# Celestial Tension - Detailed Quick Start

**Tempo:** 123 BPM | **Key:** A minor | **Mood:** Cinematic, spacious, emotional

This is an original MIDI construction kit, not a preloaded .als Set. The web audio
preview is an eight-bar, eight-stem render of this kit's MIDI through the files in
preview-assets (WAV drums, percussion and atmospheres, Serum .fxp oscillator data and
PML groove MIDI) named in RACKS_AND_ROUTING.md; it is not the exact sound of the racks you will build below.

Listening example: **Anyma & Chris Avantgarde - Consciousness**. https://soundcloud.com/anyma_ofc/anyma-chris-avantgarde-1
This is a study reference, not audio or a copied arrangement in this kit.

## 1. Import and orient

1. Start a blank Live 12 Set in Arrangement View. Set 123 BPM and 4/4 manually.
2. Import `Full_Arrangement.mid` at bar 1. If track separation or naming differs,
   import the individual `MIDI_Stems/*.mid` files at the same bar-1 origin.
3. Keep the written timing and velocity initially; do not apply another full swing
   groove on top of already-offset notes.
4. Use `BUILD_SHEET.md` for section locators and the ending. MIDI marker import is
   not assumed. The original Melodic Techno endings may include a partial bar.
5. Create the four groups and three returns below, then assign instruments.
   Start monitor/headphone volume low: the suggested faders are starting points,
   not guaranteed safe peaks for every sample or preset.

## 2. Required plugins and stock alternatives

**Required third-party plug-ins: none.** Live 12 Standard is the recommended
baseline for the complete track/rack layout. Check your installed edition's devices;
Lite/Intro users may need to consolidate parts to stay within track/device limits.

| Role | Stock starting point | Alternative / caveat |
| --- | --- | --- |
| Kick, hats and percussion | Drum Rack with Simpler one-shots | Use your own or licensed Core Library sounds. Map the exact MIDI pitch shown below. |
| Sub and bass reply | Drift | Operator or Wavetable are optional if licensed; replace rather than require them. |
| Chords, motif and answers | Drift | Any licensed stock synth with the needed voice count; Meld/Operator/Wavetable are optional, not dependencies. |
| Level, mono and width | Utility | Keep the sub centered; do not use a stereo widener to repair a weak bass patch. |
| Sidechain and bus control | Compressor | Glue Compressor is optional; its controls will not match every setting one-for-one. |
| Tone shaping | Instrument filter; EQ Three for broad balance | EQ Eight is optional for precise cleanup; EQ Three is not an identical surgical replacement. |
| Space and repeats | Reverb and Delay | Hybrid Reverb or Echo are optional if licensed; begin with the simpler stock pair. |
| Organization | Instrument Rack / Audio Effect Rack | Macros map parameters inside that rack, not arbitrary controls on other tracks. |

Do not download artist presets or copy commercial recordings into this project.
The source MIDI supplies notes, not instruments, samples, routing, FX or automation.

