[Bug 793876] h265parse: add support for 'Format range extensions profiles'

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=793876

--- Comment #14 from Guillaume Desmottes <gdesmott at gnome.org> ---
(In reply to Nicolas Dufresne (stormer) from comment #7)
> I think it would be better to match a compatible profile instead of doing an
> exact match. As an example, we know that we have bitstream using a unamed
> profile:
> 
>   {"uname", 1, 1, 1, 1, 0, 0, 0, 0, TRUE},
> 
> This is compatible with "main-422-10", just that max_8bit_constraint_flag is
> set. As per the spec a "main-422-10" compliant decoder should support this,
> there is just no specific profile name for it. In this implementation, the
> profile would not be set, and then any decoder will endup being tried.
> 
> So I think we should expose a profile name that are compatible. That's
> closer to what we do in H264 parser, where we simply check the IDC and if
> certain flags are set. To do se, we'll need to search for a profile that
> match, and then keep the one that has the least extra constraints. In our
> example, main-422-10 would have 1 extra contraints, while main-422-12 would
> have two, so main-422-10 would be a better valid profile choice for the
> custom bitstream.

In this example we actually have 2 profiles having 1 extra contraint:
main-422-10 and main-444. Which logic should the parser use to pick one rather
than the other?

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.