[mythtv-users] .nuv format, mythtv, xine, mplayer, unichrome
jules at coredevelopers.net
Fri Oct 14 10:45:17 EDT 2005
Craig Tinson wrote:
> Jules Gosnell wrote:
>> to follow up on my own posting...
>> for it to be a format issue is a bit wierd - because I thought that
>> mythbackend would just be taking the data, already encoded, straight
>> from my Nova-T and filing it, without doing any form of transcoding ?
>> But my Nova-T has not changed and my version of Myth and possibly
>> kernel, has.....
>> I guess this is where my lack of knowledge of exactly what is going
>> on under the covers stops me from doing any more digging.
>> Can anyone lend a hand ? Maybe I am seeing the interaction of several
>> related issues ?
>> Jules Gosnell wrote:
>>> I don't know much about video file formats so bear with me.
>>> I am looking into why some of my recordings do not play well (with
>>> mythtv-0.18.1) on my frontend box (Epia Via Ezra 1gz with Unichrome
>>> enabled) but play fine on another machine (Opteron 64-bit 2gz) -
>>> pretty obvious you might think :-)
>>> The recordings are made from a Hauppauge DVB Nova-T on the backend
>>> (mythtv 0.18.1) - and not transcoded, because the Nova-T kicks out
>>> MPEG2, which is what the Epia's h/w can assist with.
>>> Recordings that I made some time ago, all seem to play fine with
>>> mythtv, xine and mplayer.
>>> Some more recent recordings (perhaps since I upgraded to 18.1?) are
>>> a little odd - they exhibit sound 'stickiness' using MythTV on the
>>> Epia, but play fine on the Opteron. Initially I put this down to
>>> issues with the Unichrome driver, or the speed of the Epia, but have
>>> discounted both (xine produces a lot of video and audio artefacts
>>> when playing these recordings with AND without -Vxxmc - and, when
>>> with -V xxmc, cpu usage is usually below 10%).
>>> Since it is possible to play the recordings without issue on the
>>> opteron, I know that the artefacts are not part of the original
>>> signal. So, I began thinking in terms of data format. I played
>>> various recordings using mythtv, xine and mplayer on both boxes.
>>> Recordings that do not exhibit the sound issue running on
>>> mythtv/Epia seem to play fine in xine and mplayer no matter which
>>> box they are on.
>>> Recordings that have the sound issue, do not. Using mythtv they play
>>> fine on the Opteron, but the sounds sticks on the Epia. Using xine,
>>> they exhibit artefacts and sound stickiness on both boxes. MPlayer
>>> sometimes picked the wrong audio channel (1 instead of 0) and also
>>> sometimes suffered from lipsyncing issues...
>>> So, it looks like something has changed in my setup which is causing
>>> me to occasionally generate less portable recordings than I used to,
>>> and that these are causing problems on my Epia (probably because the
>>> unichrome h/w (which I cannot seem to prevent mythtv using) is
>>> tripping up on the format).
>>> If xine (v0.99.4.), mplayer (1.0pre7try2-3.2.3) and, possibly,
>>> unichrome are having problems with mythtv-generated MPEG2 files that
>>> mythtv can play without issue, is this a bug in mythtv, the other
>>> three, or a grey area in the MPEG2 spec - or have I misread the
>>> symptoms ?
>>> Any help with this wouldbe much appreciated. I can probably find
>>> somewhere to put files that exhibit the problem up for ftp.
>>> Thanks for mythtv,
> this is just another stab in the dark that came to mind - do you have
> any differences in the "general playback" section of setup between the
> two machines? deinterlacing playback settings for example?
I think the only difference is XvMC support selected on the Epia (and
deselecting does not seem to prevent its use!). But this does not take
into account the fact that the particular files with which MythTV has a
problem on the Epia (most play fine on both machines) also seem to be
problematic to xine, mplayer on both boxes....
keep trying :-)
> like I said just a stab in the dark - first thing that came to mind
> mythtv-users mailing list
> mythtv-users at mythtv.org
"Open Source is a self-assembling organism. You dangle a piece of
string into a super-saturated solution and a whole operating-system
crystallises out around it."
* Jules Gosnell
* Core Developers Network (Europe)
* Open Source Training & Support.
More information about the mythtv-users