trancepenta #

A sibling of trancenwaltzi (same recap/bin/trance.mjs engine, same single-mix scratch-mix.mjs post-FX) but a totally different track: 5/4 · 126 BPM steady · dorian, a ported hellsine gabber kick, a real CC0 horse-gallop / neigh / ocean scene, hats off, beat held until ~0:30. trancenwaltzi is untouched — every difference is opt-in flags, so the non-chill (trancewaltz) path stays byte-identical.

Bake (3 deterministic stages — manual; bake.mjs hardcodes the #

trancenwaltzi names so trancepenta runs the steps explicitly)

O=~/Documents/Working\ Desktop/twi-out
# 1 — engine (single all-sine multi-bus source), opt-in flags:
node recap/bin/trance.mjs --mode chill --meter 5 --master \
  --bpm 126 --scale dorian --hell 11 --chill-hats off --gallop \
  --beat-in 30 --out "$O/trancepenta-MASTER-preBright.wav"
# 2 — in-mix post-FX (scratch + restamped ID + breakbeat + slides),
#     beat-locked to the engine's own struct.json:
node pop/dance/bin/scratch-mix.mjs \
  "$O/trancepenta-MASTER-preBright.wav" "$O/.tp-scr.wav" \
  pop/dance/out/.ac-dot-stamp-vocal.mp3 \
  "$O/trancepenta-MASTER-preBright.wav.assets/struct.json"
# 3 — poppier finalize (duration-aware 18 s fade), → MASTER.wav + mp3:
D=$(ffprobe -v error -show_entries format=duration -of csv=p=0 "$O/.tp-scr.wav")
ffmpeg -y -i "$O/.tp-scr.wav" -af \
"acompressor=threshold=-19dB:ratio=2.4:attack=15:release=240:makeup=1:knee=6,\
equalizer=f=3000:t=q:w=1.2:g=2,highshelf=f=10000:g=1.5,\
loudnorm=I=-14:TP=-1.5:LRA=11,\
alimiter=limit=0.94:attack=8:release=120:level=disabled,\
afade=t=out:st=$(echo "$D - 18" | bc):d=18" \
-ar 44100 -sample_fmt s16 "$O/trancepenta-MASTER.wav"

Final: ~3:11 · −14.4 LUFS · −1.5 dBTP · 44.1 kHz/16-bit stereo.

Flags (engine, all chill-only & opt-in) #

Plus chill-general engine changes (also benefit a future trancenwaltzi re-bake): enriched piano (sustained holds + a melodic top, some reversed / pitch-bent / long-sustain), guaranteed downbeat floor (beat never fully drops), variant per-hit bass, lead harmonies, lowered "organ" pad, melodic highs tamed, cat choir −2 oct.

Sourced SFX — all CC0 (commercial-safe; provenance also in the #

gitignored pop/dance/out/.sfx-credits.txt)

The .gallop/.neigh/.waves/.foghorn/.boathorn.wav live in the gitignored pop/dance/out/ (not committed — re-fetch from the sources above; Freesound key is vault-only, see pop/SCORE.md). The gallop / neigh / ocean samples are placed via the engine's deep + echo + variable-playback-rate helper (they bend with the beat & morph under), beat-spread (random window per hit — not loop-loop).

Vocal + extra SFX (now in) #

• jeffrey vocal counterpoint — sing-penta.mjs (jeffrey-pvc say → whisper-align → WORLD score-pitch on trancepenta-hum.np G-dorian chord tones → rubberband stretch → restrained harmony stack) → pop/dance/out/trancepenta-vocal.mp3 (gitignored, cached say). The bake adds 2 stages after scratch-mix: place-penta-vocal.mjs --phrase …trancepenta-vocal.mp3 --dur <scrDur> then finalize-penta-vocal.mjs --scr …--vox …--vox-db 1 (the exact stage-3 master + a quiet pre-master vocal bus; --vox-db tunes it).
• bake flags also now: --hell 13, --beat-in 30 (subtle dampened 4-on-floor intro + a diatonic-3rd-up key that DROPS home ~1 min) ; the ~2:19 fog horn uses the pre-baked .foghorn-low.wav (−12 st, ~5 s, slow attack/smooth decay); between the two ID stamps a CC0 arpeggiated horse neigh + grenade; tamed/wavy last-30 s.
• New CC0 sources logged in out/.sfx-credits.txt: .waves/.foghorn/.boathorn + .grenade (Freesound, CC0) and .foghorn-low.wav derived from .foghorn.wav.

Sung jeffrey lead-double (v13 — now in) #

pop/dance/trancepenta.txt (lyric: "fog on the harbour / the sea is a screen / ride the slow horse home / drift / hold / glow in the grey tide / the long blue hum") + pop/dance/trancepenta-sing.np (a SUNG DOUBLE of the chill lead — G-dorian tonic-pole contour, long sustained notes). sing-penta.mjs is tuned drippin: slow stretch bpm, very high --max-stretch, big legato overlap, gentle vibrato → a warm gooey co-lead (octave-below body + soft fifth, not the dense stack) → pop/dance/out/trancepenta-vocal.mp3 (39 s, gitignored, cached say). place-penta-vocal.mjs lays 3 post-drop sung entries (~1:18 lead-in · ~1:44 full body · ~2:25 resolve) — silent in the keyed-up intro, clear of the stamp + neigh/grenade (~1:35) and the gnarly last-30 s. finalize-penta-vocal.mjs --vox-db -3 = present co-lead (~+4 dB above the old hum), mastered together. Final v13: 3:11 · −14.1 LUFS · −1.4 dBTP.

Vertical video (now in) #

build.mjs has a trancepenta CONFIG (vertical 2:3, v13 audio, dismal per-section tints, sectionsDir = gens/trancepenta-sections, no prelude, .NO_LYRICS). Section illys regenerated under a dismal / muted palette = the diegetic light, glowing PALS on the laptop backs, screen-glow on jeffrey & mark's faces matching each segment's color (gitignored under ~/Documents/Shelf/gens/trancepenta- sections/<sec>/gens/v1.png). cover-video.mjs patched for stained- glass + all struct/track data with real motion + zoom + a beat- reactive backlit bump. Outputs (build.mjs writes ~/Desktop/builds/ which auto-cleans — rescue to Shelf): twi-out/trancepenta-vertical.mp4 (1080×1920 · 3:11) and a 30 s no-lyrics IG-story cut twi-out/trancepenta-story.mp4 (~1:38→2:08: stamp/horse-grenade → drop → sung body).

Rebuild: regen illys via the section-prompt/gen-illys path, then node pop/dance/bin/build.mjs trancepenta (config carries the rest).