I-IAL9000/shrinkerr v0.3.38

on GitHub

Fixed

• Reverted -fflags +flush_packets from v0.3.37. It was added to make ffmpeg's time= field advance steadily on files with many subtitle streams, but it forces the matroska muxer to end every cluster early — creating per-packet overhead that empirically cost ~20% throughput in single-process tests (4.02× vs 5.15×) and compounded badly under concurrent NVENC sessions. Symptom: progress bar stalled for minutes on Breathless-style files even though ffmpeg was running. Kept the -max_muxing_queue_size 9999 half of the v0.3.37 fix — that's defensive and has no observable cost.

Docker images

Pinned to this release:

docker pull ghcr.io/i-ial9000/shrinkerr:0.3.38         # portable (amd64 + arm64, CPU)
docker pull ghcr.io/i-ial9000/shrinkerr:0.3.38-nvenc   # x86_64 + NVIDIA GPU (NVENC)
docker pull ghcr.io/i-ial9000/shrinkerr:0.3.38-edge    # as above but ffmpeg master
docker pull ghcr.io/i-ial9000/shrinkerr:0.3.38-edge-nvenc

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