mirror of
https://github.com/SerenityOS/serenity
synced 2026-05-09 08:32:04 +02:00
As long as possible, entire decoded frame sample vectors are moved into the output vector, leading to up to 20% speedups by avoiding memmoves on take_first.