## 3. Exact track routing and starting levels

| Track | Instrument / MIDI mapping | Audio To | Starting fader | Processing / notes |
| --- | --- | --- | --- | --- |
| Kick | Drum Rack -> Simpler on MIDI pitch 36 | DRUMS | -10 dB | Dry; no sidechain on kick |
| Clap | Drum Rack -> Simpler on MIDI pitch 39 | DRUMS | -20 dB | Short-room send only if needed |
| Closed Hat | Drum Rack -> Simpler on MIDI pitch 42 | DRUMS | -20 dB | Short-room send only if needed |
| Open Hat | Drum Rack -> Simpler on MIDI pitch 46 | DRUMS | -20 dB | Short-room send only if needed |
| Percussion | Drum Rack -> Simpler on MIDI pitch 47 | DRUMS | -20 dB | Short-room send only if needed |
| Sub Bass | Drift, one voice, triangle/sine-like fundamental | BASS | -14 dB | Utility Width 0%; external Compressor |
| Mid Bass | Drift, one voice, pulse/saw body | BASS | -20 dB | Reduce competing lows; external Compressor |
| Chords | Drift, polyphonic for chords; mono or poly for individual phrases | MUSIC | -22 dB | VOICE rack; light sidechain; returns A/B/C |
| Pluck | Drift, polyphonic for chords; mono or poly for individual phrases | MUSIC | -22 dB | VOICE rack; light sidechain; returns A/B/C |
| Arp | Drift, polyphonic for chords; mono or poly for individual phrases | MUSIC | -22 dB | VOICE rack; light sidechain; returns A/B/C |
| Lead | Drift, polyphonic for chords; mono or poly for individual phrases | MUSIC | -22 dB | VOICE rack; light sidechain; returns A/B/C |
| Atmos Cue | Drift pitched/noise texture; follow the cue gates | ATMOS + FX | -28 dB | Slow attack/release, remove unnecessary lows, gentle ducking |

For each drum track, load the one-shot into the Drum Rack pad matching its **MIDI
number**, not just a displayed octave label. The rack triggers Simpler at its intended
sample pitch. Dragging a pitched drum MIDI stem straight onto a default Simpler may
transpose the sample unintentionally; verify pitch if using that alternative.

```text
DRUMS -> Main
  Kick -> DRUMS
  Clap -> DRUMS
  Closed Hat -> DRUMS
  Open Hat -> DRUMS
  Percussion -> DRUMS
BASS -> Main
  Sub Bass -> BASS
  Mid Bass -> BASS
MUSIC -> Main
  Chords -> MUSIC
  Pluck -> MUSIC
  Arp -> MUSIC
  Lead -> MUSIC
ATMOS + FX -> Main
  Atmos Cue -> ATMOS + FX
SC GHOST -> Sends Only (A/B/C all -inf; track active)
A SHORT ROOM -> Main
B DARK SPACE -> Main
C FILTERED DELAY -> Main
```

Set ordinary track Monitor to Auto and disarm after recording. Do not route group
outputs back into their child tracks. Return sends start at -inf and returns go to
Main, never into themselves or a feedback chain.

## 4. Build this template's voices

- **Bass:** Use a clean triangle sub and a quieter pulse/saw Mid Bass. For the wide rhythm, shorten the reply before increasing sub level.
- **Melodic voice:** Start Pluck with a filtered saw/triangle and a short amplitude envelope. Use the cinematic, spacious, emotional mood as a brief; add brightness through automation, not by stacking duplicated leads.
- **Percussion:** Use a short original/licensed kick, dry clap, closed/open hats and a contrasting wood or tom one-shot. Preserve the written velocity accents.
- **Harmony:** scale-root progression `[0, 3, 6, 5]`.
  Use the written Chords MIDI rather than transposing every layer to a single root.

Set Sub Bass to one voice and keep its output mono. The sub's octave is already in
the MIDI: check the actual pitch before transposing. Give Mid Bass more upper
harmonics but less fundamental than the sub. Use at least as many synth voices as
the simultaneous notes on Chords; do not accidentally run chords through a mono patch.

## 5. Ghost sidechain - active but inaudible

Create SC GHOST manually and program a short quarter-note trigger from bar 1 to the ending; this original kit does not include a ghost stem.

1. Use a short click on a Drum Rack pad at MIDI pitch 36. Keep the track activator ON.
2. Set Audio To to **Sends Only**, with every return send at **-inf**.
3. On each receiving Compressor, enable External Sidechain and select
   **SC GHOST / Post FX**. Do not select the Main mix or the bass itself.
4. Sub Bass: start around 4:1, attack 1-3 ms, release 107 ms, makeup off.
   Adjust threshold by ear for roughly 2-4 dB of gain reduction, not to a universal dB value.
5. Mid Bass: start around 2:1, attack 5 ms, release 132 ms;
   use less reduction than the sub.
6. Pluck, Chords and return B: optional light ducking, roughly 1-2 dB.
   Keep their attack audible and release before the next quarter-note trigger.

