[mythtv-users] Re: mpeg2 playback performance issue makes no sense

Joseph A. Caputo jcaputo1 at comcast.net
Tue Oct 12 19:49:16 UTC 2004


On Tuesday 12 October 2004 15:17, Michael J. Lynch wrote:
> Grant Edwards wrote:
> >On Tue, Oct 12, 2004 at 11:06:55AM -0400, Isaac Richards wrote:
> >>>>If so, what could account for the large difference in
> >>>>processor usage between "watch videos" and "watch recordings"
> >>>>playing the same file?
> >>>
> >>>AFAICT, Myth's SW decoding/playback functions are an order of
> >>>magnitude slower than those used by xine and mplayer.
> >>
> >>All of the above use the same underlying code (libavcodec).
> >>You've got a configuration problem if you can't play things
> >>back properly.
> >
> >Any suggestions on where to look for such a configuration
> >problem?  It would be pretty cool if I could use my laptop as a
> >front end.
>
> So...since mplayer, xine, and myth all use the same underlying libs
> to do play back, isn't it reasonable to expect that playing the same
> file in any one
> of these should result in nearly identical processor usage numbers if
> configurations
> are close to the same?
>
> As Grant states, if it is a configuration problem, what are the most
> likely items that could so drastically affect the amount of processor
> usage?  My first guess would be deint after that, I wouldn't have a
> clue since I'm not using any filters or other fancy stuff.

Grant's problem may not even be related to the OP's issue.  Since he's 
using a laptop, it's entirely possible that there is no (or buggy) Xv 
support; laptop displays are notorious for these kinds of things.  
Mplayer (& xine, I think) both support alternate output methods like 
SDL, so it's possible that they might be OK where Myth wouldn't.  The 
output of mythfrontend during playback might also be helpful...

-JAC


More information about the mythtv-users mailing list