Keep timestamps aligned when joining many audio files into one. Use when:
(1) subtitles, transcript segments, chapters or bookmarks drift later and
later through a concatenated file; (2) the joined file is seconds longer than
the sum of its parts; (3) ffmpeg's concat demuxer prints "Application
provided invalid, non monotonically increasing dts to muxer"; (4) you are
stitching TTS chunks, per-chapter downloads or recorded segments and need
measured durations to add up. Explains why a lossless copy keeps each piece's
gapless padding, why frame counts still do not match, when re-encoding once
is the right trade, and how to verify drift rather than assume it.