Solo a receiving bass while the audible kick is absent: ducking should continue.
If not, verify that the ghost instrument makes a click and the source tap is Post FX.
Do not mute the ghost track to silence it; silence its destination instead.

## 6. Return effects

| Return | Devices / initial settings | Feed from | Keep out |
| --- | --- | --- | --- |
| A SHORT ROOM | Reverb, 100% wet, about 0.6-0.9s decay, modest pre-delay | Rim/clap, selected percussion, a little Pluck; start sends near -24 dB | Kick and Sub Bass |
| B DARK SPACE | Reverb, 100% wet, about 2.5-4s decay -> Compressor keyed from SC GHOST | Chords, phrase endings, replacement atmosphere; start near -28 dB | Low-end fundamentals and dense continuous hats |
| C FILTERED DELAY | Delay, 100% wet, synced eighth or dotted-eighth, feedback initially 20% and capped near 30% | Pluck and the melodic answer; start near -24 dB | Kick and sub; any self/return feedback path |

Roll unnecessary lows and excessive highs out of wet returns using their internal
filters or an available EQ. Level-match before judging whether more reverb is useful.
Do not turn a return to 100% wet and then also route dry audio through it in series.

## 7. Macro assignments - build a real rack

On **Pluck**, create one Instrument Rack named `VOICE - Celestial Tension` with one chain:

```text
Drift -> Reverb (short insert) -> Delay -> Utility
```

All four devices must be **inside the same rack chain**. Keep the track's external
sidechain Compressor after the rack. Reveal eight macros, enter Map mode, choose
each device parameter and assign its macro. Set Min/Max in the Mapping Browser.
The insert Reverb is only a small close-space layer; large space stays on return B.

| Macro | Exact device target | Suggested Min-Max | Starting value | Purpose |
| --- | --- | --- | --- | --- |
| 1 Tone | Drift -> filter cutoff | 180-7500 Hz | 900 Hz | Reveal the phrase without raising its fader. |
| 2 Decay | Drift -> amplitude-envelope decay | 80-732 ms | 244 ms | Keep note tails clear of the next attack; MIDI note-off still matters. |
| 3 Release | Drift -> amplitude-envelope release | 20-390 ms | 90 ms | Extend phrase endings, not the whole low end. |
| 4 Width | Utility -> Width | 0-120% | 90% | For this melodic rack only; never map the sub's width here. |
| 5 Short Space | Reverb -> Dry/Wet | 0-12% | 4% | Set Reverb decay to 0.7s; larger space comes from return B. |
| 6 Echo Amount | Delay -> Dry/Wet | 0-18% | 5% | Set synced delay to an eighth note before mapping. |
| 7 Echo Feedback | Delay -> Feedback | 0-30% | 15% | Keep the upper limit capped; do not create a feedback loop. |
| 8 Output | Utility -> Gain | -12-0 dB | -6 dB | Compensate for tone changes instead of chasing loudness. |

Save this Instrument Rack separately after auditioning. Duplicate it onto the
counterline if useful, but tune the second rack independently; one macro cannot
silently control another track's rack. Track return-send levels must be automated
on the track itself, not mapped as if they were parameters inside this rack.

For the bass rack, use Drift -> Utility with Width fixed at 0% on the sub, followed
by an external Compressor. Optional bass macros: filter cutoff 80-700 Hz, amp decay
80-488 ms, amp release 10-107 ms and Utility output -12-0 dB.
Leave sidechain threshold manual until the ghost and bass levels are established.

