The audible autotest used one beep for eight steps, so steps couldn't be told apart.
Each step now has its own self-made sound (synthesize.py): bell (map Baby_Cry_1), chime
(F6 post, Buoy_Bell patch 88), wood block (F8), triangle (F10), low horn (F11 landing),
click (pop rule), whistle (Digit5), shaker (Digit6), boing (honk rule). HUD: one compact
box top-left, every line fits one row, live countdown; camera return after step 3.
- Retail (the recomp, reference only): the vehicle collision callback sets the alarm only
while the car is parked (StayingParked) and the contact vector exceeds 0.1
(livingworld vehicle characteristics); any collider; every contact restarts the alarm and
parked timers; StopAlarming after 8 s; no pull-out while alarming; no impact sound.
- Engine: VehicleParked, VehicleImpact { vehicle, by, impact }, CarAlarmRule
(AlarmTuning from the setup export world_tuning.vehicle_alarm), VehicleAlarmStarted;
the alarm holds horn state 6 for whole console frames past the time.
- Mods: world_audio event 'impact', traffic option parked, sdk.world_audio.alarm_rule,
read(key).alarm. The world-audio test mod's parked taxi alarms on contact.
- No-mod e2e bench byte-identical (84 / 84); tests; doc 15, traffic spec.
Six identical taxis at one rpm and level, each detuned once by C04's own Random op,
beat at 4-8 Hz (the "lawn mower" heard in the audible autotest). Each taxi now drives its
own ring (radius 3+3i m, 2+2i m/s), headless engine-band modulation at single-taxi level.
mods/audio-content-test runs every action on a timer once the map's native audio runs
(settings autotest, autotest_step, autotest_gap, autotest_muted) and logs one
AUTOTEST <check> ok|fail line per read-back check; one bottom-left HUD box with the step,
what to listen for and a red MUTED TEST header on muted runs. Doc 16 M8b.
- sdk.world_audio.spawn(key, 'traffic'|'ped', ...) without slots gets its own instance on
the private MixMap, as mod emitters do; slots = 'shared' (alias 'retail') opts into
retail's pools. Capability world_audio 4. Engine-published objects are unchanged.
- Full private pool (16 cars + 16 peds, all mods): the nearest play, the rest in reach wait
silently; never refused, never in retail's pools. read(key).waiting / .slots,
info().own, a one-time log line.
- Dev mods: audio-content-test uses the default; world-audio-test asks for 'shared'
(retail's limits) with an "Own instances" setting.
- No mod active: e2e bench byte-identical (84 / 84); suites pass (known failures only).
Completes the "Later" list (doc 16 section M):
- Hot swap: an audio mod turning on / off is swapped into the running audio in place.
Banks keep their runtime id and constructor order, held posts are re-bound with their
payload; Splice trees, wheel streams, projects, tuning and speech swap too. Only a
MixMap or grain-bed change (or a swap error) still restarts the sound.
sdk.audio.info().swaps / last_change.
- Hot reload: editing audio.json or the files it names swaps the overlay while the
script keeps running.
- MixMap inputs: sdk.audio.set_mixmap_input / mixmap_inputs drive retail's own
controllers (validated, owned, 16 per mod / 64 total, released values restored).
- Own instances: world_audio spawn {slots='own'} plays mod cars / peds on a private
MixMap (16 + 16), retail pools untouched.
- Csis projects: audio.json add.projects ships a mod's own classes, functions and
globals; name clashes are refused; installed / removed without a restart.
- Seed: sdk.audio.seed / SKATE_AUDIO_SEED seeds all audio generators.
- Rules: play.frame = 'owner' puts offsets in the owner's axes; rule sounds follow
moving published emitters.
- Capabilities: audio 4, audio_content 3, audio_events 3, world_audio 3.
- Dev test mod: Digit1-6, a second HUD line, its own self-made dev.csi.
- No mod active: e2e bench byte-identical 84 / 84 after every step.
