v0.9.16

Fixed

• Text-sub extraction no longer times out on large files. v0.9.15 dropped the extraction time cap, so ffmpeg demuxed the whole multi-GB file and blew the 15s timeout → empty → und. Restored a 30-min cap (covers dialogue start; longer than v0.8.1's 10 min) and raised the timeout to 60s.

Added

• Configurable Whisper model via SHRINKERR_WHISPER_MODEL (default tiny). tiny is under-confident on non-English speech (Nordic tracks land ~0.5–0.58, just under the 0.6 gate); set base/small for better accuracy. Threshold remains tunable via SHRINKERR_LANG_DETECT_AUDIO_MIN.

Docker images

Pinned to this release:

docker pull ghcr.io/i-ial9000/shrinkerr:0.9.16         # portable (amd64 + arm64, CPU)
docker pull ghcr.io/i-ial9000/shrinkerr:0.9.16-nvenc   # x86_64 + NVIDIA GPU (NVENC)
docker pull ghcr.io/i-ial9000/shrinkerr:0.9.16-edge    # as above but ffmpeg master
docker pull ghcr.io/i-ial9000/shrinkerr:0.9.16-edge-nvenc

Floating tags (:latest, :edge, :nvenc, :edge-nvenc) track the most recent release of each lineage.