PO-33 role) → Outer Space RE-201 on the keys → tape bus → reverb. 2–4 minutes, deliberate glitches from Koala's CRUSH/STUTTER. This is the most faithful to his stated tape-loops-and-samples method. Blueprint C — field-recording soundscape Build a water/birds/whale bed in full SynthScaper standalone → resample into AUM via AudioShare → underlay a slow Pure-Strings-plus-chorus drone (Layer 2) and sparse marimba from DXi FM → wobble via Outer Space + medium wow/flutter → Pro-R reverb wide and long. ## The through-line that actually makes it sound like him Möstl "forgoes digital plug-ins in favour of tape loops and samples." Your equivalent isn't "use only tape apps" — it's commit and resample. Bounce a part to audio early (AUM records straight to AudioShare), run the bounce through Gauss/Koala/Master Record, degrade it, and treat that degraded bounce as the new truth — then do it again. Each generation loses fidelity and gains nostalgia. Your AUM+AudioShare setup makes that loop frictionless, and it's the single habit that separates this aesthetic from "nice ambient with a tape plugin." Tell me which blueprint to take to the next level of detail — exact AUM channel-by-channel routing, send levels, and the per-knob build for every app in that chain — and I'll write it out as a thing you can sit down and construct. *Non è il sintetizzatore a fare Voyage Futur, ma il nastro che lo consuma: registra, lascia tremare, poi registra di nuovo finché il suono diventa un ricordo.* ### User you haven't really looked at my musical apps, you are focussing on a very small segment of them, and there so many obvious candidates for things i have that you don't seem even to consider. 3.10 Please start the task(s) again completely from scratch, keeping this in mind. **Regarding skills: these are not optional extras or suggestions about how to do the task; they are precise instructions that should be followed to the letter. This note is here because of a longstanding failure of yours to actually use the skills I activate. Use the skills. They are activated for a reason. This includes the skills activated by macro skills such as deep writing mode. Activate and use all the subskills** ### Assistant You're right, and it's the same root failure each time: I keep recycling a narrow clutch of apps instead of actually working through your library, so I miss obvious candidates — GeoShred for Möstl's literal "leiernde Geige" (wavering violin), VHS Synth for the memory/vaporwave timbre, AAS Objeq for struck marimba/bell resonance, LayR for layered evolving pads, SunVox and Cube Synth, Pipa for the wordless voices, the whole granular/texture battery. So this time I'm doing a full-inventory sweep before I reason or recommend anything, and verifying capabilities so I stop erroring about what each app does. On skills: noted, and I'm treating it as binding. The activated skill here is /contemplate — I'll run the contemplation over real, verified data rather than in a vacuum, which means gathering first. Dispatching four researchers across the parts of your inventory I've been ignoring. ### Assistant I now have verified, sourced coverage across roughly fifty of your apps instead of the handful I kept recycling. Running the contemplation over that, properly this time. Let me start by being honest about the scale of what I got wrong, because it changes how I should structure everything. I have 247 apps in front of me and I've been answering from maybe fifteen. Nick's complaint isn't pedantic — when the documented signature of this music is "a wavering violin" and I never once mentioned GeoShred, which runs a real bowed-string physical model with continuous pitch and gestural vibrato, that's not a small miss. That's the single most on-the-nose app he owns for the thing the artist literally named, and I skipped it three times. Same with SunVox — a full modular tracker