[MPlayer-dev-eng] VAAPI/X11 video output driver

Mark Thompson sw at jkqxz.net
Tue Nov 24 22:57:55 CET 2015

On 23/11/15 14:22, compn wrote:
> On Sun, 22 Nov 2015 17:04:17 +0000
> Mark Thompson <sw at jkqxz.net> wrote:
>> Hi,
>> See patch attached: apply to current trunk mplayer, configure, make
>> mplayer, run with -vo vaapi on anything using recent Intel graphics
>> (or something else with VAAPI, maybe).
>> This is very much in a "works for me" state.  If it's useful to anyone
>> else then they can use it as-is, and if it or some derivative is
>> sensible to integrate into mplayer then that would be nice.  There are
>> probably lots of edge cases I haven't considered which crash it.  It's
>> also quite possible that I have broken other things with these changes
>> (especially VDPAU and XvMC in the libavcodec driver).
> wow, nice!
> rxt also posted a patch trying to port mplayer-vaapi back to mplayer
> did you see it? from october:
> http://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/2015-October/073163.html
> it needs testing. apparently you are the one to ask about testing! :)

Ok, I've had a look at that now.  The approach is substantially similar 
to what I have (put the decoder initialisation inside the video output), 
but it has much more complete support for the extra oddments (OSD, 
opengl output).  However, it doesn't have the feedback path for the 
decoder to indicate that reference frames will not be used again, so it 
will fail on any input stream with a complex GOP structure - I'm 
guessing this is the problem with some streams which the submitter 
described.  Also it lacks H.265, though that's very easy to add.

I will combine my patch with this one, and hopefully come up with 
something sensible.

> did you look at mpv's vaapi code for answers to any of your questions?
> mpv is based on mplayer, although a lot of code has been changed.
> https://github.com/mpv-player/mpv/blob/master/video/out/vo_vaapi.c
> looks like mpv used the vaapi code from mplayer-vaapi project:
> https://github.com/gbeauchesne/mplayer-vaapi

mpv has rearranged a lot of the files so I can't really compare it 
directly.  I think the internals are mostly equivalent, but the 
arrangement is more sensible (no decoder initialisation in the output, 
in particular).

> probably my mail isnt what you were expecting...
> -compn

Patch to follow soonish.


- Mark

More information about the MPlayer-dev-eng mailing list