Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Issue 2212 #2302

Merged
merged 2 commits into from
May 14, 2020
Merged

Issue 2212 #2302

merged 2 commits into from
May 14, 2020

Commits on May 14, 2020

  1. Widen useful range of BitWriter's write_quniform() 

    The versions of write_quiniform() used by the arithmetic packer and the uncompressed bitwriter vary considerably.  For some reason, the BitWriter uses a much 'narrower' implementation that overflows with inputs of more than a few bits.

This patch replaces the BitWriter's implementation with one like in the arithmetic packer.  This allows its use in coding non-uniform tile header fields.
    @xiphmont
    xiphmont committed May 14, 2020
    Configuration menu
    Copy the full SHA
    490a3c9 View commit details
    Browse the repository at this point in the history

  2. Fix for xiph#2212; Crash when using 4 tiles for 1080p 4:2:2 input 

    When doing loop filter RDO inline with the rest of the tile coding,
LRUs must align to tile boundaries.  An unexpected corner case means
that chroma LRUs must have an even superblock width in 4:2:2 video, as
LRUs must always be square.  As a result, that means tiles must also
have an even superblock width.

As tile width must be adjusted in this case, it also means we can't
use the spec's 'tile uniform spacing' mode, which would produce odd
superblock width tiles in, eg, 1080p 4:2:2 video. This patch also
implements explicit per-tile sizing the the frame OBU header.
    @xiphmont
    xiphmont committed May 14, 2020
    Configuration menu
    Copy the full SHA
    f322748 View commit details
    Browse the repository at this point in the history