Release v26.04.15.1 — Dynamic voice silence threshold #24

Merged
bvandeusen merged 0 commits from dev into main 2026-04-15 00:38:26 -04:00
bvandeusen commented 2026-04-15 00:38:10 -04:00 (Migrated from git.fabledsword.com)

Summary

  • Voice silence detection now auto-calibrates to the user's mic + room via session peak tracking. Once peak arms above -20 dBFS, "silent" means current level ≥15 dB below peak; before that a conservative static fallback (-35 dBFS) keeps dead-silent sessions from spinning. Grace period (1500 ms) at the start gives you time to begin speaking.
  • Fixes the "I always have to click the mic again to end the recording" bug on Android.

Test plan

  • Flutter analyze clean (CI)
  • Record a message in a normal-volume room — should auto-stop within ~2s of finishing
  • Record in a quiet room — should still auto-stop via fallback threshold
  • Record with background noise — dynamic peak should adapt
## Summary - Voice silence detection now auto-calibrates to the user's mic + room via session peak tracking. Once peak arms above -20 dBFS, "silent" means current level ≥15 dB below peak; before that a conservative static fallback (-35 dBFS) keeps dead-silent sessions from spinning. Grace period (1500 ms) at the start gives you time to begin speaking. - Fixes the "I always have to click the mic again to end the recording" bug on Android. ## Test plan - [x] Flutter analyze clean (CI) - [ ] Record a message in a normal-volume room — should auto-stop within ~2s of finishing - [ ] Record in a quiet room — should still auto-stop via fallback threshold - [ ] Record with background noise — dynamic peak should adapt
This repo is archived. You cannot comment on pull requests.
No Reviewers
No Label
1 Participants
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: bvandeusen/FabledApp#24