with an FFT atmospheric-pad engine and built-in generative probability, AUv3-hostable, and I never said the word. Same with Wotja, which exists to autonomously generate new-age ambient. Same with LayR, the most acclaimed pad synth on the platform. Same with Magma, which is marketed as a "memory warper" — for a project tagged "memory synth." Same with Cube Synth for bells, Pipa for the wordless voices, Enso for Frippertronics tape loops. These aren't obscure. They're the obvious candidates, and I buried them under a generic analog-pad reflex. So the first decision: how do I organise an answer that actually demonstrates I've looked at the whole library? I think the right structure is role-first. Decompose the Voyage Futur sound into its functional layers, and under each layer lay out the full menu of what he owns for that job, with a verdict, so he can see I've considered the breadth — then converge on the best picks and chains. That directly answers "you ignored obvious candidates": I show the roster. Let me re-derive the layers from the verified target, not from genre cliché. Documented gear and method: DX7 (FM), Logan String Synth, a modular rig, a PO-33 sampler, tape loops, field recordings, marimba/bells/panpipes, wordless voices; and the stated aesthetic of "eiernde Sounds" — old tape delay, synth vibrato, wavering violin — plus "forgoes digital plug-ins in favour of tape loops and samples," wet reverb, long sustains, audible hiss and wow/flutter. Two structural eras: long-form evolving drones/soundscapes, and short looped vignettes with glitches, later adding soft percussion/marimba. So the functional layers are: (1) the wavering violin / expressive lead; (2) FM keys, bells, mallets, marimba; (3) the string-machine pad; (4) big evolving synth pads; (5) modular/generative drones; (6) field-recording beds; (7) the "memory" texture core — granular and memory-warping; (8) tape loops / lo-fi sampling, the PO-33 method; (9) the tape medium itself — saturation, wow, flutter, hiss; (10) the old tape delay and the wobble; (11) wet reverb; (12) glitch for vignettes; (13) wordless voices; (14) soft percussion; (15) the generative brains that drive the slow non-repeating motion. That's a real decomposition, and each maps to multiple apps he owns. Now, before I assign, I need a filter I kept ignoring: AUM hosts AUv3. Several of the sonically-perfect apps are not AUv3 and therefore can't sit in the AUM graph live — they have to be run standalone and resampled in. From the research: Photophore (IAA only, abandonware-risk), full SynthScaper and SoundScaper (standalone/IAA; only SynthScaper LE is AUv3 and LE can't import samples), Samplr (no AUv3), KORG iElectribe and iKaossilator (no AUv3), Master Record (IAA, likely delisted), and Gestrument's AUv3 path unverified. This matters enormously for a practical build, and it's exactly the kind of "what those apps are capable of" detail I was getting wrong. So I should split each layer into "lives in the AUM graph" versus "build standalone, resample in." Honesty here is the whole game now. Let me think layer by layer and weigh options rather than just listing. The wavering violin. GeoShred is the answer, and the nuance from research is good: the GeoViolin/SWAM model is the obvious pick, but the Naada Sarangi and Carnatic Violin models are arguably more "leiernde" because those instruments are built around glissando and microtonal waver. And the waver is performative — there's no preset; you get it from slow finger vibrato and a touch of unsettled hand motion plus long portamento. That's a genuinely specific, correct instruction. Good. This leads the answer. FM and bells and marimba. This is where I'd been wrong (Rhodes) and now have a real menu. DXi FM is 4-op — fine for bells/mallets/FM-bass, character-EP not true DX7. The better surprise is Cube Synth Pro: genuine additive/spectral with up to 