## 8. Bus processing - subtle and reversible

| Bus | Chain | Starting approach |
| --- | --- | --- |
| DRUMS | Utility -> optional EQ -> Compressor | Ratio 1.5:1-2:1, attack 20-30 ms, release about 122 ms. Aim for no more than 1-2 dB reduction. Keep kick attack intact. |
| BASS | Utility -> optional EQ | Balance sub/reply first. Avoid another aggressive compressor on top of per-track ducking. Check mono and crossover overlap. |
| MUSIC | Utility -> optional EQ -> gentle Compressor if needed | Keep level and width controlled. Use roughly 1.5:1 and 1 dB reduction only if the chord/motif stack actually needs it. |
| ATMOS + FX | Utility -> optional filter/EQ | Remove unnecessary low content, automate fades and prevent textures from masking the melody. |
| Main | Utility for conservative trim; metering | Leave several dB of peak headroom, approximately -6 dBFS as a practical starting target. No mastering limiter is required to build the template. |

Optional Drum Buss, Saturator, Glue, Echo or third-party processors are enhancements,
not requirements. Do not insert them by habit. Bypass and level-match every bus change;
if it only sounds better because it is louder, undo the gain difference.

## 9. Section-by-section macro and mix moves

These are **manual automation suggestions**, not automation embedded in the MIDI.

| Section | Bars, 1-based inclusive | Suggested move |
| --- | --- | --- |
| Intro Atmosphere | 1-16 | Keep Tone near 500-900 Hz, Echo Amount near 0-5%; leave the sub out until its written entry. |
| Groove Introduction | 17-32 | Keep Tone near 500-900 Hz, Echo Amount near 0-5%; leave the sub out until its written entry. |
| Bass Establishment | 33-48 | Introduce the written layer; keep Mid Bass below the sub and expose the motif without widening bass. |
| Theme A | 49-72 | Introduce the written layer; keep Mid Bass below the sub and expose the motif without widening bass. |
| Breakdown | 73-88 | Pull audible low end back; lower Tone and cautiously raise return B. Keep SC GHOST active. |
| Build | 89-104 | Raise Tone gradually, shorten Release, and reduce reverb before the next downbeat. |
| Peak Theme | 105-136 | Open Tone toward 4875 Hz, bring the answer forward, and level-match Output. |
| Release | 137-148 | Close Echo Amount and return B; strip melodic density before the final percussion handoff. |
| DJ Outro | 149-154 | Close Echo Amount and return B; strip melodic density before the final percussion handoff. |

For the original Melodic Techno MIDI, treat the breakdown label as an arrangement cue: manually remove or automate down audible kick/bass where needed rather than assuming every baseline drum note is absent.
Keep the ghost running throughout. Load Air/Atmos Cue with the specified original
pitched/noise texture and follow its written gates; do not accidentally turn it into a loud lead.

## 10. Save and acceptance check

1. Audition the intro, full theme, breakdown, peak and ending at conservative volume.
2. Check MIDI-pad assignments, missing samples, clipping, mono bass, inactive
   sidechain sources and unexpected double-routing.
3. Confirm every macro changes its named target and returns to a usable starting state.
4. Use **Collect All and Save** for any samples you add, then **File -> Save Live Set
   as Template**. Also save reusable racks into your User Library.
5. Reopen the saved template as a new untitled Set and confirm it loads without
   missing devices or samples. Native Set/rack saving and sound auditioning must happen in Live.

References: https://www.ableton.com/en/live-manual/12/managing-files-and-sets/
and https://www.ableton.com/en/live/compare-editions/


# Celestial Tension - Pluck, Arp, Lead and Atmosphere Design

Wide two-bar bell plucks, octave-ladder arp, long ascending lead and luminous suspended swells.

These are four separately composed MIDI voices, not copies with different names or keys.
The preview gives each its own mute/solo/volume control and uses the patch envelopes below.
The preview plays these notes through the real preview-assets files named under each voice (Serum .fxp oscillator data and WAV textures); it is an approximation, not the exact Serum or Live sound.

