As you may now, I have been working a lot on reworking nova multimedia engine (avos) audio pipeline architecture to get better lipsync. This has led me to try many different architectures that were not really converging.
This release takes yet another path: instead of trying to correct A/V drift after it appears, Nova now anchors video timing directly to the audio position that is actually reaching the speakers.
The audio engine now handles AudioTrack and software tempo changes with separate backend-specific clocks, making speed changes, seeks, pauses, and resumes much less prone to accumulated lipsync errors.
- Improve audio/video sync, including passthrough (DTS-HD, TrueHD), playback speed changes, and resume/seek scenarios
- Fixed A/V desync during playback speed adjustments and after seeking
- Improved Trakt sync: minimum resume threshold to avoid recording accidental short plays, plus reliable resume saving on pause/stop
- Added Trakt sync status and skip reason visibility in settings
- Fixed OpenSubtitles login errors and added OpenSubtitles status display in settings
- Skip subtitle fetch timeout for local files to prevent double external player launch
- Fixed unintended subtitle background rendering and synced subtitle position with the player action bar
- Fixed TV player overlay visibility and TV channels not refreshing after creation
- Prevented metadata loss and unscraped files when moving files
- Improved scraper robustness for long titles, ambiguous names, and localized queries (german tranliteration)