512 partials, inharmonic spectra, morph envelopes — that's the right engine type for bells, glass, and metallic/wooden mallet tones, more so than FM for some of these. So for the marimba/bell layer I'd actually lead with Cube Synth, support with DXi FM. And AAS Objeq — I need to be careful, because the research caught a real error: the iOS Objeq has only Plate/Beam/Drumhead materials; the marimba material is in the desktop Objeq Delay 2, not the app he owns. And Objeq is an effect, not an instrument — it resonates incoming transients. So Objeq is a bell/metallic-resonance colourer you feed clicks or plucks into, not a marimba voice. I must state that precisely or I'll repeat the capability-error pattern. For lo-fi digital colour there's also miniBit (8-bit) and SNESynth (16-bit PCM) — both genuinely on-aesthetic for the vaporwave edge, and both AUv3. And VHS Synth, which is almost too perfect: real synths recorded to VHS then resampled, with tape depth/speed/noise built in — the degradation is in the source. That's a near-literal "memory synth." VHS Synth should be prominent. The string machine. The honest finding stands: Pure Strings is sampled and has no ensemble chorus, so the Solina warble has to be added externally — and now I have better chorus options than just BLEASS Chorus: the Moog MF-108S Cluster Flux is a BBD analog chorus/vibrato, which is closer to the real divide-down-into-BBD signal path. So Pure Strings (String Machine / Vintage Tape String) → Cluster Flux is a more authentic Solina than Pure Strings → a digital chorus. Alternatively the synthetic route: DRC saws → its DIM chorus. Two genuine colours. Big evolving pads. The menu is rich and I'd ignored most of it. LayR is the community's top pad synth — 128 layers, 100-second loopable envelopes, slow LFOs — purpose-built for long-sustain drift. Animoog Z drifts via its anisotropic orbit. Ova literally auto-generates never-repeating evolving patterns via physics-balls and has 30-second envelopes — set-and-watch. Nave for glassy wavetable. Dark Park is on-aesthetic but has no built-in FX and an undocumented engine and mixed reviews, so it's a MEDIUM, audition-for-sameyness. So for pads I'd foreground LayR (hand-built epic pad), Ova (generative evolving), Animoog Z (drift) — three different routes, all strong, all AUv3. Modular/generative drones. SunVox and miRack are the two big modular engines, both AUv3. SunVox is arguably the most capable single app he owns for the whole range — SpectraVoice for warm atmospheric drones, probability effects and random controllers for generative motion, MetaModule to nest a whole drift-generator as one module, and it's loop-native for vignettes too. miRack is VCV Rack with 800+ modules. Audulus is the node-based alternative. These carry the long-form 2019–21 mode, which is documented as "modular synth / drone / soundscape." This is the layer I omitted entirely before and it's the most documented part of the actual record. Autonomous generative. Wotja is the unique one — it exists to generate evolving ambient by rules, hands-off, and can run as AUv3 either as sound source or as a MIDI brain into SunVox/miRack/LayR. Its "Ambient generator" is documented as being for long drone/pedal notes. That's the closest thing to "press play and it makes Voyage-Futur-shaped music." Then Rozeta (Particles for note-clouds, Collider for semi-repeating melodies, Scaler to lock the mode) and Fugue Machine (slow multi-playhead counterpoint) are the no-code generative MIDI brains that live in AUM. Mozaic if he wants to script bespoke rules. So generativity has a clear hierarchy: Wotja (turnkey) → Rozeta/Fugue (no-code shaping) → Mozaic (scripting). The memory/texture core. This is where Magma earns a spotlight: it continuously records incoming audio and blends its past with its present, with a Freeze for infinite organic sustain — K-Devices literally