## Pluck

Bell-like answers spread over two bars.

- **Phrase:** 32 steps at 4 steps per beat (8 beats); MIDI octave basis 4.
- **Notes:** `[[0, 0, 2], [7, 4, 2], [18, 6, 3], [27, 2, 2]]`. Each entry is [step offset, scale degree, gate in steps]. Degrees are zero-based offsets from the current chord root.
- **Oscillator starting point:** Stock alternative: Drift triangle with a short bright filter envelope. Optional Operator FM can create a more inharmonic bell.
- **Amp envelope:** attack 8 ms; decay 600 ms; release 240 ms; sustain target 0%.
- **Brightness target:** 45% on the audition model. In Live, match by ear using the filter; this is not a claimed one-to-one cutoff conversion.
- **Rack focus:** map Tone to Drift cutoff, Snap to amp decay, Tail to amp release and Echo to a short Delay's wet level. Keep the transient dry and preserve the written rests.
- **Suggested file:** [`PML PL Atmo Go.fxp`](https://github.com/star6ix9/ian-ai-playground/blob/main/music-production/ableton-live-12-melodic-techno/preview-assets/PML%20PL%20Atmo%20Go.fxp) (Serum patch **PML PL Go - Pluck Atmo**); reference phrase [`PML PL Atmo Go - Em.mid`](https://github.com/star6ix9/ian-ai-playground/blob/main/music-production/ableton-live-12-melodic-techno/preview-assets/PML%20PL%20Atmo%20Go%20-%20Em.mid), clip transpose +5 st.
- **Recipe:**
    1. Browser → **Plug-ins** → VST3 (or AU) → **Xfer Records** → **Serum**; drag it onto the `Pluck` MIDI track (or into the first chain of its Instrument Rack).
    2. Open Serum, click the preset name → **Load Preset…** and choose [`PML PL Atmo Go.fxp`](https://github.com/star6ix9/ian-ai-playground/blob/main/music-production/ableton-live-12-melodic-techno/preview-assets/PML%20PL%20Atmo%20Go.fxp), or drag the .fxp from Finder onto the Serum window. The patch loads as **PML PL Go - Pluck Atmo**.
    3. What it contains: Osc A `Basic Shapes`, Osc B `Basic Shapes`. The template MIDI keeps its own notes; the preset only supplies the sound.
    4. Click the Serum device's **Configure** button, move each named macro inside Serum once, then map it to the rack: Serum Macro 1 `OPEN` (keeps the preset's own assignment) → Rack Macro 1; Serum Macro 3 `DELAY` (FX Delay Mix) → Rack Macro 3; Serum Macro 4 `REVERB` (FX Reverb Mix) → Rack Macro 4.
    5. Set Serum's Env 1 to the amp envelope above (attack, decay, sustain, release) so the written gates behave as designed; the preset's own envelope is only a starting point.
    6. Chain after the plug-in: EQ Eight (high-pass 24 dB at 150 Hz) → Echo (Sync, 3/16 both channels, Feedback 32%, Dry/Wet 18%) → Compressor (sidechain, see below); send 20% to LONG SPACE.
    7. Compressor: Sidechain on, Audio From **SC GHOST / Post FX**, Ratio 4:1, Attack 0.5 ms, Release 120 ms, Threshold for ~2–3 dB reduction (≈ the preview's 30% duck).
    8. **No Serum?** Wavetable: Basics > Basic Shapes, position ~66% (saw frame); amp match VOICE_DESIGN.md (or A 2 ms / D 280 ms / S 0% / R 180 ms); filter Filter 1 Lowpass 24 dB, Freq ~1.1 kHz, Res 20%, Env 2 → Freq +60%, D 220 ms S 0%; Unison Classic, 2 voices, Amount 12%.
- **Active bars:** 45-72, 93-140. Gates are clipped at each active-window boundary.

## Arp

Two-bar octave ladder with a missing rung.

