Skip to content

feat(prompt): bass/acid/delay Phase 2 recommendation rules + example cards #147

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
slittycode merged 1 commit into from
Jun 6, 2026
Merged

feat(prompt): bass/acid/delay Phase 2 recommendation rules + example cards #147

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 17 additions & 0 deletions apps/backend/prompts/phase2_system.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -187,6 +187,12 @@ KICK CHARACTER (kickDetail)
- kickDetail.thd — >0.2 indicates audible saturation; cite for Saturator Drive. BUT when kickDetail.isDistorted is true the source already clips — do NOT add Saturator on the kick; cite kickDetail.isDistorted for the omission.
- kickDetail.harmonicRatio — high (>0.9) = clean/sub-heavy (avoid Drum Buss Crunch, which would add unwanted grit); low (<0.5) = noisy/snappy (Drum Buss Crunch helps unify the transient).

BASS CHARACTER (bassDetail)
- bassDetail.averageDecayMs → Glue Compressor Release. The measured mean bass decay sets the release band: <100 ms (punchy) → fast Release ≈50-120 ms; 100-300 ms (medium) → ≈150-250 ms; 300-600 ms (rolling) → ≈300-450 ms; 600+ ms (sustained) → 0.6-1.2 s or Auto. Cite bassDetail.averageDecayMs with bassDetail.type. SENTINEL GUARD: when bassDetail.transientCount < 2 the decay is unmeasured (no isolable bass onsets) — do NOT emit a bass release value; state that the bass decay could not be reliably measured.
- bassDetail.type directs the dynamics device: "punchy"/"medium" → a Compressor with a fast attack to tighten transients (kick-bus style); "rolling"/"sustained" → a Glue Compressor with a longer Release for level glue rather than transient control. Cite bassDetail.type with bassDetail.transientRatio.
- bassDetail.fundamentalHz → bass tuning/EQ: cross-reference with key to name the bass root before any narrow move. Honor the schema caveat — do NOT drive a filter Q within ±15 Hz of fundamentalHz without pitchDetail corroboration (the ZCR estimate biases upward on harmonic-rich basses). Cite bassDetail.fundamentalHz with key.
- bassDetail.swingPercent / bassDetail.grooveType → Groove Pool swing when rebuilding the bass as MIDI: "straight" → none; "slight-swing" → ≈8-16%; "heavy-swing"/"shuffle" → ≈25-50%, toward the measured swingPercent. This is the only swing value scoped to the bass specifically — prefer it over grooveDetail.perDrumSwing for bass humanization. Rule-only; do not template a separate card for a Groove Pool entry.

PER-DRUM SWING (Phase 1.C #3)
- grooveDetail.perDrumSwing.{kick | snare | hihat} — swing values computed from the three beat-loudness bands (20-200 Hz / 200-4000 Hz / 4000-20000 Hz). Cite these together when recommending Groove pool entries, MIDI quantize amounts, or "humanize" automation — they let you say "humanize the snare 12% but keep the kick rigid" with the measurement attached.
- The legacy fields `grooveDetail.kickSwing` and `grooveDetail.hihatSwing` still exist and equal `perDrumSwing.kick` / `perDrumSwing.hihat` for backward compatibility — prefer the perDrumSwing object when citing two or three drum groups so the citations don't fragment.
Expand All @@ -197,6 +203,11 @@ REVERB DETAIL (Phase 1.D #5)
- reverbDetail.preDelayMs — pre-delay estimate in milliseconds. Cite directly for the Reverb device Predelay parameter (`reverbDetail.preDelayMs: 42` → "Set Reverb Predelay to 42 ms"). When close to zero, the source is dry — don't recommend a reverb with audible predelay.
- When `stemAnalysis.{stem}.reverbDetail` is present (Phase 1.D #5 per-stem), cite it for per-element reverb advice — e.g. "drums-bus Predelay 12 ms; other-bus Predelay 38 ms with longer highs (HighMids RT60 1.8 s) → Hall-style reverb with High Cut lowered for more damping." This is the cleanest way to recommend different reverb treatments per element.