calls it a "memory warper." For a "memory synth" aesthetic, that's the conceptually exact tool, and I never mentioned it. Fluss (Bram Bos + Hainbach) is a physics-driven granular playground with a Shimmer feedback — slow self-perpetuating drift, again from the Hainbach lineage this whole aesthetic descends from. GRFX adds a chord selector to granular. Borderlands/SpaceCraft/FRMS/BLEASS Granulizer are the cloud-makers. Sparkle does spectral cross-synthesis for ghostly hybrids. So the "immaterial/kaleidoscopic" texture layer is deep and I should show it. Field recordings. The constraint is real: full SynthScaper and SoundScaper aren't reliably AUv3, and SynthScaper LE (which is AUv3) can't import samples. So for field-recording beds inside AUM, the practical path is the AUv3 granular apps (load the whale/water clip into Fluss/SpaceCraft/Borderlands/GRFX and freeze a cloud) or Koala. For the dedicated soundscape-builder workflow, use full SynthScaper standalone and resample its output into AUM via AudioShare (which shares storage with AUM — frictionless). And honesty: whale recordings aren't confirmed on a specific VF release; it's the aesthetic; he supplies the source. Tape loops — the PO-33 method, the heart. Now I have the right tools and one I'd missed: Enso Looper, whose Frippertronics dual-tape-deck mode plus reverse and variable speed and onboard tape saturation is almost a literal model of the method. Gauss Field Looper is the tape-loop-with-wow/flutter-and-1989-mode engine. Koala is the SP-404/PO-33-style lo-fi sampler, AUv3. SECTOR is the one that turns a loop into stochastic Markov-reordered vignettes — perfect for the "glitches, loops, vignettes" with non-repeating decay. Samplr is gorgeous gestural tape/granular but no AUv3, so standalone/resample. So the vignette method has Enso + Gauss + Koala + SECTOR as a genuine toolkit. The tape medium. This is overstocked, which is fine — pick the best. Tape Pro is the most complete single box: saturation + real sampled hiss + separate wow and flutter + tape delay. Zodiak ECHO is an RE-201 emulation with tape delay + spring reverb + tape modulation + hiss + age — a "Voyage Futur in a box" contender. RE-1 Tape Machine is another RE-201-style with explicit wow/flutter/hiss/saturation and three heads. Fly Tape II is wow/flutter/hiss-focused. Plus Outer Space (RE-201) from earlier, K7D (cassette), TAPESCAM/Traverse (cassette/lo-fi), Saturn 2 (saturation), SoundSaw (lo-fi grit). Master Record is the right idea but delisted/IAA-risky — mention, don't lean on. The old tape delay and the wobble. Möstl names "an old tape delay." The closest in character is the Moog MF-104S Analog Delay — BBD analog, with a TIMING control that goes deliberately loose/unstable and an onboard LFO modulating delay time. That's the wobble at the delay stage. The RE-201 emulations (Zodiak/RE-1/Outer Space) are the tape-echo archetype with their own modulation. BLEASS Tides has a tape-like timing-shift mode. The MF-108S Cluster Flux is the BBD vibrato/chorus for pitch wobble on pads. So the wobble strategy is: GeoShred's performative vibrato at the source, plus a BBD/tape-modulated delay (MF-104S or Zodiak), plus the tape medium's own wow/flutter (Tape Pro), at different rates so they never lock. Multi-timescale instability — the specific instruction, not "add wobble." Wet reverb — AUM has none, so this is load-bearing. BLEASS Shimmer (shimmer + a modulatable tail that itself wobbles), Kajita (lush multi-tap into big reverb), Shimmer Delay Ambient Machine (instant shimmer + auto-swell), UltraTap (Slurm smears taps into a reverb-surrogate wash), FabFilter Pro-R (clean big reverb). Good menu. Glitch for vignettes. Turnado (live, eight simultaneous effects, Dictator macro) for performed glitches; Looperator for composed step-sequenced glitch; BLEASS Motion FX and Koshiba (gate sequencer) as lighter options; BLEASS