[vlc-devel] [PATCH v2 1/7] clock: handle the normal play time (npt) from the input
ajanni at videolabs.io
Wed Apr 7 08:31:45 UTC 2021
On Wed, Apr 07, 2021 at 09:53:08AM +0200, Thomas Guillem wrote:
> On Sat, Apr 3, 2021, at 10:22, Rémi Denis-Courmont wrote:
> > Le torstaina 1. huhtikuuta 2021, 18.30.06 EEST Thomas Guillem a écrit :
> > > vlc_clock_t will send corrected update points directly.
> > >
> > > The player won't have to fetch the normal_time from the demux and fix
> > > its timer accordingly.
> > I'm not convinced that the clock should know/care about NPT.
> The clock doesn't use it for any calculation. It is just used when passing the pts/system_date points to the clock listener (the player) and only for information purpose.
If the clock doesn´t use it, maybe it can be sent directly
from the es_out to the player there:
static void ClockUpdate(vlc_tick_t system_ts, vlc_tick_t ts, double rate,
unsigned frame_rate, unsigned frame_rate_base,
es_out_id_t *es = data;
es_out_sys_t *p_sys = container_of(es->out, es_out_sys_t, out);
input_SendEventOutputClock(p_sys->p_input, &es->id, es->master, system_ts,
ts, rate, frame_rate, frame_rate_base);
Thus, you´d just have to store the NPT in the es_out, where
it should mostly make sense, and the player won´t have to
fetch it from the demux itself either.
> > And if we need to track NPT per group, that seems rather problematic, because
> > there will be multiple NPTs for a single clock...
> Yes, NPT are per group, but there is also one vlc_clock_main per group, so if there are multiple NPTs, they will be handled by their respective clock (main).
> > --
> > 雷米‧德尼-库尔蒙
> > http://www.remlab.net/
> > _______________________________________________
> > vlc-devel mailing list
> > To unsubscribe or modify your subscription options:
> > https://mailman.videolan.org/listinfo/vlc-devel
> vlc-devel mailing list
> To unsubscribe or modify your subscription options:
More information about the vlc-devel