- **Phrase:** 16 steps at 2 steps per beat (8 beats); MIDI octave basis 4.
- **Notes:** `[[0, 0, 0.7], [2, 4, 0.7], [4, 7, 0.7], [6, 9, 0.7], [9, 6, 0.7], [11, 4, 0.7], [14, 2, 0.7]]`. Each entry is [step offset, scale degree, gate in steps]. Degrees are zero-based offsets from the current chord root.
- **Oscillator starting point:** Use Drift triangle with a low filter cutoff; optional Operator gives a purer sine.
- **Amp envelope:** attack 12 ms; decay 380 ms; release 120 ms; sustain target 0%.
- **Brightness target:** 35% on the audition model. In Live, match by ear using the filter; this is not a claimed one-to-one cutoff conversion.
- **Rack focus:** map Bite to cutoff, Gate Tail to amp release and Motion to a subtle internal modulation depth. The MIDI already contains the arp: do not add an Arpeggiator unless intentionally replacing this pattern.
- Listen with Pluck muted first, then combine them. Reduce arp level before changing its note density; it is a supporting moving voice, not a second lead.
- **Suggested file:** [`PML PL Arpy Lake.fxp`](https://github.com/star6ix9/ian-ai-playground/blob/main/music-production/ableton-live-12-melodic-techno/preview-assets/PML%20PL%20Arpy%20Lake.fxp) (Serum patch **PML PL The Lake - Arp**); reference phrase [`PML PL Arpy Lake - A#m.mid`](https://github.com/star6ix9/ian-ai-playground/blob/main/music-production/ableton-live-12-melodic-techno/preview-assets/PML%20PL%20Arpy%20Lake%20-%20A%23m.mid), clip transpose -1 st.
- **Recipe:**
    1. Browser → **Plug-ins** → VST3 (or AU) → **Xfer Records** → **Serum**; drag it onto the `Arp` MIDI track (or into the first chain of its Instrument Rack).
    2. Open Serum, click the preset name → **Load Preset…** and choose [`PML PL Arpy Lake.fxp`](https://github.com/star6ix9/ian-ai-playground/blob/main/music-production/ableton-live-12-melodic-techno/preview-assets/PML%20PL%20Arpy%20Lake.fxp), or drag the .fxp from Finder onto the Serum window. The patch loads as **PML PL The Lake - Arp**.
    3. What it contains: Osc A `Analog_BD_Sin`, Osc B `PWM LPF`. The template MIDI keeps its own notes; the preset only supplies the sound.
    4. Click the Serum device's **Configure** button, move each named macro inside Serum once, then map it to the rack: Serum Macro 1 `FILTER` (Filter 1 Cutoff) → Rack Macro 1; Serum Macro 4 `WET` (keeps the preset's own assignment) → Rack Macro 4.
    5. Set Serum's Env 1 to the amp envelope above (attack, decay, sustain, release) so the written gates behave as designed; the preset's own envelope is only a starting point.
    6. Chain after the plug-in: EQ Eight (high-pass 24 dB at 200 Hz) → Echo (Sync, 3/16, Feedback 32%, Dry/Wet 18%) → Auto Pan (Rate 1/2 bar, Amount 20%); send 20% to LONG SPACE.
    7. **No Serum?** Operator: Osc A Sine, Coarse 1 + Drift: Shape = Pulse, Filter LP ~2.5 kHz, PWM from LFO at 0.3 Hz; amp match VOICE_DESIGN.md (or A 1 ms / D 160 ms / S 0% / R 120 ms); filter Filter 1 Lowpass 24 dB, Freq ~1.4 kHz, Res 18%, Env 2 → Freq +45%, D 150 ms; Unison Classic, 2 voices, Amount 10%.
- **Active bars:** 53-72, 89-136. Gates are clipped at each active-window boundary.

## Lead

Long ascending notes ending above the opening.