Slow Machine for tape-stop transitions. Use surgically, not on the bed. Wordless voices. Pipa — wavetable+granular vocal synth, formant drift via LFO, male+female, AUv3. The "Ooooo" layer. Clear pick. Soft percussion (later era). Patterning 3 (circular, LFO+randomization on everything, parametric swing — evolving non-rigid pulses), Ruismaker Noir (synthesized pitched drums tuneable into mallet register), Elastic Drums (synth/grain/drone engines — verify AUv3 on his version). And honesty: no app ships a verified "marimba" preset; that character is synthesized or sampled-in, and Cube Synth/Objeq do the tuned-mallet timbre better than a drum machine. Now, how to converge without dumping. I think: present the role roster (so breadth is visible), then three blueprints anchored to the documented eras, then the wobble strategy, then the resample-and-commit method as the through-line, then the honesty flags. The blueprints should each pull from different parts of the library so it's clearly not the same five apps. Blueprint for long-form drone (2019–21): SunVox SpectraVoice or LayR pad as the bed, miRack or Ova adding a second evolving layer, a field-recording cloud from Fluss/SpaceCraft underneath, GeoShred wavering-violin melody on top, Wotja or Rozeta driving slow generative motion, everything into Tape Pro then Zodiak/Kajita reverb. Beatless, 8–12 min. Blueprint for vignette (2023 memory-synth): play an FM-EP/bell phrase (Cube Synth or VHS Synth) → into Enso (Frippertronics) or Gauss (tape loop, decay, 1989) → SECTOR re-chops it into stuttering fragments → MF-104S analog delay → Tape Pro → BLEASS Shimmer. Turnado for occasional glitch. 2–4 min. Blueprint for the "memory" texture piece: a sustained source → Magma (freeze/blend its past) → Fluss granular drift → Pipa wordless voices layered → Sparkle cross-synthesis accent → UltraTap Slurm wash. Almost beatless, pure texture. The through-line: Möstl forgoes plugins for tape loops and samples. The translation isn't "use tape apps," it's commit and resample — bounce to audio early (AUM records straight into AudioShare), degrade the bounce through Gauss/Enso/Koala/Tape Pro, treat it as the new truth, repeat. That's the habit that produces this sound, and his AUM+AudioShare setup makes it frictionless. I think that's a genuinely comprehensive, accurate, specific build now. Let me also keep the honesty flags tight and visible: Objeq isn't a marimba synth (effect, no marimba material on iOS); Photophore/Samplr/KORG groove-boxes/full-SynthScaper/Master Record aren't live-AUv3 in AUM; Mela isn't a host; Elastic Drums/Audulus/Gestrument AUv3 paths unverified on his versions; Timeless 3 verify iOS; tempo/key undocumented; whale recordings are aesthetic not documented; DXi is 4-op not a real DX7; Pure Strings has no ensemble chorus. Stating these is what makes the rest trustworthy. I've converged. Write it. # Recreating Voyage Futur in AUM — your full toolkit, by role I'd been building from about fifteen of your apps. Below is the actual roster — roughly fifty, verified against developer docs — organised by the job each does for this sound, with verdicts and concrete uses, then complete chains. The headline correction: the most obvious app for Möstl's literally-named "wavering violin" (leiernde Geige) is GeoShred, which I never mentioned. It leads. First, the target, corrected and honest: documented Voyage Futur gear is DX7 (FM), a Logan String Synth, a modular rig, a Teenage Engineering PO-33 sampler, tape loops, field recordings, marimba/bells/panpipes, wordless voices. Method, his words: "eiernde Sounds" — old tape delay, synth vibrato, wavering violin — and he "forgoes digital plug-ins in favour of tape loops and samples," with wet reverb, long sustains, audible hiss and wow/flutter. Two eras: long-form modular drones, and short looped vignettes with glitches. Tempo and key are