Factor the Phase 3 attribution (you_spans -> mask -> diarize far side -> merge +
name) into a shared phase3_attribute() used by both stop and reprocess. stop now
writes mic_activity.json (the 'You' spans) next to a retained audio.wav
(plaintext timing metadata, local-only, lives/dies with audio.wav); reprocess
prefers that timeline over the voiceprint fallback so re-transcription keeps
correct per-stream 'You'/'Speaker' attribution.
stop_recording now, when the mic is on, takes 'You' spans straight from the
MicActivity timeline and diarizes the recording with those ranges zeroed, so
sherpa only ever clusters the far side into Speaker N. Merges the span lists and
names them via the existing voiceprint map (You + Speaker 2..). Falls back to
the whole-signal pass + voiceprint match when the mic is off. Adds mask_ranges
helper + tests.
Records, per 100ms of audio.wav frame time, when the mic (the user) was
speech-level — captured in the loopback writer where mic and loopback exist
separately in the recording's own timebase. you_spans() collapses active
windows into merged 'You' ranges. The far side will be diarized with these
ranges masked out, so clustering never sees the mic. Struct + span logic land
first (tested); capture wiring + stop attribution follow.
WASAPI loopback delivers no packets when the render endpoint is idle, so the
recording only advanced while system audio played. The bridged mic (the user
talking over silence) was lost to MicBridge overflow, and audio.wav lost those
stretches entirely — corrupting diarization and dropping the user's solo
remarks in real meetings. When loopback is idle, synthesize silence for the
elapsed gap and fold the mic in, keeping audio.wav continuous. Loopback +
mic-enabled only.
VoiceSample kept the first N samples unconditionally, so a user silent for the
first 30s got a voiceprint of room noise/a cough. That noise embedding
spuriously matched a cluster and mislabeled every speaker 'You', even during
silence. Skip sub-threshold (VOICE_ENERGY_FLOOR) chunks so the window tracks
actual speech; until ~1s of real voice accumulates, the match returns empty
(no premature 'You').
Reprocess re-diarizes and upserts the new labels but never deleted the old
ones, so get_meeting (Participants pane) kept returning the discarded run's
labels (e.g. 83) while the transcript showed the new 5. Clear the meeting's
speaker rows before finalize re-inserts the fresh set.
Phase 2 backend. The 15s provisional pass now (1) runs the mic voiceprint
match every tick and merges 'You'/'Speaker N' into the session name map without
overwriting user-set names, and (2) re-emits every committed segment whose
speaker changed via transcript://segment (replace-by-id) so the live transcript
reflects the refined labels instead of a frozen 'S1'.
Phase 1. stop_recording writes voiceprint.wav next to a retained audio.wav
(ADR-0009 gated, sealed at rest with the vault); reprocess_transcript now
re-diarizes the fresh transcript and rebuilds the speaker name map from the
voiceprint instead of reusing the stale original-run names, fixing the
collapse to a single 'You' speaker on re-transcription.
Phase 0 of the speaker-accuracy plan. sherpa fast-clustering threshold 0.5
over summed mic+loopback mono exploded a 2-speaker call into 83 clusters;
raise to 0.7 (larger threshold -> fewer clusters) and drop sub-700ms spans
whose embeddings are unstable and only churn labels in assign_by_overlap.
A migration-checksum mismatch (an edited-after-shipping migration) made
SqliteStore::connect() return Err, which .expect() turned into a panic=abort
fast-fail: the window never opened and the OS reported only a 0xc0000409 crash.
Replace the expect with a blocking MessageBoxW that names the failure and points
at wa.db, then exit(1) cleanly.
iamdoubz
iamdoubz