- **Phrase:** 24 steps at 1 steps per beat (24 beats); MIDI octave basis 5.
- **Notes:** `[[0, 0, 3], [6, 2, 2], [11, 5, 3], [18, 6, 4]]`. Each entry is [step offset, scale degree, gate in steps]. Degrees are zero-based offsets from the current chord root.
- **Oscillator starting point:** Start with Drift's triangle oscillator; keep the second oscillator quieter.
- **Amp envelope:** attack 140 ms; decay 1200 ms; release 600 ms; sustain target 50%.
- **Brightness target:** 28% on the audition model. In Live, match by ear using the filter; this is not a claimed one-to-one cutoff conversion.
- **Rack focus:** map Presence to cutoff, Bloom to amp attack, Sustain to amp sustain and Space to a short rack reverb. Use an independent rack, not the pluck's macro defaults.
- Start mono for single-note phrases. The lead has longer note gates and larger rests than the arp; do not quantize those away.
- **Suggested file:** [`PML LD Short Lake.fxp`](https://github.com/star6ix9/ian-ai-playground/blob/main/music-production/ableton-live-12-melodic-techno/preview-assets/PML%20LD%20Short%20Lake.fxp) (Serum patch **PML LD The Lake - Lead Main**); reference phrase [`PML LD Short Lake - A#m.mid`](https://github.com/star6ix9/ian-ai-playground/blob/main/music-production/ableton-live-12-melodic-techno/preview-assets/PML%20LD%20Short%20Lake%20-%20A%23m.mid), clip transpose -1 st.
- **Recipe:**
    1. Browser → **Plug-ins** → VST3 (or AU) → **Xfer Records** → **Serum**; drag it onto the `Lead` MIDI track (or into the first chain of its Instrument Rack).
    2. Open Serum, click the preset name → **Load Preset…** and choose [`PML LD Short Lake.fxp`](https://github.com/star6ix9/ian-ai-playground/blob/main/music-production/ableton-live-12-melodic-techno/preview-assets/PML%20LD%20Short%20Lake.fxp), or drag the .fxp from Finder onto the Serum window. The patch loads as **PML LD The Lake - Lead Main**.
    3. What it contains: Osc A `Basic Mg`, Noise `BrightWhite`, 1 embedded custom wavetable frames. The template MIDI keeps its own notes; the preset only supplies the sound.
    4. Click the Serum device's **Configure** button, move each named macro inside Serum once, then map it to the rack: Serum Macro 1 `FILTER` (Filter 1 Cutoff) → Rack Macro 1; Serum Macro 2 `FILTER ENV` (keeps the preset's own assignment) → Rack Macro 2; Serum Macro 3 `DELAY` (FX Delay Mix) → Rack Macro 3; Serum Macro 4 `REVERB` (FX Reverb Mix) → Rack Macro 4.
    5. Set Serum's Env 1 to the amp envelope above (attack, decay, sustain, release) so the written gates behave as designed; the preset's own envelope is only a starting point.
    6. Chain after the plug-in: EQ Eight (high-pass 24 dB at 180 Hz) → Echo (Sync, 3/16, Feedback 32%, Dry/Wet 18%) → Hybrid Reverb (Dry/Wet 28%, Decay 3.5 s) or a 30% send to LONG SPACE.
    7. **No Serum?** Drift: Shape = Saw, Drift ~12%; amp match VOICE_DESIGN.md (or A 15 ms / D 600 ms / S 70% / R 350 ms); filter Filter 1 Lowpass 12 dB, Freq ~2.5 kHz, Res 15%, Env 2 → Freq +20%; Voices Mono for single-note phrases, Glide 40 ms; add Drift Noise at ~10% high-passed above 2 kHz to stand in for the preset's `BrightWhite` noise oscillator.
- **Active bars:** 57-68, 109-132. Gates are clipped at each active-window boundary.

## Atmos Cue

Two luminous swells with a suspended upper tone.

