OK. So basically this means 3 possible upstreams: optirun, Qt or VLC, since I guess this is beyond Debian.
I think you can close the issue. Actually I managed to find a workaround that I will report at https://forum.videolan.org/viewtopic.php?f=13&t=144411 as I still believe that there is a regression in VLC 3.0.2. 2018-06-10 18:42 GMT+02:00 Sebastian Ramacher <sramac...@debian.org>: > On 2018-06-05 10:13:19, Antoni Marcinek wrote: > > > VA-API was not automatically probed in VLC 2.2. This is expected. > > > VDPAU fails, and then VLC falls back to software decoding. > > > > > > In VLC 3.0, VA-API probing fails due to insufficient hardware > > capabilities, so > > > VLC uses software decoding as well. So no differences. > > > > > > -- > > > Rémi Denis-Courmont > > > > Sorry, I didn't notice this message earlier. Thank you. > > > > OK, so it is not about hardware acceleration. However something did > change > > between VLC versions in Debian what changes behaviour of software > decoding. > > The title of the bug report is wrong now, but I guess the issue of > software > > decoding should still be addressed. Should I send output of vlc -vvv for > > VLC 3.0.2? > > > > > > Still it is not clear who should address the problem of optirun + VLC-Qt. > > As mentioned in another subthread: please seek help from someone who's > familiar > with optirun, ideally someone who's also familiar with Qt and VLC. > > Cheers > -- > Sebastian Ramacher >