- Mod emitters get their own instances by default (ModEmitterSlots::Extra); "shared" stays
selectable, SKATE_AUDIO_MOD_EMITTER_SLOTS overrides either way.
- Mod WAVs route through the native mixer by default when the native audio runs and a voice
is free, else the Bevy path; native = true / false forces one. Positional native sounds
default to a 40 m squared reach.
- Replace / layer rule sounds are positional: `at` = owner (default, followed each frame,
optional offset), world (fixed position) or centre; optional falloff; default reach is the
owner's retail one (emitter record, cars 40 m, peds 50 m, skaters 30 m).
- Dev test mod, Skyline and audio-example use the defaults; SDK docs and doc 16 updated
(section L questions written as decisions).
- No mod active: e2e bench byte-identical, 84 / 84; new tests for each default.
- Mod WAVs can play through the game's own mixer (per-mod mixer banks, retail emitter
words, a private MixMap), beside the existing Bevy path (mod_voices.rs).
- Tuning can change while the game runs: sdk.audio.set_tuning / tuning / tuned, applied
between passes, exact restore on nil or mod disable (tuning.rs, audio_tuning.rs).
- Mods add world emitters and reverb zones through the published WorldEmitter /
ReverbZoneVolume components (world_audio kinds).
- Sound rules: mods mute, replace or layer the game's own sounds by identity, from
audio.json `rules` or sdk.audio.rule (mod_rules.rs, audio_rules.rs); empty or
rules-only overlays no longer restart the sound.
- Capabilities: audio 3, audio_tuning 1, audio_events 2, audio_content 2, world_audio 2.
- Dev test mod: F8 native voices, F9 tuning, F10 emitter + reverb zone, F11 a mute rule.
- Headless e2e bench byte-identical to the #36 head after every step; doc 16 section L.
- Mod crossfade banks play: crossfade layouts come from the bank's program or from
`add.crossfade_layouts` (crossfade.rs, crossfade_layouts.rs); dev test mod fade sample.
- Speech clips and takes are checked at merge time and by check_mod.
- Empty Lua tables (`{}`) were parsed as maps and rejected by list fields; one helper
(lua_list.rs) now reads them as empty lists in every list field of every mod command.
- Traffic lights answered in the world-audio spec (Skate 3 has none).
- Doc 16 and PULL-REQUESTS.md updated.
Brings optimisation pass 2, the tools split (#37) and the world gaps
(tazer, ped body falls, speech echo / voice details, main-cast speech,
NPC loose-board slide) under the audio modding layer:
- main-cast speech takes go through the content overlay (archive
"maincast"), like the living world's;
- ped speech (phone ring), body-fall and foot-plant Splice starts and the
NPC skaters' starts are recorded through the observe-only event wrapper;
tazer posts get a slot name; new tags "tazer" and "body_fall"; main-cast
line starts are speech rows with class "maincast";
- the new world tuning (ped_objects, speech_voice) is overlay-reachable
through tuning.world like the rest of the world tuning.
Tazer and PedBodyFall ported from the retail mechanism and checked against the recordings; ped speech gets its PEAK filter, the voice float, the speech inputs, the phone and photographer lines, first-free-stream choice and the queue clock; the speech echo send (pre-gain send -> high pass -> distance delay -> low pass -> environment bus); the main-cast speech channel for pros, NPC skaters' own speech and message pairs, with its opt-in setup decode; the NPC skater's board slide. Engine and mod surfaces extended (reactions, tazer, body fall, phone / photo events); retail values are setup-data defaults. Local player output byte-identical.
- Ped speech plays: speech manager -> library -> the living world's two streams, with queue / interrupts and the
near / far lines by distance; levels and filters from the retail mechanism (filters equal to the recomp's, gain
median within 0.5 %). Silent and logged when the opt-in speech decode is absent.
- NPC skater: wheels and clothing on its own instance and a bail grunt; tricks, treatment and footsteps stay off for
NPCs as in retail.
