On Fri, Apr 18, 2014 at 12:27 PM, Janne Grunau <[email protected]> wrote: > On 2014-04-17 20:38:27 +0200, Anton Khirnov wrote: >> >> On Tue, 25 Mar 2014 11:13:37 +0100, Janne Grunau <[email protected]> >> wrote: >> > On 2014-03-24 21:20:27 +0100, Luca Barbato wrote: >> > > On 24/03/14 20:26, wm4 wrote: >> > > > In practice, it's called when the decoder is flushed, i.e. on seeking. >> > > > And yes, this is a bit of a bother. Freeing all hardware surfaces and >> > > > reallocating them etc. makes seeking quite a bit slower this way. >> > > >> > > I'm not sure how to make it more optimal, if you have ideas I'm all >> > > hears. >> > >> > I don't think it necessary to call get_format() unconditionally after >> > seeking. Fixing it in h264.c might be not so nice though. >> > >> > I guess h264.c is not the only decoder which needs fixing. I'll look >> > into it. >> > >> >> Any progress? > > Not fully convinced from patch below. Without changes in the stream > get_format will be only called once. That might break applications > using the codec context from avformat_find_stream_info. > One the plus side it's non-ugly and adding support for all other codecs > is a triviality. >
If its supposed to be the all-mighty place to init hwaccels, you also need to call it when frame size changes, but pixel format stays the same, so that the hwaccel gets a chance to re-init. Not sure if some hwaccels have other requirements to be notified of certain changes. - Hendrik _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
