Files

T

jpmschweitzerandClaude Opus 5.5 ddce4441a9 refactor(tooling): T-1290 — the assets domain, where OFF is the normal case

tooling/db/ (a misnamed directory: connectors, not database work),
trellis-batch.sh and synth_ui_sounds.py become `reach assets`:
audio {health,generate,batch,post {convert,normalize,trim,pipeline}},
image {health,generate}, trellis {health,generate,batch}, and synth-ui.
The four audio bash wrappers are retired, and tooling/db/ is gone.

Parity, from baselines taken before anything moved:

- the four UI-sound WAVs and the harmonic-synth WAVs (exponential and linear
  decay) are byte-identical
- the ffmpeg pipeline's decoded PCM is identical. Its .ogg bytes are not,
  even between two runs of the OLD code: Ogg picks a random stream serial,
  so the encoded file was never the right thing to compare
- the network success paths can't be run in a gate (Stable Audio and Trellis
  are kept off, Gemini costs money), so tooling/test_assets.py stands up a
  fake Gradio and pins every payload: the audio submit, Trellis's six-call
  session sequence with its 9-input image_to_3d, and the Gemini body. It
  failed when one Trellis value was mutated (7.5 → 7.0)

Failure classification, in endpoints.py, is the point of the port. The
services are OFF by design (VRAM on tower-of-joy, D-17), and the topology doc
warns against "fixing" one by restarting it. So a refused connection says OFF
and asks for the service to be turned on rather than restarted; a 4xx/5xx says
the request was rejected; 401/403 says credentials; 429 says quota; and an
unreachable Gemini blames the network, not VRAM.

Behaviour changes, each a failure that used to read as success or crash:

- audio batch and trellis batch exited 0 with failures in their summaries;
  they now print the summary and exit 1
- trellis generate on a missing image crashed with a TypeError
  (print(..., indent=2)); it now names the file, and checks it before the
  service so a typo is not reported as an outage
- the ffmpeg pipeline left its intermediates behind when a step failed

Structure: the connectors called each other as subprocesses (batch spawned
the connector, which spawned audio_post) and parsed each other's stdout. They
are now function calls, and ffmpeg is the only exec, through core/process.
ensure_venv() is removed: it os.execv'd into .venv, which D-263's exec rule
forbids, and reach declares the dependencies itself. config.json moved into
the domain deliberately, and the local-services rule follows it.

Output contract: results are still JSON on stdout with the same keys, so skill
readers keep working. Failures are an exit status with a Fix line, never
{"ok": false}. The audio-gen, glb-gen and image-gen skills, Araminta's agent
file and the allow-list are updated to match. glb-gen's "trellis-batch.sh is
hardcoded to one category" caveat is gone: batch takes --input-dir or --names.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

2026-09-23 19:49:20 +02:00

7.9 KiB

Raw Permalink Blame History

title, description, type, status

title	description	type	status
UI Audio Assets	Interface sound assets for cursor, implant, monologue chimes, and weapon aim with generation specs	design	active

UI — Audio Assets

Interface sounds triggered by player interaction, insert systems, and cognitive events. Non-positional (AudioStreamPlayer). Routes to UI Sounds bus.

Generation Approach

All UI sounds: Generated via Stable Audio Open with sonic family prefix prompts, then trimmed/normalized/converted via reach assets audio post pipeline.
Monologue chimes: Replaced in Sprint 10 (#327) with manual synthesis. Insert-tech aesthetic: pure sine harmonics, mathematical envelope, no SAO. Previous S9 SAO versions were acknowledged placeholders per D-038 amendment.

Assets

ID	Filename	Status	Bus	Method	Duration	Prompt/Notes	Sprint
UI-001	`cursor_hover.ogg`	done	UI Sounds	SAO	77ms	Subtle tick/ping on entity hover (D-056). Insert-tech: clean, precise, synthetic. Trimmed + normalized to 44.1kHz in S9 #453.	S7 #440, S9 #453
UI-002	`implant_open.ogg`	done	UI Sounds	SAO	217ms	Rising tone when insert UI opens (radial menu, inventory, stance). Neural lattice powering up. Trimmed + normalized in S9 #453.	S7 #440, S9 #453
UI-003	`fog_recognition.ogg`	done	UI Sounds	SAO	400ms	Warm organic chime at onset of cognitive delay (D-060). Re-generated in S9 — original was silent. Organic prefix + Chimes template.	S7 #440, S9 #453
UI-004	`weapon_aim.ogg`	done	UI Sounds	SAO	127ms	Harder click/lock for weapon aim state. Mechanical, deliberate. Trimmed + normalized in S9 #453.	S7 #440, S9 #453
UI-005	`sfx_monologue_chime.ogg`	done	UI Sounds	manual synthesis	800ms	Production-quality manual synthesis in S10. Insert-tech: 1200Hz fundamental + harmonics (2400Hz -12dB, 3600Hz -24dB, 6000Hz -30dB). 15ms attack, 0.2 sustain ratio, exp decay. Peak -15dB. No SAO — pure mathematical precision.	S7 #440, S9 #453, S10 #327
UI-006	`sfx_monologue_chime_urgent.ogg`	done	UI Sounds	manual synthesis	800ms	Production-quality manual synthesis in S10. Insert-tech: 1220Hz fundamental (detuned +20Hz for tension) + stronger harmonics (2440Hz -6dB, 3660Hz -18dB, 5087Hz -24dB, inharmonic 5087Hz -28dB). 8ms attack, 0.3 sustain ratio. Peak -6.3dB.	S7 #440, S9 #453, S10 #327
UI-007	`sfx_weapon_aim_lock.ogg`	done	UI Sounds	SAO	505ms	Sharp targeting lock-on confirmation for weapon aim. Downsampled to 44.1kHz + normalized in S9 #453.	S8 #440, S9 #453
UI-008	`sfx_stance_change.ogg`	done	UI Sounds	SAO	349ms	Soft mechanical mode-switch click for stance toggle. Downsampled to 44.1kHz + normalized in S9 #453.	S8 #440, S9 #453

