Release v1.9.0
Release v1.9.0
feat: content-derived credits scroll duration, Whisper artifact filtering, and stop state cleanup
- Compute EndCredits scrollDurationMs dynamically from content height (totalTravelPx / scrollSpeedPxPerMs) and fade in chained tracks smoothly.
- Filter out Whisper noise/music artifact tags ([Music], [NOISE]) and silence hallucinations in SherpaSTTEngine.kt to prevent invalid user turns.
- Add minimum audio duration (0.45s) and RMS energy validation in LiveVoiceOverlay.kt before running STT.
- Remove [stopped] string concatenation and prevent duplicate assistant response bubble creation on cancellation in ChatViewModel.kt.
- Bump version to 1.9.0 (versionCode 59).