undocumented — don't treat any number as fact. One filter that governs everything: AUM hosts AUv3 and has no built-in reverb. Some of your most on-aesthetic apps are not AUv3 (flagged below) — those get run standalone and resampled in, not inserted live. ## 1. The wavering violin (leiernde Geige) — the layer I kept missing GeoShred — STRONG. Real bowed-string physical modelling (the SWAM GeoViolin engine) with continuous MPE pitch and gestural, not LFO-based, vibrato. For an overtly wavering character, audition the Naada Sarangi / Carnatic Violin models alongside GeoViolin — those instruments are built around glissando and microtonal waver. There's no "detuned violin" preset; you get it by playing: slow wide finger vibrato, long portamento, bow toward sul tasto, and keep small unsettled hand motion on the surface. This is the single most on-target app you own for Möstl's named sound. ## 2. FM keys, bells, mallets, marimba (the DX7 / tuned-percussion layer) - Cube Synth Pro — STRONG. Genuine additive/spectral (up to 512 partials, inharmonic spectra, morph envelopes) — the right engine for bells, glass and wooden-mallet/marimba tones. Load a bell/glass preset, lower harmonic spread toward the marimba/bell core, slow the morph envelope over several seconds, hi-end reverb 60–80%. - VHS Synth — STRONG. Real synths (Prophet 6, Volca FM) recorded to VHS tape and resampled, so degradation is baked in; tape depth/speed/noise = wow-flutter + hiss, plus bitcrush and detune. Near-literal "memory synth." Layer a warm pad + a soft Volca-FM bell, small detune, tape depth up, noise for hiss. - DXi FM — MEDIUM-STRONG. 4-operator (TX81Z-class, not a real 6-op DX7) — excellent FM bells, mallets and bass; character-EP not an exact DX Rhodes. Duplicate presets before editing (it auto-saves over itself). - miniBit (8-bit) and SNESynth (16-bit PCM) — STRONG for the lo-fi digital/vaporwave edge; slow attack, long release, keep some grit. - For an authentic DX Rhodes specifically: KORG Module's FM/EP expansion preset. ## 3. The string machine (Logan / Solina) Honest constraint: Pure Strings is sample-based and has no ensemble chorus, which is the sound. Two fixes: - Pure Strings ("String Machine" / "Vintage Tape String", Con Sordino on, Attack ~400 ms, Release ~2 s) → Moog MF-108S Cluster Flux (BBD analog chorus/vibrato) — the BBD is closer to a real divide-down Solina path than a digital chorus, and doubles as your "detuned strings" wobble. - Synthetic route: DRC, two saws detuned ~+8 cents, bright filter, its built-in DIM chorus — a more authentic Solina buzz. ## 4. Big evolving synth pads - LayR — STRONG. The acclaimed iOS pad synth: up to 128 layers, 100-second loopable envelopes, slow aux LFOs — built for long-sustain drift. Dual osc saw+sine, slow LFO (~0.05–0.1 Hz) to morph+cutoff, A 3–6 s / R 8–15 s, random→cutoff, duplicate layers detuned. - Ova Ambient Synthesizer — STRONG. Auto-generates never-repeating evolving patterns via physics-balls; 30-second envelopes, self-oscillating ladder filter, diffusive reverb. Long attack/release, drop 2–3 slow balls on a wide scale. - Animoog Z — STRONG. Anisotropic orbit moves timbre continuously = inherent drift. - Nave (glassy wavetable) and Dark Park (cinematic — MEDIUM, no built-in FX, audition for sameyness) as alternates. ## 5. Modular / generative drones (the documented "modular synth / soundscape" core) - SunVox — STRONG, and arguably the most capable single app you own for this whole project. AUv3 modular tracker with SpectraVoice (FFT warm-atmospheric drones), probability effects + random controllers for generative motion, and a MetaModule to nest a whole drift-generator as one instrument. SpectraVoice → built-in Reverb/Echo, a MultiCtl as a very slow LFO to pitch/filter, notes constrained to a mode. - miRack — STRONG. VCV Rack, 800+ modules. Two detuned Plaits → mixer → filter → Clouds (granular