ACID / FILTER MOVEMENT (acidDetail) and DELAY
- acidDetail.isAcid → Auto Filter on the bass-like element. When isAcid is true, recreate the 303 squelch with Auto Filter: a Lowpass Filter Type, high Resonance scaled by acidDetail.resonanceLevel (higher level → higher Resonance), raised LFO Amount, and Drive for grit; the LFO speed should track acidDetail.centroidOscillationHz. Cite acidDetail.resonanceLevel with acidDetail.centroidOscillationHz; hedge ("acid-style movement may be present") when acidDetail.confidence < 0.5, and omit the move entirely when isAcid is false.
- Delay / Echo BPM-sync: when the material is sparse enough to support audible repeats (a lead, stab, or chop — not a dense wall), add Echo or Delay with Sync on and the repeat division derived from bpm via the GROOVE CALCULATION RULE (a dotted-eighth repeat suits dub/house; a quarter-note repeat suits sparse leads). Keep Feedback modest (≈20-40%) so the tail decays within a couple of bars. Cite bpm; when rhythmDetail.tempoCurve is non-flat prefer Echo (it re-syncs to tempo drift) and say so.
- effectsDetail.gatingDetected → Beat Repeat / Gate: when gatingDetected is true, recreate the chop with Beat Repeat — match the repeat grid to effectsDetail.gatingRate (a "16th" rate → a 1/16 grid) and raise the repeat chance when effectsDetail.gatingRegularity is high (machine-like) versus a lower chance for irregular/glitch feel. For a plain on/off chop use Gate instead. Cite effectsDetail.gatingRate with effectsDetail.gatingRegularity; omit when gatingDetected is false.

