Native video support for AV1 #7557
Conversation
Deployed docs
|
emilk
force-pushed
the
emilk/native-av1-decode
branch
from
a03b115
to
d98b785
Compare
| @@ -0,0 +1,268 @@ | |||
| //! Video demultiplexing. | |||
There was a problem hiding this comment.
todo note to self: review this later again, checking what has moved / is new etc.. This stuff changed quite a bit so there's a risk of regressing progress, should keep an eye on this
There was a problem hiding this comment.
I can make a separate PR where I just move the code on main, and then we can rebase and compare
There was a problem hiding this comment.
much simpler decoder now, nice!
Looking good, but I stick with my usual line of complaining about seqcst: Seqcst is a bad default and hard to deal with imho (Mara has a really nice section about that in her book if you want to learn more check https://marabos.nl/atomics/memory-ordering.html#common-misconceptions -> Myth: Sequentially consistent memory ordering is a great default and is always correct.)
# Conflicts: # crates/store/re_video/src/lib.rs # crates/store/re_video/src/mp4.rs # crates/viewer/re_viewer_context/src/cache/video_cache.rs
What
What
Supports native decoding of AV1 videos.
Downsides: it is extremely slow in debug builds
In release builds it is ok, but there is still A LOT of performance on the table.
TODO before merging
nasmtopixi.toml, because it is needed to compilerav1drav1dopt-in so users don't neednasmto compilererunrav1dis difficult on some platformsFix performance of debug builds, if possibledav1dis always fast, butrav1dis super-slow in debug buildsProof
native-video.mp4
Checklist
mainbuild: rerun.io/viewernightlybuild: rerun.io/viewerCHANGELOG.mdand the migration guideTo run all checks from
main, comment on the PR with@rerun-bot full-check.