Mix Specs (not audio files)

These are AudioBus parameter changes, documented in docs/audio/dialogue-ambient-dip.md:

Spec	Trigger	Effect	Sprint
Dialogue dip	Dialogue box opens	Ambient -6 to -8dB, 300ms ease-in, 500ms ease-out	S7 #440
Confrontation dip	Confrontation dialogue	Ambient -10 to -12dB + LP (800Hz), World SFX -4 to -6dB, 500ms ease-in, 1000ms ease-out	S7 #440

Detailed Entries

UI-001: cursor_hover

Sonic family: insert-tech
Frequency range: 3-4kHz primary
Design principle: "Like breathing — present, functional, invisible." Player shouldn't describe it if asked, but would notice absence. Fatigue test critical — plays 300+ times per session.
Generation: Pure synthesis. Sine tone at 3.2kHz, exponential decay over 40-50ms. No SAO — precision and purity are the point.
Debouncing: AudioManager enforces 100ms cooldown between plays (rapid mouse movement).
Integration: CursorRenderer.cursor_state_changed signal → AudioManager.play_ui("cursor_hover")
Cross-reference: D-056 (cursor states)

UI-003: fog_recognition

Sonic family: organic
Frequency range: 300Hz-1.5kHz
Design principle: "The sound you hear internally when something clicks into place." Not a chime — more like a soft exhalation of tone. Sits BETWEEN insert sounds and ambient — not UI, not world, cognition.
Timing (Q-014 resolved): Fires at ONSET of cognitive delay. Sequence: hear something in fog → chime plays → 0.6s delay → monologue during delay → blob transitions to D-033 color. Chime is "your character is processing," color transition is "recognition complete."
Integration: Cognitive delay onset event → AudioManager.play_ui("fog_recognition")
Cross-reference: D-060 (cognitive delay), D-018 (three-range sound, medium range)

UI-007: sfx_weapon_aim_lock

Sonic family: insert-tech
Frequency range: 1-3kHz primary
Design principle: "Crosshair snapping to bone." Clinical confirmation that the weapon system has acquired a target. Harder and more aggressive than cursor_entity_lock — this is the weapon variant. Still UI feedback, not a combat sound. Should not startle.
Generation: SAO 5s generation, trimmed to best 0.5s transient. Insert-tech prefix + UI template. CFG 8, 100 steps.
Integration: cursor_state_machine.gd weapon_aim_lock signal → AudioManager.play_ui("sfx_weapon_aim_lock")
Cross-reference: D-056 (cursor states), weapon_aim.ogg (aim enter, different from lock)

UI-008: sfx_stance_change

Sonic family: insert-tech
Frequency range: 1-2kHz primary
Design principle: "A precision dial clicking one notch." Subtle neural insert feedback on stance toggle. Player may toggle stances many times while positioning — must be short, pleasant, fatigue-safe. No long tails, no reverb. Nearly subliminal.
Generation: SAO 5s generation, trimmed to best 0.35s transient. Insert-tech prefix + UI template. CFG 8, 100 steps.
Debouncing: AudioManager should enforce 150ms cooldown (rapid stance toggling).
Integration: stance_indicator.gd stance_changed signal → AudioManager.play_ui("sfx_stance_change")
Cross-reference: D-048 (neural insert feedback aesthetic), D-053 (movement stances)

UI-005 / UI-006: Monologue Chime Design Brief

Target emotional register (achieved Sprint 9 #453):

The monologue chimes are the two poles of augmented cognition made audible:

Variant	Meaning	Feel	Target
Normal	"My lattice surfaced this"	A thought arriving — not a notification. Like your own attention shifting.	Player never consciously notices it. Pavlovian: text appeared, brain registered, moved on.
Urgent	"My lattice flagged a contradiction"	Productive discomfort. "Wait, what did I miss?"	Micro-spike of alertness. Not alarm. The delta from normal must be SMALL but unmistakable.

Relationship to fog recognition chime:

Fog recognition = organic (biological cognition, warm, 300-800Hz)
Monologue normal = insert-tech (crystalline, clean, higher register, ~1.2kHz)
Monologue urgent = insert-tech (sharper, brighter, adds second harmonic ~2.4kHz)
Same cognitive family, different branches. Cousins, not twins.

Quality bar: If the urgent chime feels like a quest marker ping, it has failed. If it feels like Ubisoft, it has failed. The delta between normal and urgent = "someone thinking quietly" vs "someone's eyes going slightly wide."

Sprint 9 generation spec:

Normal: SAO 5s generation (insert-tech prefix + chimes template), CFG 7, 100 steps. Trimmed to 0.8s (attack + initial decay). LUFS normalized to -16.
Urgent: SAO 5s generation (insert-tech prefix + chimes template, sharper/brighter descriptors), CFG 8, 100 steps. Trimmed to 0.8s. Stronger attack (-6.3dB peak vs -15.3dB normal), more sustained energy.
Both at 44.1kHz stereo, Vorbis q6 encoding.

Evaluation gate: First internal playtest: "Does the normal chime feel like a thought arriving? Does the urgent chime create a physiological response?" If no → iterate with manual synthesis.

7.9 KiB Raw Permalink Blame History