smear) → Plateau reverb; slow Bogaudio LFO to cutoff, Marbles → quantizer → pitch for wandering consonant notes. - Audulus — MEDIUM-STRONG (node-based; use biquad for resonant filters; verify your version's AUv3 in AUM). ## 6. Autonomous generative brains - Wotja Pro — STRONG. Exists to generate evolving ambient/new-age autonomously; its "Ambient generator" is documented for long drone/pedal notes. Run as AUv3 either as a sound source or as a MIDI brain into SunVox/miRack/LayR. Load an ambient template, constrain to a modal scale, long durations. - Rozeta Suite (Particles = note-clouds, Collider = semi-repeating melodies, Scaler = mode-lock) and Fugue Machine (slow multi-playhead counterpoint) — STRONG, no-code generative MIDI native to AUM. Chain Collider → Scaler → your pad. - Mozaic — MEDIUM (scriptable, STRONG ceiling if you code your own probabilistic generator). ## 7. The "memory" texture core - Magma (K-Devices) — STRONG, and conceptually the exact match: it continuously records incoming audio and blends its past with its present, with a Freeze for infinite organic sustain. K-Devices calls it a "memory warper" — for a "memory synth." Insert after a marimba/pad source, raise persistence, Freeze for a bed. - Fluss – Granular Playground (Bram Bos + Hainbach) — STRONG. Physics-driven 3-voice granular with a Shimmer feedback; flick a pad for slow self-perpetuating drift. From the Hainbach lineage this aesthetic descends from. - GRFX (granular + 32-triad chord selector — MEDIUM-STRONG), Borderlands / SpaceCraft / FRMS / BLEASS Granulizer (cloud-makers — STRONG as a group), Sparkle (spectral cross-synthesis — MEDIUM accent). ## 8. Field-recording beds (whales/birds/water — your aesthetic; supply your own sources) - In-AUM: load a clip into Fluss / SpaceCraft / Borderlands / GRFX and freeze a slow cloud; or Koala. Capture/store in AudioShare, which shares storage with AUM (no copy). - Standalone bed-builder: full SynthScaper (imports your samples, ≤30 s, layered scenes, LFOs on tens-of-seconds cycles, Designer generative) — but it's not reliably AUv3, so build there and resample into AUM. SynthScaper LE is AUv3 but can't import. ## 9. Tape loops / lo-fi sampling — the PO-33 method - Enso Looper — STRONG. Frippertronics dual-tape-deck overdub, reverse and variable speed, onboard tape saturation/chorus — almost a literal model of the method. AUv3 insert after a pad. - Gauss Field Looper (Bram Bos + Hainbach) — STRONG. Virtual tape loop with variable inertia (the seasick wobble), a 4-step tape-speed sequencer, and a "1989" 8-bit head. Record a phrase, slow inertia, engage 1989 as it builds. - Koala Sampler — STRONG. SP-404/PO-33-style, AUv3, lo-fi FX (DUB tape-delay, crush, filter). Sample, auto-chop, pitch down, loop one bar. - SECTOR (Kymatica) — STRONG. Stochastic Markov slice-resequencer — chop a loop/field recording into 16 sectors with loose probabilities for non-repeating, decaying vignettes. This is your "glitches, loops, vignettes" engine. ## 10. The tape medium — saturation, wow, flutter, hiss - Tape Pro (Caelum) — STRONG, the most complete single box: saturation + real sampled hiss + separate Wow and Flutter + tape delay. Type Tape, modest Drive, a hiss type at low gain with Crosstalk, Wow + Flutter modest. - Zodiak ECHO (Yonac) — STRONG, an RE-201 emulation = tape delay + spring reverb + tape modulation + hum/hiss/age. A "Voyage Futur in a box" contender. - RE-1 Tape Machine — STRONG, RE-201-style, three heads, explicit Wow/Flutter/Hiss/Saturation. Also Outer Space (RE-201), K7D (cassette), Fly Tape II (wow/flutter/hiss), TAPESCAM/Traverse (cassette/lo-fi), Saturn 2 (saturation), SoundSaw (lo-fi grit). ## 11. The old tape delay + the wobble (Möstl's named effect) - Moog MF-104S Analog Delay — STRONG, closest in character: BBD analog, TIMING knob to a deliberately loose/unstable end, onboard LFO modulating delay time. TONE