CHORD TIMELINE (Phase 1.D #2)
- chordDetail.chordTimeline is an array of `{startSec, endSec, label, labelLong, confidence}` segments produced by a 25-state (12 major + 12 minor + "N" no-chord) Viterbi decoder over `librosa.feature.chroma_cqt`. `label` is short form (`"Cm"`, `"Eb"`, `"N"`) and `labelLong` is the readable form (`"C minor"`, `"Eb major"`, `"N"`) — cite either. Use for arrangement-aware harmonic claims: "the build sits on Cm from 0:48 to 1:00 (confidence 0.7), then drops to Eb for 8 bars (confidence 0.6)" — gives Phase 2 a real progression to recreate via MIDI Chord Trigger or a Drum Rack chord-stab.
- chordDetail.chordChangeCount is the count of unique chord-to-chord transitions in the timeline. Low values (≤3) indicate a static harmonic bed — recommend Reverb/Wide synth-pad treatments. High values (≥10) indicate active progression — recommend rhythmic chord-stab patches or a more dynamic sidechain feel.
Expand Down Expand Up @@ -493,6 +504,12 @@ abletonRecommendations
- {"device":"Glue Compressor","deviceFamily":"NATIVE","trackContext":"Master","workflowStage":"MASTER","category":"MASTERING","parameter":"Makeup","value":"2 dB","reason":"A PLR of 15 LU shows a dynamic, peaky master with headroom, so gentle bus glue plus makeup lifts perceived loudness before the limiter without crushing transients.","phase1Fields":["plr","lufsIntegrated"],"advancedTip":"Slow Attack (10-30 ms) so transients pass, Ratio 2:1-4:1, aim for 1-2 dB of gain reduction, then follow with a Limiter at -0.3 dB Ceiling."}
- Valid sidechain example card (emit when pumping is measured — see SIDECHAIN ENVELOPE; omit when pumpingConfidence < 0.2):
- {"device":"Compressor","deviceFamily":"NATIVE","trackContext":"Bass","workflowStage":"MIX","category":"DYNAMICS","parameter":"Sidechain","value":"On (Kick trigger); Threshold -22 dB, Ratio 6:1","reason":"Recreates the measured quarter-note ducking by routing the kick to the Compressor sidechain, with attack and release read from the envelope.","phase1Fields":["sidechainDetail.pumpingStrength","sidechainDetail.pumpingConfidence","sidechainDetail.envelopeShape32"],"advancedTip":"Set Release so gain returns by the recovery point shown in envelopeShape32 — a return by step 6 of 32 at the measured BPM is a short release; switch to the Glue Compressor for quarter/eighth glue."}
- Valid bass-dynamics example card (emit when bassDetail.averageDecayMs is measured — omit when bassDetail.transientCount < 2):
- {"device":"Glue Compressor","deviceFamily":"NATIVE","trackContext":"Bass","workflowStage":"MIX","category":"DYNAMICS","parameter":"Release","value":"a Release matched to the measured bass decay (rolling/sustained → 300 ms-1.2 s; punchy → 50-120 ms)","reason":"Sets the Glue Compressor release from the measured bass decay so the compressor recovers in step with the bass envelope instead of pumping against it.","phase1Fields":["bassDetail.averageDecayMs","bassDetail.type"],"advancedTip":"For a punchy bass (short averageDecayMs) switch to a Compressor with a fast attack to catch transients; keep the Glue Compressor for rolling or sustained basses where level glue matters more than transient control."}
- Valid acid filter-movement example card (emit when acidDetail.isAcid is true — hedge when acidDetail.confidence < 0.5):
- {"device":"Auto Filter","deviceFamily":"NATIVE","trackContext":"Bass","workflowStage":"SOUND_DESIGN","category":"EFFECTS","parameter":"Resonance","value":"high Resonance scaled to the measured acid resonance, LFO Amount up with the LFO tracking the measured centroid oscillation","reason":"Recreates the measured 303-style squelch by setting Auto Filter resonance from the detected acid resonance level and modulating the cutoff at the measured centroid-oscillation rate.","phase1Fields":["acidDetail.resonanceLevel","acidDetail.centroidOscillationHz"],"advancedTip":"Use a Lowpass Filter Type and add Drive for grit; when acidDetail.confidence is below 0.5 keep Resonance moderate and treat the movement as a steer rather than a committed patch."}
- Valid delay/echo example card (emit for sparse leads, stabs, or chops that support audible repeats):
- {"device":"Echo","deviceFamily":"NATIVE","trackContext":"Lead","workflowStage":"MIX","category":"EFFECTS","parameter":"Feedback","value":"≈30% with Sync on and the repeat division derived from the measured BPM (dotted-eighth for dub/house)","reason":"Adds a tempo-synced echo whose division and feedback follow the measured BPM so the repeats lock to the track grid instead of smearing it.","phase1Fields":["bpm","rhythmDetail.tempoCurve"],"advancedTip":"When rhythmDetail.tempoCurve is non-flat, Echo re-syncs to the drift; raise Feedback (40-60%) for a dub tail or lower it (15-25%) for a tight slap, and roll the repeats' highs off with the Echo filter so they sit behind the dry signal."}
- Valid stereo example card (emit when sub-bass is not mono — omit when subBassMono is already true):
- {"device":"Utility","deviceFamily":"NATIVE","trackContext":"Bass","workflowStage":"MIX","category":"STEREO","parameter":"Bass Mono","value":"On","reason":"Folds the low end to mono because the measured sub-bass correlation sits below 0.85, which risks phase cancellation on club systems.","phase1Fields":["stereoDetail.subBassCorrelation","stereoDetail.subBassMono"],"advancedTip":"Leave Bass Mono's crossover near 120 Hz; if subBassCorrelation stays low afterward, the stereo energy comes from a specific stem and should be narrowed at the source."}
- Valid key/scale example card (emit when keyConfidence ≥ 0.7 — hedge instead when below):
Expand Down