[MPlayer-dev-eng] Re: [MPlayer-users] ssa reader for mplayer using vf_overlay

[Robert Henney]

On Thu, Sep 29, 2005 at 10:06:05AM -0400, Jason Tackaberry wrote:
> > ok, now for an odd one.  If the slave command get_time_pos is sent while
> > mplayer is paused, the video starts playing again.  while I think this
> > may be an mplayer bug, something interesting occurs if the overlay
> 
> Yes, mplayer can't respond to slave commands while it's paused, and the
> result of trying is that the video suddenly unpauses.  With vf_overlay,
> the exception to this is the overlay slave command.
> 
> I'd like to see mplayer handle more commands while paused, especially
> seeking and loadfiles.  (My local copy of mplayer has this ability.)
> get_time_pos I suppose would be nice too, but it's not necessary,
> strictly speaking.  Your app should know mplayer's state, and should
> know that when it's paused, the current time is whatever time was last
> outputted on the status line.  (For my applications, I just parse the
> status lines to get current time/pos information.)

I've been trying to get away with using "-quiet", since my app can
follow its own clock for the timing of all subtitle events and every so
often does a get_time_pos to resync its clock with the player.

this looks like its just not going to be possible though.  other than
through the realtime status line, pause states and jumps forward and
backward do not get told to the program connected via the slave channel.

So I'll be removing "-quiet" and adjusting my code to work off the
status line for timing.

> > happens to be visible.  the video plays extremely jumpy.  to get rid of
> > the jumpyness, one needs only hit space twice to re-pause and then
> > continue playing.  could vf_overlay be maintaining a state during pause
> > mode that it doesn't get out of because the unpause was caused by an
> > irregular event it might not have heard?
> 
> This one sounds weird for sure.  Which vf_overlay patch are you using?
> (What's the date in the filename?)

vf_overlay-2005-09-14.diff

I've also not being able to reproduce this state through manual entry in
slave mode, so there's more to it apparently than what I described
before.  I will have to try and figure it out before I change my code to
use the realtime status line, as it will be a non-issue after that.

-- 
robh