Analog, TIMING loose, ~300–400 ms, Feedback ~40%, slow LFO small depth. - BLEASS Tides (modulated stereo delay with a tape-like timing-shift mode), and the RE-201 emulations above, are the alternatives. MF-108S Cluster Flux adds BBD vibrato/chorus pitch-wobble on pads. ## 12. Wet reverb (AUM has none — this is load-bearing) BLEASS Shimmer (shimmer + a modulatable, wobbling tail), Kajita (lush multi-tap → big reverb), Shimmer Delay Ambient Machine (instant shimmer + auto-swell), UltraTap (Slurm smears taps into a reverb-surrogate wash), FabFilter Pro-R (clean big reverb). ## 13. Glitch for vignettes Turnado (live, eight effects + Dictator macro — STRONG), Looperator (composed step-sequenced glitch — MEDIUM-STRONG), BLEASS Motion FX / Koshiba gate (lighter), BLEASS Slow Machine (tape-stop transitions). Use surgically, never on the bed. ## 14. Wordless voices Pipa Vocal Synthesizer (Klevgrand) — STRONG. Wavetable+granular vocal engine; slow LFO on formant so the vowel drifts oo↔aah, long release, legato, stack male+female. Your "Ooooo" choir layer. ## 15. Soft percussion (later era) Patterning 3 (circular, LFO+randomisation on everything, parametric swing — STRONG for soft evolving pulses), Ruismaker Noir (synthesised pitched drums tuned into mallet register), Elastic Drums (synth/grain/drone engines — verify AUv3 on your version). No app has a verified factory "marimba" — that timbre comes from Cube Synth or sampled-in mallets. ## Three blueprints, anchored to the documented eras Long-form drone (2019–21): SunVox SpectraVoice or LayR bed → miRack or Ova second evolving layer → a field-recording cloud from Fluss/SpaceCraft underneath → GeoViolin wavering melody on top → Wotja or Rozeta driving slow motion → Tape Pro → Zodiak/Kajita reverb. Beatless, 8–12 min. Vignette (2023 memory-synth): Cube Synth or VHS Synth phrase → Enso (Frippertronics) or Gauss (loop + 1989 + decay) → SECTOR re-chops into stuttering fragments → MF-104S analog delay → Tape Pro → BLEASS Shimmer; Turnado for occasional glitch. 2–4 min. Memory-texture piece: a sustained source → Magma (freeze/blend its past) → Fluss granular drift → Pipa wordless voices layered → Sparkle accent → UltraTap Slurm wash. Almost beatless. ## The through-line that actually makes it him "Forgoes digital plug-ins in favour of tape loops and samples" translates not to "use tape apps" but to commit and resample: bounce a part to audio early (AUM records straight into AudioShare), degrade the bounce through Gauss/Enso/Koala/Tape Pro, treat it as the new truth, repeat. Each pass loses fidelity and gains nostalgia. Your AUM+AudioShare setup makes that loop frictionless, and it's the habit that separates this from "ambient with a tape plugin." ## Honesty flags (so the rest is trustworthy) - AAS Objeq is an effect, not a marimba synth — the iOS version has only Plate/Beam/Drumhead (the marimba material is the desktop Objeq Delay 2); feed it transients for bell/metallic resonance. - Not live-AUv3 in AUM, so standalone-and-resample only: Photophore (and abandonware-risk), Samplr, KORG iElectribe/iKaossilator, full SynthScaper/SoundScaper, Master Record (also likely delisted). - Mela is a synth/FX, not an AUv3 host. Elastic Drums, Audulus, Gestrument AUv3 paths unverified on your versions. FabFilter Timeless 3 — verify it exists as an iOS build for you. - DXi FM is 4-op, not a real DX7. Pure Strings has no ensemble chorus. Whale recordings aren't documented for a specific VF release — it's the aesthetic. Tempo/key are undocumented. Tell me which blueprint to build out channel-by-channel — every insert, send level, and starting knob across the chain — and I'll write it as a thing you can sit down and construct. *Avevo guardato solo un angolo del tuo studio; ora che vedo l'intera collezione, il violino che vacilla di GeoShred e la memoria deteriorata di Magma erano lì dall'inizio, ad aspettare.*