AutoSubSync v6.5

• LAPSE Sync Tool Integration: Added LAPSE (Language-Agnostic Playback Synchronization Engine) v2.0.1 as a first-class synchronization engine in both the GUI and Headless CLI (assy-cli):

  Features Auto, Split (with adjustable split penalty for commercial cuts and scene edits), No Split (single constant shift), and OLS (linear regression for progressive framerate drift) modes, along with embedded subtitle track detection and voice profile caching. Cross-platform auto-downloading and runtime Silero VAD neural network integration via ONNX Runtime are supported out of the box across Windows, macOS, and Linux. See BENCHMARKS.md for details.
  Supports selecting a specific audio track (--audio-track N) and subtitle track (--sub-track N) for synchronization. Mentioned in issue #84.

• ffsubsync Upstream Upgrades & PGS Timing Extraction:
  • Added support for piecewise alignment split penalty slider (--split-penalty), PGS subtitles reference (--pgs-ref-stream), multi-segment fast sync (--multi-segment-sync), and remote streaming URLs (http://, https://, rtmp://, rtsp://, ftp://).
  • Implemented standard Matroska Show/Clear packet timing reconstruction in call_ffsubsync to support PGS subtitle streams where container packet durations are omitted (duration_time=N/A), enabling high-accuracy, fast (~7s) reference synchronization across MKV, M2TS, and Blu-ray media.
  • Added seamless automatic fallback to audio Voice Activity Detection (VAD) if reference PGS timings are unreadable or missing. Fixes issue #10 by @babywin.
• ffsubsync Silero VAD Support: Implemented an ONNX Runtime fallback engine in call_ffsubsync to enable high-accuracy Silero VAD speech detection without requiring heavy PyTorch installations. Fixes issue #72 by @tyhawk.
• CLI Embedded Subtitle Extraction: Added embedded subtitle extraction support to the headless CLI (assy-cli), allowing command-line sync pipelines to automatically extract and synchronize against internal subtitle tracks in video containers. Contributed by @vicente-higino in PR #92
• ALASS Rename Prompt Options: Added settings to customize and disable the filename prompt when ALASS encounters bracketed filenames ([ / ]). Contributed by @vicente-higino in PR #91
• File Access & Lock Validation: Added pre-sync file accessibility and permission validation. When a subtitle or reference file is locked or open in another application (such as Subtitle Edit or across SMB network shares), AutoSubSync now logs clear, localized error messages in the log window instead of silently failing. Mentioned by @enchained in issue #89
• Autosubsync .ass Subtitle Conversion: Fixed an issue where syncing .ass subtitles with the autosubsync engine produced 0-byte output files. AutoSubSync now automatically converts .ass and unsupported formats to .srt before synchronization. Mentioned by @bgtsai in issue #95
• Settings & Config Persistence: Fixed an issue where GUI dropdown selections (such as save locations and default sync tool) could revert to defaults upon application restart due to localized language string mappings.
• Custom Suffix Handling with Overwrite Mode: Fixed custom suffix behavior when configured with overwrite mode to save the synced subtitle alongside the original file with the custom suffix instead of clobbering the original.
• Build System & Docker Improvements: Added uv fast package manager support in build scripts and CI/CD workflows, made PyInstaller build hooks cross-platform for Python 3.12, and made the Docker build self-contained. Contributed by @ale94mleon in PR #83

Downloads

• Windows (x64): AutoSubSync-windows-amd64.zip
• macOS (Apple Silicon - M1/M2/M3): AutoSubSync-macos-arm64.zip
• macOS (Intel): AutoSubSync-macos-amd64.zip
• Linux (AppImage): AutoSubSync-linux-amd64.AppImage
• Linux (Tarball): AutoSubSync-linux-amd64.tar.gz

MacOS users please read the instructions.

Included:
ffmpeg: 7.1
ffsubsync: 0.5.1
lapse: 2.0.1
alass: 2.0
autosubsync: 1.0.1