yt-dlp's `--audio-format alac` writes lossy AAC

Fact

When FFmpegExtractAudioPP re-encodes, it replaces the codec's own ffmpeg options with the quality arguments (more_opts = self._quality_args(acodec)). ALAC is the one codec whose encoder lives only in those options — 'alac': ('m4a', None, ('-acodec', 'alac')) — so ffmpeg is handed no codec at all and picks the .m4a default: AAC. The file gets the extension ALAC would have, and nothing warns. The same line drops -f adts for aac, which then comes out as MP4 rather than raw ADTS.

Present in yt-dlp 2026.08.19 and on master as of 2026-09-23.

Why it matters

The failure is silent: someone asking for lossless audio gets lossy audio with a plausible file name, and only a codec probe tells them apart.

How to apply

Pass the encoder again as output arguments for the extract step; they are appended after yt-dlp's own, so they win, and they stay harmless once the bug is fixed:

yt-dlp -x --audio-format alac --ppa "ExtractAudio+ffmpeg_o:-acodec alac" URL

In the Python API: postprocessor_args={"extractaudio+ffmpeg_o": ["-acodec", "alac"]}. Check lossless output by codec, never by extension: ffprobe -v error -show_entries stream=codec_name -of csv=p=0 FILE.