mirror of
https://github.com/fergalmoran/ladybird.git
synced 2025-12-28 04:08:08 +00:00
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.