kajisho5/ffmpeg-skill v2.22.4 on GitHub

Automated release: version and notes generated from pull requests merged since 2.22.3.

export.py keeps subtitle and data streams. It mapped nothing, so ffmpeg's default selection wrote one video and one audio stream: on 2.22.1, an .mp4 with a mov_text track exported with --preset youtube or --preset copy came out h264 + aac, and the result said nothing. Found by an agent that muxes soft subtitles before a platform export and had to refuse the pair. Every preset but gif now keeps the track: copied when the container holds it, or converted to the container's text track (mov_text in .mp4/.mov, SubRip in .mkv). A platform trim (-t) cuts it with the picture. A data stream the container refuses (a camera's timecode in .mp4) is retried away on its own, so it does not take the subtitles with it, and the retry is a non_av_streams_dropped fallback naming what it dropped. The result gains non_av_streams: {kept, dropped} (each source stream, its output_codec or the reason it is missing, read back from the written file) and dropped_non_av_streams, with a notes line when something was dropped. When the source has such a stream, the picture and the sound are mapped as the first non-cover video and the first audio stream (what probe describes); a source without one gets the same command as before.
fix(export): keep subtitle and data streams, and report them (#380)

What's Changed

fix(export): keep subtitle and data streams, and report them by @kajisho5 in #380

Full Changelog: v2.22.3...v2.22.4