-`config/perfect-machine.example.json` declares the contract shape for six cameras, six microphones, reservoir typed views, native authorities, outputs, and the demotion of bridge scripts to tooling.
-`native/reservoir` is the first native Rust crate. It implements one
shared-edge five-second rolling buffer with typed views and proves that
-`config/audio-field.example.json` declares the actual distributed six-mic rig: two Kiyo mics, two PS Eye mics, local Focusrite shielded cardioid, neighbor Focusrite shotgun on the co-streamer, two local speakers, placeholder geometry, clock domains, capture policy, calibration sweep settings, and FOA AmbiX ACN/SN3D output.
-`scripts/audio_field.py` lists devices, validates shared or distributed profiles, checks local distributed sources, generates calibration sweeps, summarizes clock-domain sync requirements, preserves shared-input capture helpers, and encodes already aligned offline WAVs to FOA.
- The live visual producer writes a multi-LOD scene cache with source kind and priority. Real Leap frames are promoted as the highest-priority visual timing/spatial evidence; Leap fallback frames remain lower-priority diagnostics.
- Live clap calibration is wired into the visual producer. It keeps a rolling frame window from the Kiyo pair plus Leap, reads the live spatial audio frame for transient candidates, publishes `clap-events.msgpack`, injects clap calibration markers into the render frame, and writes clap evidence into the LOD cache. Kiyo stereo owns the current rough 3D solve; Leap owns the best visual timing witness until its geometric model is calibrated.
- Clap peaks now update a per-camera clock sync model. The point-cloud builder can sample camera history buffers at `oracle_time - camera_offset` instead of using whichever frame arrived last.
- The live visual producer now uses an explicit five-second spatiotemporal reservoir. RGB/Leap frame history expires from the newest shared sample, the multi-LOD cache only accepts evidence inside that same window, and `RenderFramePacket.source_time_min_ns/source_time_max_ns` declares the reservoir slice being rendered.
- Multi-LOD cells now carry first-pass material estimates: weighted albedo plus roughness/metallic hints derived from source evidence. Aquarium should treat this as relightable material evidence to refine, not as the final BRDF solve.
- The Spout/audio overlay boundary clamps stale visual packets to the current audio reservoir edge. If camera geometry ages out, it is dropped instead of being broadcast out of sync.
- Live fallback-only RGB/Leap mode exists for deadline operation when OpenCV/MSMF/DirectShow reads block longer than the five-second reservoir. It keeps the reservoir and OBS output alive while real camera capture is moved to nonblocking ingress.
- OpenCV camera reads now run behind `LatestFramePump` workers so blocking drivers cannot freeze the fusion hot loop. The live deadline command uses lower per-frame sample density (`--rgb-room-step 16 --leap-step 16 --points 64`) so the reservoir/TAA path stays fresh enough for Spout to render visual points inside the five-second budget.
-`localcast.sensor_fusion.chirp_pose` converts live phase-field delay meaning for Kiyo/PS Eye microphone sources into camera-body range constraints and per-camera pose-correction estimates. It can load camera-mic and speaker geometry from the audio-field profile, with the example profile used as the live fallback until measured local geometry exists. The diagnostic visual producer can inject those constraints as `camera-chirp:*` and `camera-pose-correction:*` render points plus `chirp-camera-pose` LOD evidence.
-`scripts/setup_obs_synced_program.py` derives OBS-controllable stems from an aligned program audio timeline: host voice, co-streamer voice, ambient, transients, co-streamer loopback, and local loopback.
-`scripts/capture_co_streamer_surfaces.py` captures neighbor Focusrite and neighbor loopback with local loopback ground truth, estimates the late remote-family offset, and writes aligned co-streamer surfaces for the stem packer.
-`scripts/wasapi-loopback-capture.ps1` is the direct primary-playback loopback path. It must run in the neighbor's interactive console session; SSH-only capture sees the device but receives no render packets.
-`audio_field.phase_meaning` extracts actionable delay, coherence, confidence, suppression, correction-energy, and active-probe need from internal phase/chirplet evidence.
-`localcast.audio.phase_field` is a typed CultCache document at `calibration/runs/audio-phase-field.msgpack`.
-`localcast.diagnostics.audio_phase_field` can replay an aligned mic field plus known program/loopback reference into the phase-field document for diagnostics. It is not a live bridge authority.
-`localcast.diagnostics.audio_phase_field` can write `audio-phase-field-status.json` during replay so estimator behavior remains inspectable while the native ingest path is built.
-`audio_field.active_probe` wires low phase-field confidence into `ActiveProbeOptimizer`, emits bounded chirplet WAVs plus `active-probes.jsonl`, and can play probes through the default output device.
-`localcast.diagnostics.audio_live_field` preserves the old local WASAPI/PortAudio capture experiment as diagnostic code only. Production local mic and loopback ingest belongs in native capture workers that append typed handles through `LocalcastProducer`.
- Active probe maintenance only targets source ids backed by live local capture devices. Missing distributed channels remain stable placeholders in the mic/phase field, but they do not consume chirp budget.
- Active probe artifacts are bounded: emitted chirps rotate through a fixed slot set and the manifest rotates at a byte cap. The calibration loop is allowed to be noisy; it is not allowed to become an unbounded filesystem leak.
-`localcast.audio.mic_field` publishes aligned six-mic float32 blocks at `calibration/runs/audio-mic-field.msgpack`.
-`faust/localcast_voice_separation.dsp` is the first Aquarium-hosted graph surface for host voice, co-streamer voice, ambient, transient, and loopback stems.
