isu-admin/hearbit-ai-app
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 2 Wiki Activity

fix: meeting mode triggers only on system audio, not speech

Browse Source 
Reverts speech-based fallback trigger. Meeting mode now strictly
requires system audio energy > 0.005 to trigger auto-start.
The system_queue always has data (even silence) from ScreenCaptureKit,
so checking queue emptiness caused false positives. Speech alone
must never trigger meeting mode.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
This commit is contained in:
michael.borak
2026-02-06 10:45:53 +01:00
parent b911eefe72
commit 2c93afce9e
1 changed files with 7 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
src-tauri/src/audio_processor.rs
View File

@@ -256,27 +256,20 @@ impl AudioProcessor {
// MODE-SPECIFIC TRIGGER LOGIC: // MODE-SPECIFIC TRIGGER LOGIC:
// "voice" -> Trigger if user speaks (VAD) // "voice" -> Trigger if user speaks (VAD)
// "meeting" -> Trigger ONLY if system audio energy detected. // "meeting" -> Trigger ONLY on system audio energy above threshold.
// Speech alone is NOT enough (prevents false triggers when // Speech alone NEVER triggers meeting mode to prevent
// user talks near the mic without being in a call). // false starts when user speaks near mic without a call.
// However, speech + ANY system audio (even below threshold) // The threshold (0.005) is low enough to catch ringtones,
// counts as a trigger, to catch Electron apps (e.g. Nextcloud Talk) // call audio, and notification sounds from any app.
// that may produce low-level system audio.
let sys_queue_has_data = if let Ok(q) = self.system_queue.lock() { !q.is_empty() } else { false };
let system_has_any_audio = max_system_energy > 0.0001 || sys_queue_has_data;
let trigger = if self.recording_mode == "voice" { let trigger = if self.recording_mode == "voice" {
self.is_speech_active self.is_speech_active
} else { } else {
// Primary: system audio above threshold system_active
// Fallback: speech detected AND system audio stream has ANY data
// (catches low-volume call audio from Electron/WebRTC apps)
system_active || (self.is_speech_active && system_has_any_audio)
}; };
if trigger { if trigger {
// Trigger Detected! // Trigger Detected!
println!("Auto-Start: Call detected (SysEnergy: {}, Speech: {}, SysHasData: {}). Flushing pre-roll...", max_system_energy, self.is_speech_active, system_has_any_audio); println!("Auto-Start: Trigger! (Mode: {}, SysEnergy: {:.4}, Speech: {})", self.recording_mode, max_system_energy, self.is_speech_active);
self.waiting_for_speech = false; self.waiting_for_speech = false;
// Flush Ring Buffer (Orderly: from ring_pos to end, then 0 to ring_pos) // Flush Ring Buffer (Orderly: from ring_pos to end, then 0 to ring_pos)