- Per-model ped data (shoe class, voice, far threshold) and traffic engine records per model exported by setup;
traffic relative-speed input and front / rear sound points; world hosts process before the mixer tick.
- Test mod: opt-in (off by default), higher world-object limits (48 per mod, 128 total), a per-second summary log.
- Specs published in docs/hails-additions/audio-specs/; code comments point at them; test data paths behind env vars
(no .local defaults).
Review fixes:
- The audio host takes each physics step exactly once (one-step pulses latched, nothing while paused or silenced,
at most 4 steps per frame), so pushes and other one-frame events are never dropped or doubled. e2e at 60 fps and
above is byte-identical; 30 / 45 fps change only by the latch and the step pair.
- Removed the emitter PROFILES fallback, the stray trace dump and the audition script; data-gated tests are
#[ignore = "needs the private install data"] and fail loudly when run without it; per-post log lines behind
SKATE_AUDIO_TRACE; checked bank parsing, finite output, camera-cut Doppler reset, fewer render allocations.
- The 15 verdict-settled A/B switches are removed; their "on" behaviour is the only one (byte-identical).
World / NPC audio, hookable by engine systems and mods:
- Map-change fix: world and NPC hosts reset when the map's banks are unloaded, so surviving objects sound again.
- Engine surface (skate-game::world_audio): TrafficAudio / PedAudio / NpcSkaterAudio components, horn / alarm /
speech messages, read-back of who is audible; retail limits (4 cars within 40 m, 15 peds within 50 m with
footsteps for the nearest 3, one NPC skater within 30 m), measured in the recompilation.
- Mod surface (Lua API 2, `sdk.world_audio.*`); test mod `mods/world-audio-test/` (cars, peds, a ghost skater).
- Per-skater audio state builder (local player output identical); remote players take the NPC skater slot; an
opt-in non-retail "more audible" setting.
- Docs: doc 15 "Hooking up world audio", doc 11 sections, PR plan.
- 09-water.md: physics, camera, splash and look after in-play testing
- 10-sky-shader-test.md, README index
- PULL-REQUESTS.md: PR H prepared against upstream 60efdef
- water-test-teleport: F7 Aletown canal, F5 position readout, the
"Teleported" text clears after 5 s
- PULL-REQUESTS.md: pending edits to the upstream PR plan
- mods/water-test-teleport: dev-only mod for testing water (F2/F3/F4
drop into University fountain, University reservoir, DownTown fountain;
Shift+F2/F3/F4 stand beside them). Not intended for upstream.
Add generalized engine access and challenge mods, repair SKATE turn handling, and improve collision prediction and physics transitions. Reduce mod snapshot and material update overhead, preserve authored GLB materials with lighting and shadows, and improve CPU profiling.
Apply probe SH to online customiser characters after remote bind, restore received sun shadows, and ship Skyline driving, fullscreen, and renderer improvements on the branch.
Co-authored-by: Cursor <[email protected]>
Respawn the chassis when a session starts so solid-bridge replication publishes a fresh body, and show local publish counts in replication status.
Co-authored-by: Cursor <[email protected]>
Branch channel now reports already-current and missing-release cases clearly instead of a generic 404 error, and remote smoke/skids initialize without requiring local F10 spawn.
Co-authored-by: Cursor <[email protected]>
Expose append-only mesh buffers for mod skid/smoke rendering, sync rear-wheel VFX over multiplayer, rebuild Skyline audio assets, and let the updater install from any CI branch.
Co-authored-by: Cursor <[email protected]>
Un-ignore bundled Skyline GLBs, ignore local docs and the BUILD-RELEASE.bat wrapper while keeping scripts/Build-Release.ps1 in tree.
Co-authored-by: Cursor <[email protected]>
Sanitize non-finite SIMD lanes in offboard air selection, add solid dynamics and mod replication APIs, improve setup refresh without re-extracting assets, and harden Build-Release.ps1 for older PowerShell hosts.
Co-authored-by: Cursor <[email protected]>
Hails