- **Phrase:** 24 steps at 1 steps per beat (24 beats); MIDI octave basis 4.
- **Notes:** `[[1, 4, 8], [15, 6, 5]]`. Each entry is [step offset, scale degree, gate in steps]. Degrees are zero-based offsets from the current chord root.
- **Oscillator starting point:** Start with Drift's triangle oscillator; keep the second oscillator quieter.
- **Amp envelope:** attack 1100 ms; decay 3000 ms; release 1300 ms; sustain target 75%.
- **Brightness target:** 18% on the audition model. In Live, match by ear using the filter; this is not a claimed one-to-one cutoff conversion.
- **Texture:** the preview blends 12% relative noise into a pitched texture. In Live, begin with a very quiet Drift noise source or your own noise sample, then add the pitched layer.
- **Cue semantics:** each MIDI note deliberately controls the onset, pitch and held length of a swell, impact or bed. It is no longer just the same one-bar root at every section boundary.
- **Rack focus:** map Rise to amp attack, Tail to amp release, Air to noise level and Distance to reverb wet level. Keep every mapped device inside that Atmos rack.
- Route to ATMOS + FX, start its fader around -28 dB, remove unnecessary low frequencies, and use gentle ghost-sidechain ducking if the texture masks the kick or melody.
- **Suggested file:** [`PML_MTSP1_Atmo_001_124BPM_Am.wav`](https://github.com/star6ix9/ian-ai-playground/blob/main/music-production/ableton-live-12-melodic-techno/preview-assets/PML_MTSP1_Atmo_001_124BPM_Am.wav) (A minor, transpose 0 st).
- **Recipe:**
    1. Drag [`PML_MTSP1_Atmo_001_124BPM_Am.wav`](https://github.com/star6ix9/ian-ai-playground/blob/main/music-production/ableton-live-12-melodic-techno/preview-assets/PML_MTSP1_Atmo_001_124BPM_Am.wav) (16 bars, A minor) onto an empty MIDI track to create a Simpler.
    2. Simpler **Classic** mode, Warp **on** in **Complex Pro** (Formants 100%, Envelope 128), Loop on, Transpose **0 st** so the texture sits in A minor.
    3. Simpler amp envelope Attack 1100 ms, Decay 3000 ms, Sustain 75%, Release 1300 ms (this kit's Atmos Cue design); Filter Lowpass 12 dB around 3 kHz; Spread 40%; Voices 4.
    4. Chain EQ Eight (high-pass 24 dB at 200 Hz) → Auto Filter (LFO 1/2 bar, Amount 15%) → Compressor sidechained from SC GHOST (2–3 dB). Start the fader at -28 dB and send 35% to LONG SPACE.
    5. Transition layer: load Serum with [`PML FX Downlifter Without.fxp`](https://github.com/star6ix9/ian-ai-playground/blob/main/music-production/ableton-live-12-melodic-techno/preview-assets/PML%20FX%20Downlifter%20Without.fxp) on a separate `ATMOS + FX` track and drag [`PML FX Downlifter Without.mid`](https://github.com/star6ix9/ian-ai-playground/blob/main/music-production/ableton-live-12-melodic-techno/preview-assets/PML%20FX%20Downlifter%20Without.mid) onto it at the section boundary. Keep it 6–8 dB under the atmos texture.
- **Active bars:** 1-16, 73-136, 149-ending. Gates are clipped at each active-window boundary.

## Compare and save

1. Solo Pluck, Arp, Lead and Atmos separately in the web preview. Then combine them at reduced volume.
2. Import their individual MIDI stems into Live at bar 1 and load the four independent racks.
3. Check the lead's space against the arp and the atmosphere's low-frequency content against the bass.
4. Save each voice rack after matching its envelope; save the full native Set only after auditioning.

The drum, bass and chord foundation, tempo, key and five-minute format are intentionally retained.